Mechanistic aspects of the selective reduction of NO by propene over alumina and silver-alumina catalysts

Article abstract of DOI:10.1006/jcat.1999.2622

The SCR of NO with C₃H₆ in the presence of a large excess of O₂ was studied over γ-Al₂O₃, 1.2% Ag/γ-Al₂O₃, and 10% Ag/γ-Al₂O₃ catalysts. NO₂ formation during the C₃H₆-SCR of NO over γ-Al₂O₃ was achieved not through direct oxidation of NO with O₂ but through an alternative mechanism involving the formation of organo-nitrite species followed by their decomposition/oxidation. The promoting role of low loadings of silver on activity for N₂ production was attributed to the higher rate of formation of inorganic ad-NO_x species, which further reacted with the reductant or a derived species to yield various organo-NO_x compounds. It was suggested that organo-nitro and organo-nitroso compounds and/or their derivatives (isocyanate, amines, cyanide, and NH₃) react with NO or the organo-nitrite and/or its derivative NO₂ to form N₂. Without a reductant, the 1.2% Ag/γ-Al₂O₃ catalyst was poisoned by strongly bound nitrates and its activity for NO₂ formation was similar to that observed over the alumina.

Full text of DOI:10.1006/jcat.1999.2622

Journal of Catalysis 187, 493–505 (1999)
Article ID jcat.1999.2622, available online at http://www.idealibrary.com on
Mechanistic Aspects of the Selective Reduction of NO by Propene
over Alumina and Silver–Alumina Catalysts
1,2
1
F. C. Meunier, J. P. Breen, V. Zuzaniuk, M. Olsson, and J. R. H. Ross
Centre for Environmental Research, University of Limerick, Limerick, Ireland
Received April 28, 1999; revised June 28, 1999; accepted June 28, 1999
which hinders the development of improved materials in
The selective catalytic reduction of NO with C3H6 in the pre- terms of activity, selectivity, hydrothermal stability, and sul-
sence of a large excess of O2 (i.e., C3H6-SCR (selective catalytic phur resistance. Burch et al. have contributed to the eluci-
reduction)) was studied over � -Al2O3, 1.2% Ag/� -Al2O3, and 10%
Ag/� -Al2O3 catalysts. The � -Al2O3 and the low-loading silver ma-
terial exhibited high conversions to N2 whereas the high-loading
sample predominantly yielded N2O. Surprisingly, a comparison of
actual NO2 yields to thermodynamically predicted yields of NO2
is kept reduced and a decomposition-type mechanism oc-
showed that the formation of NO2 during the C3H6-SCR of NO over
dation of the mechanism of the SCR reaction over Pt/Al2O3
materials; in particular, the different effect of using propene
or propane on the reaction pathways has clearly been es-
tablished (3, 4). With propene, the surface of the platinum
curs on the noble metal, leading to the formation of sig-
nificant amounts of N2O along with N2. With propane, the
surface of the noble metal is covered with oxygen species
�
-Al2O3 was not achieved through the direct oxidation of NO with
O2. An alternative mechanism involving the formation of organo-
nitrite species followed by their decomposition/oxidation was sug-
gested to be the main route for the formation of NO2. The promot- which favours the oxidation of NO to NO2 and/or ad-NOx
ing role of low loadings of silver on alumina on the activity for N2 species which subsequently react with the reductant over
production was attributed to the higher rate of formation of inor- the alumina, selectively yielding N (the so-called “NO2
2
ganic ad-NOx species (e.g., nitrates) as evidenced by in situ DRIFTS
and thermogravimetric analyses. It was proposed that these inor-
ganic ad-NOx species further react with the reductant or a derived
species to form various organo-NOx compounds. In particular,
organo-nitro and organo-nitroso compounds and/or their deriva-
Alumina is one of the most active single-metal oxides for
tives (e.g., isocyanate, cyanide, amines, and NH3) were suggested to
reactwithNOortheorgano-nitriteand/oritsderivativeNO2 toyield
route”). The latter route underlines the importance of the
alumina, in the selective reduction reaction scheme, which
isnotmerelyasupportfortheplatinumbutalsoparticipates
directly in the reduction reaction.
the C3H6-SCR of NO and can be further promoted by a
wide range of metal oxides such as cobalt (5), copper (6),
or silver (7). The exact role of these compounds on the en-
N2. When no reductant was present, the low-loading Ag/� -Al2O3
material was poisoned by strongly bound nitrates and its activity
for NO2 formation was similar to that observed over the alumina. hancement of the catalytic activity of the alumina is not triv-
ial. In the case of Co/Al2O3, it has been suggested that the
Key Words: NO; NO2; propene; silver; alumina; DRIFTS; role of the promoter is to oxidise NO with O2 to form NO2
�
c 1999 Academic Press
de-NOx; reaction mechanism.
which subsequently reacts with the alkene on the alumina
to yield N2 (5, 8). This reaction mechanism would thus be
similar to the NO2 route observed over Pt/Al2O3 in the case
of the C3H8-SCR of NO. This assumption relies mostly on
the fact that (1) over the alumina, the C3H6-SCR of NO2
to N2 proceeds much faster than that of NO and (2) the
C3H6-SCR of NO2 over alumina proceeds much faster than
the C3H6-SCR of both NO and NO2 over the Co-promoted
sample. However, other results published on the oxidation
of NO to NO2 over Co/Al2O3 materials apparently contra-
dict this view (9). The measured activity for NO oxidation
to NO2 was found to be too low to be able to account for
the related SCR activity. It has been suggested that the de-
activation of the catalyst by strongly bound surface nitrates
is possible when no reductant is available (10). The work by
Yan et al. also stresses the importance of various phases of
1
. INTRODUCTION
In spite of much research work over the last 10 years
or so (1, 2), the selective catalytic reduction (SCR) of NO
with light hydrocarbons or oxygenated molecules is not yet
a practical method to control NOx emissions in the pres-
ence of a large excess of oxygen. This failure is probably re-
latedtothecomplexityofthereactionmechanismsinvolved
1
To whom correspondence should be addressed: Fax: +353-61-
2
02602. E-mail: John.Paul.Breen@ul.ie. Fax: +49-89-289-13544. E-mail:
meunier@thor.tech.chemie.tu-muenchen.de.
2
Current address: Technische Universit a¨ t M u¨ nchen, D-85748
Garching, Germany.
493
0021-9517/99 $30.00
Copyright �c 1999 by Academic Press
All rights of reproduction in any form reserved.

