NO reduction by propylene over Pt/SiO2: An in situ FTIR study

Article abstract of DOI:10.1006/jcat.2000.2950

The SCR of NO by propylene (C₃H₆) over a 1 wt % Pt/SiO₂ catalyst was studied using in situ FTIR spectroscopy at 250°C. Introduction of NO at 250°to an oxidized Pt/SiO₂ sample did not result in the formation of any significant amounts of adsorbed NO_x-type species, associated with either the Pt catalyst or the silica support. On the contrary, introduction of C₃H₆ to the same catalyst resulted in the formation of surface carboxylates. Experiments conducted under reaction conditions resulted in the formation of a surface cyanide (-CN) species, and a gas phase/weakly adsorbed N₂O species. The surface cyanide was the product of the reaction between an oxygen-activated hydrocarbon derivative and NO, and was reactive toward NO, NO₂, and O₂, resulting in the formation of N₂ and N₂O. Its behavior was consistent with a role as a potential intermediate in the SCR of NO by C₃H₆ over the Pt/SiO₂ catalyst.

Full text of DOI:10.1006/jcat.2000.2950

Journal of Catalysis 194, 222–232 (2000)
doi:10.1006/jcat.2000.2950, available online at http://www.idealibrary.com on
NO Reduction by Propylene over Pt/SiO : An in Situ FTIR Study
2
1
Dinyar K. Captain and Michael D. Amiridis
Department of Chemical Engineering, University of South Carolina, Columbia, South Carolina 29208
Received December 9, 1999; revised May 31, 2000; accepted June 5, 2000
face chemistry involved in this reaction is still not clearly
The selective catalytic reduction (SCR) of nitric oxide (NO) by understood and is highly debated. This study aims toward
propylene (C3H6) over a 1 wt.% Pt/SiO2 catalyst was studied by the further development ofsuch an understandingand com-
the use of in situ Fourier transform infrared (FTIR) spectroscopy
at 250 C. At this temperature, introduction of NO to the catalyst
FTIR studies were conducted with a Pt/SiO2 catalyst at
plements our earlier work in this area (4, 5, 19). In situ
�
did not result in the formation of any significant amounts of ad-
sorbed NOx-type species on either the catalyst or the SiO2 support.
On the contrary, introduction of C3H6 to the same catalyst resulted
in the formation of surface carboxylates. Experiments conducted
under reaction conditions resulted in the formation of a surface
cyanide (–CN) species, and a gas phase/weakly adsorbed nitrous
�
2
50 C. Our previous kinetic studies indicate that at this
temperature the catalyst is active but has not yet reached
its maximum activity (5). Spectra were collected at atmo-
spheric pressure over both the Pt/SiO2 catalyst and the
SiO2 support under a variety of gas compositions. Stud-
15
ies with isotropically labeled nitric oxide ( NO) were also
oxide (N2O) species. The surface cyanide is the product of the re-
action between an oxygen-activated hydrocarbon derivative and conducted to further qualify the identification of the surface
NO and is reactive toward NO, NO2, and O2, resulting in the for- species detected.
mation of N2 and N2O. Its behavior, as observed in this study, is
consistent with a role as a potential intermediate in the selective
EXPERIMENTAL
reduction of NO by propylene over the Pt/SiO2 catalyst.
�c 2000
Academic Press
Catalyst Preparation
The Pt/SiO2 catalyst was prepared by incipient wetness
impregnation of the SiO2 support (Davison Syloid 74) with
an aqueous solution of H2PtCl6 � 6H2O (Aldrich Chemi-
INTRODUCTION
NOx emissions are regulated by a variety of stringent cal Co.). Prior to impregnation the support was calcined
�
laws at both the federal and local levels. Current com- overnight at 500 C in air. Following impregnation, the sam-
�
mercial catalytic technologies for the control of NOx emis- ple was dried at 120 C for 2 h and then calcined in air for 5 h
�
2
sions include the nonselective reduction by carbon monox- at 500 C. The calcined catalyst had a BET surface area of
ide and hydrocarbons over noble metal-based “three-way” 280 m /g and a Pt loading of approximately 0.9 wt% (ICP-
catalysts, and the selective reduction by ammonia over Galbraith Laboratories). Hydrogen chemisorption mea-
vanadia/titania-based catalysts. Since its introduction in surements indicate a Pt dispersion of approximately 60% .
1
991 (1, 2), the selective reduction of NO by hydrocarbons
Infrared Spectroscopy
has attracted a lot of interest, because it has certain advan-
tages over both current technologies (3).
Infrared spectra were collected with a Nicolet 740 FT-IR
spectrometer equipped with an MCT-B detector. Transmis-
sion spectra were collected in the single beam mode with
Supported platinum catalysts have shown significant ac-
tivity toward the selective catalytic reduction of NO by hy-
drocarbons at low temperatures (3–10), and they are not
significantly affected by the presence of water in the exhaust
stream (1, 3, 7, 11, 12). Their drawbacks include a signifi-
cant selectivity toward nitrous oxide (N2O) and a narrow
temperature window of operation (3–10, 13).
�
1
a resolution of 2 cm . Reference spectra of the clean sur-
faces in flowing He were collected separately and differ-
ence spectra between the samples and the corresponding
references are shown herein. A stainless steel IR cell with
NaCl windows cooled by flowing water and a 10-cm path
length was used. A heating element wrapped around the
cell allowed collection of in situ spectra at elevated tem-
peratures. The temperature was monitored through a ther-
mocouple located in the cell in close proximity with the
catalyst sample.
A number of studies have focused on the performance of
Pt-based catalysts (2–18). However, the fundamental sur-
1
To whom correspondence should be addressed. E-mail: amiridis@
engr.sc.edu.
222
0021-9517/00 $35.00
Copyright �c 2000 by Academic Press
All rights of reproduction in any form reserved.