4
94
MEUNIER ET AL.
cobalt species present on the alumina-based material (e.g., NO2 yield observed. To obtain more detailed information
oxidic particles of different sizes, cobalt aluminate, and iso- about the mechanism of the C3H6-SCR of NO over alu-
lated Co2 species).
+
mina and Ag/Al2O3 catalysts, the present study compares
The Ag/Al2O3 materials seem to exhibit a different be- the catalytic activity of alumina and two silver-based ma-
haviour as compared to that of the Co/Al2O3 catalysts. terials with low (i.e., 1.2%) and high (i.e., 10%) loading of
For the latter, increasing the cobalt content, on one hand, silver and relates this to the surface species observed during
only favours the nonselective combustion of the reductant. in situ DRIFTS experiments.
On the other hand, the use of increasing contents of sil-
ver on the alumina seems also to favour a different route
for the reduction, with greater proportions of N2O be-
2
. EXPERIMENTAL
ing formed along with N2 (11). A similar trend has been 2.1. Catalyst Preparation and Characterisation
observed on Ag/TiO2–ZrO2 materials (12). Bethke and
Kung have also reported a synergistic effect between the
The � -Al2O3 utilised was supplied by Alcan (AA400)
2
� 1
with a total surface area of 148 m g . For the preparation
of the silver-promoted materials, an appropriate amount of
silver nitrate (BDH, analytical grade) was dissolved in a
volume of deionised water equal to that of the porous vol-
ume of the alumina. The solutions were then deposited on
the alumina by dry impregnation at room temperature. The
Al2O3 and the Ag/Al2O3 catalysts (11). They have postu-
lated the existence of a short-life intermediate going from
one phase to the other. In the same work, differences in
the nature of the catalysts have been noticed, depending
on the silver content. Under reaction conditions, metallic
silver particles are thought to prevail on a 6% Ag/Al2O3
�
+
samples were dried for 14 h at 120 C and then calcined at
catalyst whereas Ag species are thought to prevail on
�
6
30 C for 6 h in synthetic air. N2 adsorption at 77 K using
a 2% Ag/Al2O3 sample. In the latter case, a higher dis-
persion and a greater interaction with the alumina would
explain the stabilisation of an oxidised state of the silver
a Micromeretics system was used to measure the surface
area of the samples. Prior to these measurements, the sam-
�
ples were each outgassed for 2 h at 200 C under a dynamic
(13). The parallelism to the results of Burch et al. for the
vacuum (i.e., with a residual pressure lower than 20 Pa).
Atomic absorption spectroscopy measurements were per-
formed to determine the silver content of the catalysts. The
silver loadings on the alumina are reported in wt%.
Pt/Al2O3 discussed above is tempting (4): one could pos-
tulate a decomposition-type mechanism over the metallic
+
silver particles whereas the Ag phase would favour a NO2
route involving both the silver and the alumina phases.
Over a 5% Ag/Al2O3, Sumiya et al. (14) reported that a
CxHyNOz compound was produced from NO, O2, and C3H6
which subsequently decomposed to an isocyanate species.
2
.2. Catalytic Tests
A quartz flow microreactor (3-mm internal diameter)
The isocyanate was found to react readily with NO, O2, was used for the catalytic tests, the catalytic bed being held
and especially NO + O2 to produce N2 and carbon oxides. in place by quartz wool plugs. The temperature of the reac-
The authors concluded that organo-NOx compounds and tionwasmeasuredinsidethereactor, justbeforethecatalyst
isocyanate were crucial intermediates of the SCR reaction bed, by a thermocouple enclosed in a quartz tube. Unless
and that the rate-determining step was the formation of the otherwise stated, the temperature of the reactor furnace
�
�
isocyanate. It has been observed that the use of high calci- was reduced from 600 to 200 C in 50 or 25 C intervals,
nation temperatures for some alumina-promoted catalysts dwelling at each temperature for 1 h. The data points re-
was beneficial to the SCR reaction (15, 16). This effect has ported were taken 5 and 20 min before the end of the
been assigned to the formation of aluminate phases which dwelling stage at each temperature. The reactant gases used
are less active for the total combustion of the reductant than were high-purity 1% NO/He (BOC), 1% C H /He (Air
3
6
the corresponding alumina-supported oxides (17, 18, 8).
2
Products), O (BOC, 99.9%), and Ar (BOC 99.99%). The
We have recently reported unexpectedly high activity for actual feed compositions used in each of the experiments
the oxidation of NO to NO2 over (nonpromoted) � -Al2O3 reported in this paper are shown in the legends of the ap-
during the C3H6-SCR of NO (19). Conversions to NO2 propriate figures. An analysis of the reaction products was
greater than that allowed by the thermodynamics of the carried out using a Nicolet 550 FT-IR spectrophotometer
1
3
reaction NO + O2 ⇔ NO2 could be obtained when com- fitted with a gas cell of volume 0.22 dm . The gas cell and
2
�
plete consumption of the propene was achieved. This ob- the lines of the system were heated at 90 C. The concen-
servation questions the validity of the accepted model of tration of a given species was measured by integrating the
the SCR of NO over some alumina materials in which the peaks in selected regions of its absorbance spectrum and
NO2 is believed to be formed by O2 oxidation of NO (5, 8). comparing these to a calibration curve. The integration in-
On the basis of these results, we have proposed alternative tervals were selected to avoid, wherever possible, overlap
SCR routes through a NO disproportionation reaction or a between the different species (Table 1). However, NO had
2
selective oxidation of the reductant which could explain the to be integrated in a region of the spectrum where C H6
3

PROPENE-SCR OF NO OVER ALUMINA-BASED CATALYSTS
495
TABLE 1
Integration Regions Used for the Quantification of Gaseous Species
Molecule
NO
N2O and CO
area
NO2 and C3H6
area
NH3
Peak measurement technique
Integration interval (cm 1)
area
height
931a
�
1878.7–1872
2220–2100
2960–2900
a
� 1
.
Peak extends over 940–924 cm
also absorbs and these products were therefore quantified temperature. Unless otherwise stated, the spectra reported
together using matrix-based calculations (QuantIR soft- here were taken after dwelling at a given temperature for
ware); the same applied to the combination of N2O and CO. 40 min. The absorbance measured in the presence of the re-
A relative precision better than � 2% was estimated for the action stream over the catalyst relative to that of the same
measured concentrations of the different species. As N2 is material at the same temperature under a stream of argon
notIRactiveandhencecouldnotbedetected, itsconcentra- is reported. Usually 64 or 128 scans were recorded at a
�
1
tion was calculated assuming that there was a mass balance resolution of 2 cm . The assignments of the absorption
of 100% for all the nitrogen-containing molecules. The rel- bands reported were based on those reported in the lit-
ative precision obtained on the value of N2 was estimated erature and other experiments (not shown) in which the
to be better than 6% and therefore lower than that of the reaction of propene and NO were investigated separately
other N-containing products which were directly quantifi- in the presence of a large excess of dioxygen.
able. Hydrogen cyanide, which was only observed at higher
temperatures over the alumina catalyst, was not quantified
3
. RESULTS
but this was thought to affect the N balance only to a mi-
nor extent (20). The yields were calculated on a nitrogen-
atom basis, i.e., N2 yield = 20%, meaning that 20% of the
NO molecules was converted to N2. On one hand, blank
3
.1. Activity of � -Al2O3 and Ag/� -Al2O3 Catalysts
for the C3H6-SCR of NO
experiment using an empty reactor showed no significant
conversion of NO up to 600 C at the lowest flow rate em- yields of N , NO , N O, and NH for the title reaction over
Figure 1 shows the propene and NO conversions and the
�
2
2
2
� 1
3
2
2
� 1
ployed (feed was 0.05% NO + 0.05% C3H6 + 2.5% O2/He; � -Al O (148 m g ), 1.2% Ag/� -Al O (141 m g ), and
2
3
2
3
�
1
2
� 1
total flow = 22.5 ml min ). On the other hand, the conver- 10% Ag/� -Al O (120 m g ). The alumina was active and
2
3
sion of propene was ca. 6% at this temperature.
selective for the formation of N at higher temperatures,
2
�
i.e., above 400 C, under these experimental conditions. It
2
.3. Thermogravimetric Analysis
is interesting to note that some N2O and especially high
concentrations of NO2 were observed over the alumina
Weight changes associated with passing NO and O2 over
�
after complete propene conversion, i.e., above 565 C. On
different materials were measured as a function of time
using an intelligent gravimetric analyser (IGA) supplied by
Hiden Analytical and capable of detecting changes in mass
of � 0.1 �g. Samples of 100 mg were first annealed in a He
the contrary, some NH3 could be observed but only be-
fore complete propene conversion. The 1.2% Ag/� -Al2O3
yielded similar conversions to N2 as those obtained over
the alumina but at lower temperatures. Low concentra-
3
� 1
�
�
stream of 100 cm min at 450 C for 3 h. The temperature
2 2 3
tions of N O, NO , and NH were also obtained in this case.
was then reduced to 400 C and the system was flushed with
3
� 1
The activity of the high-loaded silver catalyst was signif-
icantly different from that of the � -Al2O3 and the 1.2%
Ag/� -Al2O3. Complete combustion of the reductant was
a 400 cm min stream of 2.5% O2/He for 1 h. At this stage
00 ppm of NO was introduced into the stream and the
weight changes were recorded.
5
�
achieved at 350 C over the 10% Ag/� -Al2O3, in contrast
to the temperatures of 100% combustion of reductant over
2
.4. Diffuse Reflectance FT-IR Analysis
�
1
.2% Ag/� -Al2O3 and � -Al2O3 of 500 and 565 C, respec-
The diffuse reflectance FT-IR measurements were car-
tively. In addition, the N2 yield remained significantly lower
than that of N2O (obtained at the lower temperatures) and
NO2 (obtained at the higher temperatures). Over the 10%
Ag/� -Al2O3 and at the higher temperatures, the conversion
to NO2 was limited by the thermodynamics of the reaction
ried out in situ in a high-temperature cell (Spectra-Tech)
fitted with ZnSe windows. The sample for study (ca. 30 mg)
was finely ground and placed in a ceramic crucible, the tem-
perature of which could be varied from 20 to 800 C. All the
samples were calcined in situ at 630 C prior to analysis.
�
�
(
see dotted line in Fig. 1):
The temperature of the sample was increased from room
temperature in 100 C steps, dwelling at least 1 h at each
�
1
2
NO + O2 ⇔ NO2.
[1]

4
96
MEUNIER ET AL.
FIG. 2. Effect of the W/F on the C3H6-SCR of NO over � -Al2O3 at
�
5
90 C: propene conversion (᭿), NO conversion (4), N2 yield (᭺), NO2
yield (᭡), N2O yield (� ) and NH3 yield (❏). Feed: 0.05% NO + 0.05%
C3H6 + 2.5% O2/He.
�
NO over � -Al2O3 at temperatures above 565 C, this being
the temperature above which complete propene conver-
sion was obtained (Fig. 1). Higher than expected concen-
trations of NO2 have also recently been reported by our
laboratory for the same reaction and catalyst under con-
ditions where 100% propene conversion was achieved by
�
steadily increasing the W/F at a fixed temperature of 540 C
(
19). Figure 2 shows the results of a similar study car-
�
ried out at 590 C and over a broader range of W/F val-
ues than reported previously. The results show that, at
�
3
a W/F = 0.03 g s cm , propene conversion was 90% and
FIG. 1. C3H6-SCR of NO over � -Al2O3 (� ), 1.2% Ag/� -Al2O3 (᭺)
and 10% Ag/� -Al2O3 (᭹) catalysts as a function of temperature. Feed:
no NO2 could be observed while the N2 yield was 22%.
�
3
�
3
At W/F = 0.06 g s cm , propene conversion was completed
0
5
.05% NO + 0.05% C3H6 + 2.5% O2/He, W/F = 0.06 g s cm (GHSV �
�
1
and the NO yield reached a maximum at 32% while the N
2
2
0,000 h ). The dotted line in the plot giving the NO2 yield represents the
thermodynamic limit associated to the reaction NO + O2 ⇔ NO2.
1
2
yield was 40%. With further increases in the W/F value, the
N2 yield remained constant whereas the NO2 yield steadily
decreased, the latter being converted back to NO. Figure 3
shows the experimental molar ratio NO2/NO as a function
One of the striking features of the catalytic data reported
in Fig. 1 was the sharp increase in the NO2 yield obtained
overthealuminaassoonascompleteconversionofpropene
�
was achieved, at ca. 565 C. The values of NO2 yield ob-
tained at these temperatures were significantly higher than
that allowed by the thermodynamics of Eq. [1]. It has to be
stressed that identical plots were obtained independently
of using increasing or decreasing temperature profiles and
no change in the yield of NO2 was observed after several
hours on stream. A NO2 yield in slight excess of the ther-
modynamic limit was also observed over the 1.2% Ag/� -
Al2O3, although the difference between these two values
was within the experimental margin of error.
3
.2. Effect of the W/F on the C3H6-SCR of NO
over � -Al2O3 at 590 C
�
FIG. 3. Effect of the W/F on the C3H6-SCR of NO over � -Al2O3 at
Conversions to NO2 in excess of the thermodynamic
�
90 C: NO2/NO experimental ratio (᭺) and theoretical thermodynamic
5
limits of Eq. [1] were obtained during the C3H6-SCR of ratio (dotted line). Feed: 0.05% NO + 0.05% C3H6 + 2.5% O2/He.

PROPENE-SCR OF NO OVER ALUMINA-BASED CATALYSTS
497
�
of the W/F at 590 C, related to the data in Fig. 2. The dot-
ted line represents the thermodynamic NO2/NO ratio (i.e.,
0
.055) associated with Eq. [1] at the same temperature.
Upon complete conversion of propene, the experimental
ratio of NO2/NO, on one hand, increased from zero to well
beyond the thermodynamic limit, reaching a value of 1.7
�
3
at W/F = 0.06 g s cm . On the other hand, with further in-
creases in the W/F, the value of the ratio decreased toward
the thermodynamic figure associated with Eq. [1]. It has
to be stressed that identical plots were obtained indepen-
dently of using increasing or decreasing W/F values.
3
.3. Effect of the Partial Pressure of Propene on the
C3H6-SCR of NO over � -Al2O3 at 590 C
FIG. 5. Conversion to N2 during the C3H6-SCR of NO (open sym-
bol) and NO2 (filled symbol) over � -Al O (circle) and 1.2% Ag/� -
�
2
3
Al2O3 (triangle) as a function of temperature. Feed: 0.05% NOx + 0.05%
The effect of using increasing propene concentration in
� 3
C3H6 + 2.5% O2/He, W/F = 0.06 g s cm
.
�
the reaction feed was investigated at 590 C (Fig. 4). In
these experimental conditions, complete propene conver-
sionwas obtained ata feed concentrationbelow ca. 900 ppm
3
.4. Comparison of the Catalytic Activity for the
C3H6-SCR of NO and NO2
(
propene conversion was 97.5% at 1000 ppm). Without any
propene, NO was converted only to a small extent to NO2
i.e., yield of ca. 2%). With addition of only 100 ppm of
The � -Al2O3 and 1.2% Ag/� -Al2O3 appeared to be the
(
most interesting catalyst with regard to the reduction of NO
to N2 (Fig. 1). The beneficial effect of using NO2 instead of
NO as a starting NOx molecule has already been reported
for the C3H6-SCR over � -Al2O3-based catalysts (5, 11).
This is exemplified in Fig. 5 which shows the N2 yield ob-
tained over our � -Al2O3 and 1.2% Ag/� -Al2O3 using both
NO and NO2. On one hand, the most active combination of
catalyst and feed was � -Al2O3–NO2, which gave the high-
est N2 yields over the broadest temperature range. On the
other hand, the least active combination was � -Al2O3–NO.
Interestingly, the combinations using NO2 gave similar con-
versions to N2 at lower temperatures over both the � -Al2O3
and the 1.2% Ag/� -Al2O3. Similar to the case of Co/� -
Al2O3 materials (5), these results could suggest that the
only role of silver (at this low loading) was to act as a pro-
moter for the oxidation of NO to NO2, which would be the
rate-determining step of the C3H6-SCR reaction over the
alumina.
propene into the feed, more NO2 and also some reduction
products, i.e., N2 and N2O, were observed. With 300 ppm
of propene in the feed, the NO2 yield was at its maximum
value. It must be stressed that below 300 ppm, the yield of
NO2 was significantly higher than that of N2. With higher
propene concentrations, the NO2 yield decreased and N2
became the main product of the reaction. The NO2 con-
centrations obtained were higher than that expected from
the Eq. [1] when using initial concentrations of propene
higher than zero and not exceeding 1000 ppm. At the
higher propene concentrations investigated, the N2 yield
decreased, while significant proportions of N2O and NH3
were observed.
1
2
3
.5. Catalytic Activity for the Reaction: NO + O2 ⇔ NO2
To assess the role of the silver in the reaction scheme,
the activity of the � -Al2O3 and the 1.2% Ag/� -Al2O3 for
the oxidation of NO to NO2 was measured (Fig. 6). For
completeness, the data for the oxidation activity of the 10%
Ag/� -Al2O3 is also shown. This sample displayed high activ-
ity for the formation of NO2 which reached the limit of yield
�
imposed by the thermodynamics of Eq. [1] at 500 C. Inter-
estingly, the NO2 yield obtained over the 1.2% Ag/� -Al2O3
was similar to that obtained over the alumina. Moreover,
both catalysts exhibited NO2 yields significantly lower than
the N2 yields observed in the corresponding SCR reaction
over the whole temperature range investigated (see Fig. 1).
These results apparently contradict the model suggested
FIG. 4. Effect of the feed concentration of C3H6 on the C3H6-SCR
of NO over � -Al2O3 at 590 C: N2 yield (᭺), NO2 yield (᭡), N2O yield
�
(
� ) and NH3 yield (❏). Feed: 0–0.2% C3H6 + 0.05% NO + 2.5% O2/He,
� 3
W/F = 0.06 g s cm
.

4
98
MEUNIER ET AL.
TABLE 2
Bands Observed on Alumina and Ag/Alumina Material dur-
ing DRIFTS Experiments and the Corresponding Surface Species
and Vibrations To Which They Were Assigned (s = Symmetric,
a = Asymmetric � = Stretching, and � = Bending)
Wavenumber
Surface
species
�
1
(
cm
)
Vibration
References
25, 27
�
a
1
1
585
460
Carboxylate COO
�
OCO
s
�
OCO
�
a
3
2
1
1
005
910
595
390
Formate HCOO
�
+ �CH
25, 26, 27
OCO
�CH
a
�
OCO
�CH
� Os CO
FIG. 6. Conversion of NO to NO2 over � -Al2O3 (� ), 1.2% Ag/� -
1380
�
Al2O3 (᭺)and10%Ag/� -Al2O3 (᭹)catalystsasafunctionoftemperature.
1
1
550
250
Nitrate NO3
�N == O
21, 22
21, 22
Feed: 0.05% NO + 5% O2 in Ar, W/F = 0.06 g s cm�
3
.
a
�
ONO
�
1590
Nitrate NO3
�N == O
a
earlier, in which the oxidation of NO by O2 to NO2 was
proposed to be one of the reaction steps. This step should
be at least as fast as the rate-determining step of the overall
SCR reaction. These apparent low activities for the forward
reaction of Eq. [1] in the absence of any reductant could be
explained by the poisoning of the catalysts by NO2, as was
suggested in the case of Co/� -Al2O3 catalysts (10). Figure 7
shows the activity of the same materials for the partial de-
1305
�ONO
�
a
1235
Nitrite NO2
�
21
ONO
1560
1300
Ad-NOx
23, 24
�
bulk-like NO2 ?
1645
1
Organo-nitrite or oxime? �N == O or �N == O
14, 31
24
a
380
Organo-nitro?
Isocyanate –NCO
Cyanide –CN
�
ONO
2230
2135
29
30
1
2
composition of NO2 to NO and O2. The 10% Ag/� -Al2O3
catalyst achieved equilibrium yields at temperatures higher
�
than 450 C, whereas the alumina and 1.2% Ag/� -Al2O3 ex-
to the NO/O2 mixture, with the intensity of the absorption
band remaining unchanged after 3 h on stream (Fig. 8a). Af-
hibited much lower activities.
�
ter several hours of reaction, the formation of nitrate NO3
3
.6. In situ DRIFTS and Thermogravimetric Studies of the
Reaction: NO + O2 ⇒ ad-NOx
� 1
species (1550 cm , (21, 22)) was also observed. The in situ
1
2
DRIFTS analysis over the 1.2% Ag/� -Al O revealed that
2
3
The formation of adsorbed NOx (ad-NOx) species over nitrate species were formed only minutes after exposure
the Al2O3 and the Ag/� -Al2O3 materials was studied by to a NO/O2 stream and that the intensity of their absorp-
�
in situ DRIFTS and thermogravimetric analyses at 400 C. tion bands continued to grow gradually with time on stream
�
� 1
A nitrite NO species (1235 cm , Table 2) was observed (Fig. 8b). Two types of bidentate nitrates were formed, one
on the surface of the � -Al2O3 only minutes after exposure species with bands at 1550 and 1255 cm and another one
2
�
1
�
1
with bands at 1590 (shoulder) and 1305 cm . Figure 8c
shows the in situ DRIFTS spectra obtained over 10% Ag/� -
Al2O3. Several ad-NOx species were formed within minutes
of the introduction of the NO/O2 stream; the bands around
�
1
1
560 cm are probably similar to the bidentate nitrates ob-
served over the materials discussed above. The intensity of
�
1
the band at 1300 cm was higher than that expected for the
�
1
band associated with the bidentate nitrate at 1590 cm . As
�
1
a result, this band at 1300 cm can probably be attributed
to bulk-like nitrite species (23, 24). Figure 9 reports the rel-
ative weight uptake obtained for the same catalysts under
the same experimental conditions. The weight uptake was
significantly faster with the 1.2% Ag/� -Al2O3 than it was
with the alumina. This set of data shows that both the alu-
mina and the 1.2% Ag/� -Al2O3 were able to oxidise NO to
ad-NOx species and that the latter was significantly more
active for the adsorption than the former. The similarly low
FIG. 7. Conversion of NO2 to NO over � -Al2O3 (� ), 1.2% Ag/� -
Al2O3 (᭺)and10%Ag/� -Al2O3 (᭹)catalystsasafunctionoftemperature.
Feed: 0.05% NO2 + 5% O2 in Ar, W/F = 0.06 g s cm�
3
.

PROPENE-SCR OF NO OVER ALUMINA-BASED CATALYSTS
499
�
1
00 C. Thespectraobtainedafter40-mindwellingatagiven
temperature are shown. Unlike the case of the catalytic data
for which steady-state measurements were usually obtained
within minutes of a change in the experimental conditions,
it was observed in other experiments (not shown) that the
achievement of steady-state in situ DRIFTS spectra could
require significantly longer periods. This was due to slower
achievementofequilibriumsurfacecoverageof“spectator”
species associated with slower side reactions. Therefore,
the data reported here should be regarded as bearing in
them a character of temperature time-programmed analy-
sis. In addition, the W/F during the catalytic and DRIFTS
�
3
� 3
experiments were different, typically 60 � 10 g s cm in
�
3
� 3
the former case and approximately 18 � 10 g s cm in the
latter. Therefore, the DRIFTS experiments provided essen-
tially a qualitative description of the nature and reactivity
of the species observed at the catalyst surface at a given
temperature.
Figure 10 shows the in situ DRIFTS spectra obtained
during the title reaction over � -Al2O3. The spectra at all
temperatures were relatively simple and were essentially
composed of the bands associated with a formate species
�
1
(
i.e., bands at 3005, 2910, 1595, 1390, and 1380 cm ) and
�
1
a carboxylate species with bands at 1585 and 1460 cm
(
25–27) (Table 2). Whereas the carboxylate observed at the
highertemperaturesismostlikelyafreecarboxylatespecies
precursor of CO2), the carboxylate group observed at the
lower temperatures could also be that of an acetate species
(
�
� 1
(
28). At 300 C, a nitrite species (1235 cm (21)) was read-
ily formed upon admission of the reaction stream over the
sample but was gradually displaced with time on stream.
Figure 11 shows the in situ DRIFTS spectra obtained
during the same reaction over the 1.2% Ag/� -Al2O3. In
FIG. 8. In situ DRIFTS analysis of the formation of ad-NOx species
�
at 400 C over (a) � -Al2O3, (b) 1.2% Ag/� -Al2O3, and (c) 10% Ag/� -
Al2O3. For each catalyst, time on stream was 15 min (lower spectrum),
6
0 min (middle spectrum), and 180 min (upper spectrum). Feed: 0.05%
NO + 2.5% O2/He.
activity of these two samples for the oxidation of NO to NO2
reported in the previous section suggested that the ad-NOx
species formed on the 1.2% Ag/� -Al2O3 were stable and,
once formed on the surface of the catalyst, could not readily
be desorbed as NO2. Therefore, it appears likely that the
surface of the 1.2% Ag/� -Al2O3 catalyst can be poisoned
by strongly bound nitrates which form in the presence of
NO/O2. The low steady-state NO2 formation observed in
Fig. 6 could, hence, mostly be attributed to the alumina,
which was also probably poisoned to some extent by the
ad-NOx species.
3
.7. In situ DRIFTS of the C3H6-SCR of NO
FIG. 9. Thermogravimetric analysis of the formation of ad-NOx
species at 400 C over � -Al2O3 (� ), 1.2% Ag/� -Al2O3 (᭺), and 10% Ag/� -
For the experiments reported in this section the reaction
�
�
temperature was increased from 100 to 600 C in intervals of Al2O3 (᭹) as a function of time: Feed: 0.05% NO + 2.5% O2/He.

5
00
MEUNIER ET AL.
FIG. 12. In situ DRIFTS of 1.2% Ag/� -Al2O3 during the C3H6-SCR
�
�
of NO at 400 C. The sample was brought to 400 C in Ar and then exposed
to the stream of reaction; the spectra were recorded after 2 h. Feed: 0.05%
NO + 0.05% C3H6 + 2.5% O2/He.
FIG. 10. In situ DRIFTS of � -Al2O3 during the C3H6-SCR of NO.
Feed: 0.05% NO + 0.05% C3H6 + 2.5% O2/He.
this case, essentially no formate species were present at any
temperature but instead mainly nitrites (bands at 1235 and
which was probably masked at lower temperatures by the
nitrates bands. In an experiment made on an isothermal
�
1
1
1
320 cm ) and nitrates species (bands at 1255, 1300, and
�
dwelling at 400 C and over longer durations (Fig. 12), the
�
1
�
545 cm ) were observed (Table 2, (21)). At 300 C, ab-
� 1
band at 1645 cm was more clearly resolved and appeared
�
1
sorption bands assigned to isocyanate (2230 cm , (29)) and
cyanide (2135 cm , (30)) species were also observed. At
� 1
to be formed parallel to the band at 1380 cm . The absorp-
tion peak centred at 1645 cm could correspond to that of
�
1
� 1
�
4
00 C, the isocyanate band disappeared, whereas a band
organo-nitrite compound (14, 31) or oxime species (6) and
�
1
at 1460 cm appeared, indicating the presence of surface
carboxylate groups (the other associated band at 1585 cm
being masked by that of the nitrates species). A minor band
� 1
that at 1380 cm to organo-nitro compounds (24).
�
1
Figure 13 shows the in situ DRIFTS spectra obtained
�
over the 10% Ag/� -Al2O3. At 200 C, the surface, on one
�
1
at 1380 cm was also observed. At this temperature, some
gas-phase CO2 could be observed, which indicated that sig-
nificant propene conversion was attained. The coincidence
of the disappearance of the isocyanate band at the tempera-
ture at which the N2 yield started to be significant (see Figs. 1
and 5) suggested that the isocyanate could be an interme-
diate of the SCR reaction, as already proposed by other
authors for the same system (14). Eventually, the cyanide
species was also displaced but with longer dwelling times
hand, was covered by carboxylates species (ca. 1595 and
� 1
1
455 cm ) and other compounds which were not readily
determinable. On the other hand, at the higher tempera-
tures, only ad-NOx species were observed, similar to those
obtained under a NO/O2 stream (see Fig. 8c).
3
.8. In Situ DRIFTS of the C3H6-SCR of NO2
The C3H6-SCR of NO2 over � -Al2O3 was also inves-
�
�
�
� 1
tigated (Fig. 14). At 300 C, the in situ DRIFTS analysis
at 400 C. At 500 C, a band at 1645 cm was observed,
FIG. 11. In situ DRIFTS of 1.2% Ag/� -Al2O3 during the C3H6-SCR
of NO. Feed: 0.05% NO + 0.05% C3H6 + 2.5% O2/He.
FIG. 13. In situ DRIFTS of 10% Ag/� -Al2O3 during the C3H6-SCR
of NO. Feed: 0.05% NO + 0.05% C3H6 + 2.5% O2/He.

PROPENE-SCR OF NO OVER ALUMINA-BASED CATALYSTS
501
mation of NO2. On the basis of thermodynamic calculations
and reports from the literature, two alternative routes could
be proposed. One of them is based on a disproportionation
reaction (32):
4
NO ⇔ N2 + 2NO2.
[2]
The stoichiometry of Eq. [2] implies that the N2 and NO2
yields (based on nitrogen) should be equal. In fact, as NO2
could further readily react with the propene to give N2, the
expected NO2 yield should be lower than that of N2. How-
ever, NO2 yields significantly higher than that of N2 could
be obtained when using low concentrations of reductant
in the feed (Fig. 4), thus discarding Eq. [2] as the major
route for NO2 formation. A second possibility for NO2 for-
mation involves the selective oxidation of the reductant by
both NO and O2 to form a organo-nitro or -nitrite com-
pound (Eq. [3a]). Organo-nitro and -nitrite species have
often been quoted as possible intermediates of SCR reac-
tions (31, 33–35). The combustion of this molecule by O2
would lead to NO2 (Eq. [3b]):
FIG. 14. In situ DRIFTS of � -Al2O3 during the C3H6-SCR of NO2.
Feed: 0.05% NO2 + 0.05% C3H6 + 2.5% O2/He.
indicated the presence of both formate (i.e., bands at 3005,
0
�
1
R + NO + O2 →R –ONO + COx /H2O,
[3a]
[3b]
2
910, 1595, 1390, and 1380 cm ) and nitrate (1555, 1305,
�
1
0
00
and 1255 cm ) species at the surface of the catalyst. At
4
carboxylate group (1585 and 1465 cm ) and gas-phase
CO2. This observation should be related to the fact that the
rate of N2 formation was essentially nil at 300 C, whereas
R –ONO + O →R + NO + CO /H O.
2
2
x
2
�
00 C, only the formate species remained, along with the
�
1
In earlier work (19), nitro-methane was used as an exam-
ple for the organo-NOx compound. Thermodynamic data
for this molecule is available at high temperatures; as a
result, the free enthalpy of reactions with this molecule
can be calculated. Thus, whether the thermodynamics will
allow reactions [3a] and [3b] to proceed under the condi-
tions of interest and the possible relevance of this reaction
scheme for the formation of high concentrations of NO2
can be ascertained. However, if such a route applies to a
major extent during the C3H6-SCR of NO over � -Al2O3,
it is likely that the intermediate compound could corre-
spond to some other organo-nitro or -nitrite compounds
and possibly to a species which is only present in an ad-
sorbedstate. Obuchietal. reportedthepresenceofcarbona-
ceous radical species over the surface of alumina during the
C3H6-SCR of NO (36). The reactivity of such compounds
should be high toward O2 (paramagnetic) and especially
NO which is a free radical. Otsuka et al. provided further
support for the relevance of organo-nitrogen species to
�
�
the reduction proceeded to a significant extent at 400 C
�
(Fig. 5). Hence, the nitrates species observed at 300 C were
not reactive at this temperature. A similar spectrum to that
�
�
at 400 C was obtained at 500 C, with an increased inten-
sity of the bands associated with the carboxylate and a de-
creased intensity of the bands associated with the formates.
The fact that no ad-NOx species could be detected on the
�
surface of the alumina above 300 C was rather surprising
and suggests that the NO2 may be readily reacting with an
adsorbed species to yield N2.
4
. DISCUSSION
4
.1. Origin of NO2 during the C3H6-SCR of NO
over Alumina
High surface area � -Al2O3 is one of the most active single de-NOx chemistry (37). These authors found that nitro-
oxides for the formation of N2 during the SCR of NO with alkane compounds were formed during the gas-phase ox-
light hydrocarbons such as ethene, propene, and propane, idation of light alkanes in the presence of NO. Although
although it is only active at temperatures typically above alkyl-nitrite could not be directly observed, these authors
�
4
00 C. A widely accepted model for the reaction mecha- suggested that nitrite species were important intermediates
nism suggests that the initial step is the O2 oxidation of in the reaction scheme involved. Interestingly, Smith et al.
NO to NO2, which is followed by the reduction of NO2 (38) reported the homogeneous formation of the concen-
with the hydrocarbon. However, the data reported here tration of NO2 in excess of the limits determined by Eq. [1]
(
Figs. 1 and 3) and earlier published works show that yields from mixtures containing NO, O2, and hydrocarbon species
�
ofNO2 inexcessof thethermodynamic limit associatedwith at temperatures above 500 C. These authors invoked the
Eq. [1] could be obtained during the C3H6-SCR of NO. This occurrence of free radicalar reactions including HO2 and
observation rules out Eq. [1] as the main route for the for- alkyl radicals which would thermodynamically justify the

5
02
MEUNIER ET AL.
0
high proportion of NO2 observed. The space velocities used
R + NO → –CH –NO,
[3g]
[3h]
2
�
1
by these authors (up to 4,000 h ) were, however, signifi-
cantly lower than that used in the present work (typically
NO or NO2 + –C == N == O or –NH2 or NH3 → N2.
�
1
5
0,000 h ). Although the NO conversions measured in an
0
empty reactor were not significant in our case, the occur- The species R and R quoted in Eqs. [3f] and [3g] derive
rence of gas-phase reactions possibly triggered by the alu- from propene and could be oxygenated compounds, or pos-
mina surface can nevertheless not be excluded.
sibly radicals as suggested by Obuchi et al. (36). The in situ
Other results obtained over alumina showed that the DRIFTS analysis during the SCR reaction over the alumina
compound from which the NO2 originated was more likely only revealed formate and carboxylate species (Fig. 10),
to be an organo-nitrite species rather than an organo-nitro similar to those observed during the simple combustion
species (39). We have looked for such species by in situ of propene by O (not shown). The absence of any band
2
DRIFTS measurements over the alumina, both in steady assignable to organo-NO compounds suggests that the
x
state, and transient analyses at the high temperatures typ- concentration of these species were too low to be observed.
ical of C3H6-SCR reactions, but no such species could yet
Many of the potential products of reaction of organo-
be detected. An explanation for this could be that their nitrite and organo-nitro/nitroso compounds could further
concentration and lifetime is very small. This is supported react to yield N ; these possibilities are gathered in Eq. [3h].
2
by the findings of Yamaguchi (34) who reported that, for NH is a well-known reductant of NO and NO in O -rich
3
2
2
instance, nitro-methane was readily decomposed over alu- conditions over many catalytic materials (43). The interme-
�
mina at temperatures lower than 200 C. Another possibility diacy of NH in the hydrocarbon-SCR reaction has been
3
would be that these species are mostly short-lived gas-phase suggested over zeolitic catalysts (44, 45). Figures 1, 2, and
compounds.
4 show that NO was essentially observed only when NH3
2
was not present and vice versa. This fact strongly supports
4
.2. Mechanistic Considerations of the C3H6-SCR of NO
over Alumina
the idea that a significant proportion of the N formed
2
arose from the reaction between NO and NH , or at least
2
3
from the corresponding adsorbed species from which these
molecules were formed. In addition, a parallel route for N2
formation could also involve the reaction of NO with an ad-
sorbed isocyanate species (i.e., –NCO + NO → N2 + CO2),
similar to the one proposed to be involved in CO/NO re-
action over Rh catalysts (27). Over silver–alumina cata-
lysts, isocyanate species were even shown to react to form
N2 more readily with NO + O2 than with a O2-free NO
stream (14). More work is needed to clarify which of these
steps are possible over the alumina and ultimately if any of
those is actually occurring during the C3H6-SCR of NO. A
simplified scheme summarising the different reaction steps
proposed is given in Fig. 15.
For the rest of the discussion, we shall assume, on one
hand, the relevance of Eqs. [3a] and [3b] to the reaction
mechanism of the C3H6-SCR of NO over alumina with
regard to the formation of NO2. On the other hand, this sys-
tem needs to be completed by another set of equations to
take into account the formation of NH3. This molecule was
observed in significant proportions during the C3H6-SCR of
NO, but only when the conversion of propene was incom-
plete (Figs. 1, 2, and 4). NH3 can be obtained from reaction
of nitromethane over alumina (34, 41), probably through
the tautomerisation to the corresponding oxime followed
by dehydration to a nitrile N-oxide (Eq. [3c]) which iso-
merises to an isocyanate before yielding a primary amine
and NH3 by hydrolysis (Eq. [3d]), as suggested over zeolitic
materials (40, 41). Over alumina, the possibility of forming
NH3 from reaction of organo-nitrile N-oxides species was
confirmed by Obuchi et al. (42). The same authors pro-
posed that the organo-nitrile N-oxide were formed from
organo-nitroso compounds, via enol and cyanide forma-
tion (Eq. [3e]). Therefore, the formation of organo-nitro
(Eq. [3f]) and organo-nitroso (Eq. [3g]) compounds ap-
pears to be relevant to the mechanism of the C3H6-SCR
of NO over alumina:
–
–
–
CH2–NO2 → –CH == NO(OH) → –C == N == O,
C == N == O → –N == C == O → –NH2 → NH3,
CH2–NO → –CH == N(OH) → –C� N
[3c]
[3d]
FIG. 15. Simplified reaction scheme of the C3H6-SCR of NO over � -
Al2O3 giving the nature of the different species likely to be involved. See
→
–C == N == O,
[3e]
a
1
2
R + NO + O2 → –CH2–NO2,
[3f] Ref. (38).

PROPENE-SCR OF NO OVER ALUMINA-BASED CATALYSTS
503
4
.3. Utilisation of NO2 for the C3H6-SCR of NOx
over Alumina
Surprisingly, the in situ DRIFTS spectra of the C3H6-
�
SCR of NO and NO2 were similar at 400 and 500 C, tem-
peratures at which N2 was being formed in both cases (see
Figs. 5, 10, and 14). The absence of nitrate species on the
�
alumina surface at temperatures above 300 C when using
NO2 as a feed indicates that this molecule reacted from
the gas phase or was decomposed immediately upon ad-
sorption. Different roles for NO2 could be envisaged in the
reaction mechanism. NO2 could provide the oxygen nec-
essary for the activation of the reductant and thus be re- whereas Ag species favours the oxidation of NO to ad-NOx species
duced to NO; the latter molecule was indeed observed un- which subsequently react through the intermediacy of organonitrogen
FIG. 16. The different roles of Ag during the C3H6-SCR over Ag/� -
�
Al2O3: large Ag particles catalyse the decomposition-reduction of NO
+
compounds.
der these conditions. The activated reductant would then
react with NOx/O2 to form the organo-NOx compounds.
The formation of oxygenated molecules from propene can
�
to Ag2O at temperatures higher than 130 C. Interestingly,
be obtained over oxide catalysts pretreated with NO2 (46).
the 10% Ag/� -Al2O3 material was significantly more active
However, alumina displayed a poor ability for the selective
for the formation of NO2 during the SCR reaction (when
oxidation of propene, in contrast to titania, for instance. Al-
propene conversion was completed) than during the simple
though not directly related to the SCR reaction, this obser-
O2 oxidation of NO (see Figs. 1 and 3). A sample previ-
vation effectively rules out the role of NO2 as an activator
ously used for the SCR reaction or one prereduced in H2
of the propene to a specific CxHyOz intermediate. A more
�
at 400 C gave higher activities for the O2 oxidation of NO
straightforward possibility would be that the NO2 could re-
than a calcined sample (not shown). As no deactivation nor
act with the reductant or a derived species (CxHy or CxHyOz
any hysteresis were observed after use in strongly oxidative
formedwithO2)todirectlyformtheorgano-NOx speciesas-
atmospheres, it is thought that the reduction (in situ with
sociated with the reaction scheme described above. The for-
propene or ex situ with H2) irreversibly modified the disper-
mation of alkyl-nitrite compounds from light alkanes and
sion of the metal over the alumina. The 10% Ag/� -Al2O3
NO2 was suggested as being a crucial step in homogeneous
catalyst displayed a catalytic behaviour somewhat similar to
selective oxidation reactions, at temperatures as low as
that of Pt-based formulations (47). The related mechanism
of reaction involves the decomposition of NO into elemen-
tal nitrogen and oxygen species over a surface which is kept
reduced by the reductant (3) (Fig. 16). In spite of significant
activity for NO reduction at low temperatures, the high se-
lectivity to N2O observed on the 10% Ag/� -Al2O3 catalyst
makes this type of material of rather limited interest for the
C3H6-SCR of NO.
�
4
50 C (37). Therefore, the formation of an organo-NOx
compound from propene and NO2 is imaginable under our
experimentalconditions. Interestingly, noammoniawasob-
served during the C3H6-SCR of NO2. This suggests that
the formation of the organo-nitro/nitroso compounds from
which ammonia derives would be the rate-limiting step of
the reaction when using NO2 as the reactant in place of NO.
4
.4. Mechanistic Considerations of the C3H6-SCR of NO
4.5. Mechanistic Considerations on the C3H6-SCR of NO
over 10% Ag/� -Al2O3
over 1.2% Ag/� -Al2O3
The results reported here regarding the effect of the load-
At the lower loading of 1.2% Ag/� -Al2O3 the silver was
ingofsilveroverthealuminaareconsistentwiththefindings probably kept in an oxidised state, even at the higher tem-
of Bethke et al. (11). These authors suggested that the differ- peratures under reaction conditions, this being due to a
ence in catalytic activity for the C3H6-SCR of NO between strong interaction with the support (11). The possibility
a 6 and a 2 wt% Ag/� -Al2O3 was due to the different dis- of forming a silver aluminate was reported when using
�
persions and oxidation states of the silver on the alumina, high temperature of calcination (i.e., 800 C) in a steam-
2
the surface area of the latter being in the range 200–250 m
containing atmosphere (17). The activity of the 1.2% Ag/� -
g . The high activities for propene combustion and N2O Al2O3 was slightly modified by prereduction with H2 at
and NO2 formation (in the absence of propene) which were 400 C; the activity for the O2 oxidation of NO was increased
�
1
�
observed on our 10% Ag/� -Al2O3 can therefore be most by ca. 50% and the activity for the SCR activity was de-
likely attributed to the metallic character of large silver creased, probably due to an increase in the direct combus-
particles on the alumina. For nonsupported silver materials tion of propene. The modification upon H2 reduction could
and under our conditions of O2 partial pressure, the metal- be assigned to a decrease of the dispersion of the silver. Re-
lic state of silver is thermodynamically favoured as opposed gardless of the exact state of the silver over the (calcined)

5
1
04
MEUNIER ET AL.
.2% Ag/� -Al2O3 during the SCR reaction, the associated mina. These assumptions are supported by the findings of
catalytic properties were significantly different from that of Sumiya et al. who observed the formation of an IR band
�
� 1
the10%loadingsample. At300 C, therateforpropenecon- at 1655 cm assigned to an organo-NOx species over a
version of the former was ca. 200-fold lower than that of the 5% Ag/Al2O3 material under a NO/O2/C3H6 flow at room
latter. The 1.2% Ag/� -Al2O3 sample appeared to have cata- temperature (14). During a temperature-programmed re-
lytic properties for the C3H6-SCR of NO somewhat similar action, this compound was converted to isocyanate species,
�
1
� 1
to those of the � -Al2O3 but shifted toward lower temper- both on the silver (2230 cm ) and the alumina (2260 cm ).
atures (Fig. 1). However, contrary to the case of the alu- These authors reported that the isocyanate species were re-
mina, the NO2 yield over the 1.2% Ag/� -Al2O3 was never acting with NO, O2, and especially with NO/O2 to yield N2.
in significant excess of the limit set by Eq. [1] when com- However, no clear conclusions could be drawn as to which
plete propene conversion was achieved, either by varying of the Al–NCO or Ag–NCO was reacting faster. Following
�
the temperature (Fig. 1) or the W/F at 590 C (not shown). our results and those of the literature, a simplified model
Although both samples had identical activities for O2 oxi- of the mechanism of the C3H6-SCR of NO over the 1.2%
dation of NO to NO2 (Fig. 6), the silver-promoted material Ag/� -Al2O3 material can be obtained by adding the set of
displayed greater activity for the formation of surface ni- equations [4a]–[4d] to those already described for the alu-
trates (Figs. 8 and 9). This observation suggests that the role mina (Eqs. [3a]–[h]) (Fig. 16):
of the silver phase was to oxidise NO to inorganic ad-NOx
NO + O2 → ad-NOx ,
R + ad-NOx → R–ONO,
R + ad-NOx → R–NO,
R + ad-NOx → R–NO2.
[4a]
[4b]
[4c]
[4d]
species during the SCR reaction. This is supported by the
insituDRIFTSresultswhichshowedthepredominantpres-
ence of nitrate species on the surface of the silver-promoted
catalyst during the SCR reaction (Fig. 11), in contrast to the
alumina which was covered with formate species (Fig. 10).
It is likely that these nitrates would be too strongly bound
to the surface to desorb and would, eventually, if no re-
ductant was available, poison the catalyst. Hence, the low
activities measured for Eq. [1] were very similar to those of
the alumina. A similar explanation for the low activity of
Co/� -Al2O3 for the oxidation of NO to NO2 was proposed
by Yan et al. (10).
Hence, the promoting effect of the silver that we would
like to propose over the 1.2% Ag/� -Al2O3 is characterised
by increased rates of formation of organo-NOx compounds
Eqs. [4b]–[4d]) as compared to the case of the � -Al2O3
Eqs. [3a], [3f], and [3g]). However, it is not clear yet if
the decomposition/reactions of these organo-NOx species
occur on both the silver and the alumina phases or predom-
inantly on one of those. The precise nature of the reacting
species derived from propene still needs to be determined.
(
(
The C3H6-SCR of NO2 gave similar N2 yields over the
�
-Al2O3 and the 1.2% Ag/� -Al2O3 over a broad range of
temperatures (Fig. 5). This suggests that the reduction abil-
ity could essentially be attributed solely to the alumina.
5
. CONCLUSIONS
(The lower N2 yield observed over the silver-promoted ma-
terial at the higher temperatures was probably due to a
The formation of NO2 during the C3H6-SCR of NO over
higher rate of propene combustion over this sample com- � -Al2O3 is not achieved through the oxidation of NO with
pared to that of the alumina, the combustion reaction O2. A mechanism (possibly partly homogeneous) involving
competing with the SCR reaction for the reducing agent.) the formation of organo-nitrite species followed by their
Similar to the reaction scheme proposed for the alumina, decomposition/oxidation is suggested to be the route ac-
the nitrates formed over the silver phase could be reacting counting for the formation of NO2. The promoting role of
with the reductant or a derived species to form organo-NOx low loadings of silver (i.e., 1.2 wt%, probably in an oxi-
species which would then react to yield N2. The isocyanate dised state) on the activity for N2 production is probably
and cyanide species observed over the silver-promoted ma- due to the higher rate of formation of inorganic ad-NOx
terial could be formed from the species with the IR band species of the silver phase. It is proposed that these inor-
�
1
at 1645 and 1380 cm (Figs. 11 and 12), probably organo- ganic ad-NOx species further react with the reductant or a
nitro- or -oxime species. Interestingly, the isocyanate band, derived species to form various organo-NOx compounds.
on one hand, disappeared readily when the temperature In particular, organo-nitro and organo-nitroso compounds
�
was increased from 300 to 400 C, coinciding with the light- and/or their derivatives (e.g., isocyanate, cyanide, amines,
off of the SCR reaction. On the other hand, the cyanide and NH3) are suggested to react with NO or the organo-
band was more stable. This suggests that the isocyanate nitrite and/or its derivative NO2 to yield N2.
species would be the most reactive reaction intermediate.
When no reductant is present, the 1.2% Ag/� -Al2O3 ma-
The promoting effect of silver on the formation of the terial is poisoned by strongly bound nitrates and its ac-
organo-NOx species would increase their surface concen- tivity for NO2 formation is similar to that observed over
tration to a noticeable extent when compared to the alu- the alumina. A material with a high loading of silver (i.e.,

PROPENE-SCR OF NO OVER ALUMINA-BASED CATALYSTS
505
1
0 wt%) on the alumina displays significantly different cata- 19. Meunier, F. C., Breen, J. P., and Ross, J. R. H., Chem. Comm. 3, 259
(
1999).
0. Radtke, F., Koeppel, R. A., Minardi, E. G., and Baiker, A., J. Catal.
67, 127 (1997).
1. Schraml-Marth, M., Wokaun, A., and Baiker, A., J. Catal. 138, 306
1992).
22. Kijlstra, W. S., Brands, D. S., Poels, E. K., and Bliek, A., J. Catal. 171,
08 (1997).
lytic properties as compared to that of a low-loading ma-
terial; those are probably associated with a metallic state
of the silver in the former case and result in the forma-
tion of mostly N2O at low temperatures during the C3H6-
SCR of NO.
2
2
1
(
2
2
2
2
2
2
3. Cross, A. D., “Introduction to Practical IR Spectroscopy.” Butter-
sworth, London, 1964.
ACKNOWLEDGMENTS
4. Colthup, N. B., Daly, L. H., and Wiberley, S. E., “Introduction to
Infrared and Raman Spectroscopy.” Academic Press, Boston, 1990.
5. Busca, G., Lamotte, J., Lavalley, J. C., and Lorenzelli, V., J. Am. Chem.
Soc. 109, 5197 (1987).
6. Nakamoto, K., “Infrared and Raman Spectra of Inorganic and Coor-
dinatio Compounds,” 4th ed. Wiley-Interscience, New York, 1986.
7. Mathyshak, V. A., and Krylov, O. V., Catal. Today 25, 1 (1995).
Part of this project was funded by the European Community, through
the Environment and Climate Programme, subprogramme “Technologies
for Rehabiliting the Environment.” Contracts E5V5-CT94 and ENV4-
CT97-0658 and by Forbairt, Contract SC/1997/519. The referees are grate-
fully acknowledged for their comments on gas-phase reactions.
REFERENCES
28. Shimizu, K., Kawabata, H., Satsuma., A., and Hattori, T., Appl. Catal.
B 19, L87 (1998).
1
2
3
4
. Farrauto, R. J., Ed. Appl. Catal. B 10, 1 (1996).
29. Ukisu, Y., Miyadera, T., Abe, A., and Yoshida, K., Catal. Lett. 39, 265
(1996).
30. Bamweda, G. R., Obuchi, A., Ogata, A., and Mizuno, K., Chem. Lett.
2109 (1994).
. Fritz, A., and Pitchon, V., Appl. Catal. B 13, 1 (1997).
. Burch, R., and Watling, T. C., Catal. Lett. 43, 19 (1997).
. Burch, R., Fornasiero, P., and Watling, T. C., J. Catal. 176, 204
(
1998).
31. Tanaka, T., Okhuhara, T., and Misono, M., Appl. Catal. B 4, L1 (1994).
5
. Hamada, H., Kintaichi, Y., Inaba, M., Tabata, M., Yoshinari, T., and 32. Smits, R., and Iwasawa, Y., Appl. Catal. B 6, L201 (1995).
Tsuchida, H., Catal. Today 29, 53 (1996). 33. Yokoyama, C., and Misono, M., J. Catal. 150, 9 (1994).
. Radtke, F., Koeppel, R. A., Minardi, E. G., and Baiker, A., J. Catal. 34. Yamaguchi, M., J. Chem. Soc., Faraday Trans. 93, 3581 (1997).
6
1
67, 127 (1997).
35. Cowan, A. D., Cant, N. W., Haynes, B. S., and Nelson, P. F., J. Catal.
176, 329 (1998).
7
8
. Miyadera, T., Appl. Catal. B 2, 199 (1993).
. Okazaki, N., Tsuda, S., Shiina, Y., and Tada, A., Chem. Lett. 51 36. Obuchi, A., Ogata, A., Mizuno, K., Ohi, A., Nakamura, M., and
1998). Ohuchi, H., J. Chem. Soc., Chem. Commun. 247 (1992).
. Yan, J., Kung, M. C., Sachtler, W. M. H., and Kung, H. H., J. Catal. 172, 37. Otsuka, K., Takahashi, R., Amakawa, K., and Yamanaka, I., Catal.
78 (1997). Today 45, 23 (1998).
0. Yan, J. Y., Kung, H. H., Sachtler, W. M. H., and Kung, M. C., J. Catal. 38. Smith, J., Phillips, J., Graham, A., Steele, R., Redondo, A., and Coons,
(
9
1
1
1
75, 294 (1998).
J., J. Phys. Chem. A 101, 9157 (1997).
1
1
1. Bethke, K. A., and Kung, H. H., J. Catal. 172, 93 (1997).
2. Haneda, M., Kintaichi, Y., Iniba, M., and Hamada, H., Appl. Surf. Sci.
39. Zuzaniuk, V., Meunier, F. C., and Ross, J. R. H., J. Chem. Soc., Chem.
Comm. 815 (1999).
1
21, 391 (1997).
40. Cant, N. W., Cowan, A. D., Doughty, A., and Nelson, P. S.,
Catal. Lett. 46, 212 (1997).
41. Lombardo, E. A., Sill, G. A., d’Itri, J. L., and Hall, W. K., J. Catal. 173,
440 (1998).
42. Obuchi, A., Wogerbauer, C., Koppel, R., and Baiker, A., Appl. Catal.
B 19, 9 (1998).
43. Bosch, H., and Janssen, F., Catal. Today 2, 369 (1988).
1
1
1
1
3. Kung, M. C., Bethke, K. A., Yan, J., Lee, J. H., and Kung, H. H., Appl.
Surf. Sci. 121, 261 (1997).
4. Sumiya, S., He, H., Abe, A., Takezawa, N., and Yoshida, K., J. Chem.
Soc., Faraday Trans. 94, 2217 (1998).
5. Okazaki, N., Tsuda, S., Shiina, Y., Tada, A., and Iwamoto, M., Chem.
Lett. 429 (1998).
6. Hamada, H., Haneda, M., Kakuta, N., Miura, H., Inami, K., Nanba, 44. Centi, G., Galli, A., and Perathoner, S., J. Chem. Soc., Faraday Trans.
T., Hua, W. Q., Ueno, A., Ohfune, H., and Udagawa, Y., Chem. Lett.
87 (1998).
7. Nakatsuji, T., Yasukawa, R., Tabata, K., Ueda, K., and Niwa, M., Appl.
Catal. B 17, 333 (1998).
92, 5129 (1996).
8
45. Poignant, F., Saussey, J., Lavalley, J. C., and Mabilon, G., Catal. Today
29, 93 (1996).
46. Ueda, A., Ejima, K., Azuma, M., and Kobayashi, T., Catal. Lett. 53, 73
(1998).
1
1
8. Shimizu, K., Maeshima, H., Satsuma, A., andHattori, T., Appl. Catal. B
1
8, 163 (1998).
47. Burch, R., and Watling, T. C., Appl. Catal. B 11, 207 (1997).