NO REDUCTION BY PROPYLENE OVER Pt/SiO2
223
Catalyst samples were prepared as self-supported wafers, gas chromatograph (Hewlett–Packard 5890) equipped with
1
2 mm in diameter and with a “thickness” of about 20 mg/ a thermal conductivity detector and a three-column sys-
2
cm . Prior to each experiment the samples were pretreated tem (Molecular Sieve 5A, Porapak Q, and Graphpac GC)
at 250 C in a flowing 2% O2 in He mixture for 2 h. Certified for analysis of N2, N2O, CO, CO2, and C3H6. Calibration
�
analyzed mixtures of 1.0% NO in He (Matheson), 1.0% curves were obtained for each of these species and the re-
NO2 in He (Matheson), 1.0% C3H6 in He (National Spe- producibility of the measurements in each case was within
cialty Gases), 10.0% O2 in He (National Specialty Gases), � 5% . A thermocouple placed in the catalyst bed was used
and a 99.9999% He carrier gas (National Specialty Gases; to monitor the temperature of the catalyst. Each run uti-
further purified by an oxygen trap) were used to prepare lized approximately 500 mg of the Pt/SiO2 catalyst in the
the different gas mixtures. Typical concentrations used were form of 60/85 mesh particles.
2
000 ppm of NO/NO2 and C3H6, and 2.0% of O2. Gases
were mixed at the appropriate amounts by the use of a sys-
RESULTS AND DISCUSSION
tem of needle valves and flow meters and preheated prior
to the introduction to the cell. The volumetric flow rate of NO Adsorption
3
the gas mixtures was held constant at 100 cm /min (1 atm,
Pt/SiO2. The spectrum of the Pt/SiO2 catalyst follow-
ing exposure to a mixture of 2000 ppm NO in He at 250 C
�
2
5 C).
�
for 45 min is shown in Fig. 1a. The gas phase NO doublet
Reactor Studies
�
1
(
centered at 1875 cm ) is the only feature in this spec-
Transient reactor studies were conducted in a quartz, trum, indicating the absence of any significant amounts of
single-pass, fixed-bed, flow reactor. The volumetric flow adsorbed species either on Pt or the silica support. This is
3
rate of the reacting mixture was held at 100 cm /min (1 atm, consistent with previous literature reports indicating that
�
2
5 C). Inlet and outlet gas mixtures were analyzed using a adsorption of N O -type species on silica is only observed
x
y
�
FIG. 1. In situ infrared spectra at 250 C of an initially oxidized Pt/SiO2 catalyst: (a) 2000 ppm NO in He for 45 min; (b) 2000 ppm NO, 2% O2 in
He for 20 min; (c) flushing of (b) with He for 2 min.

2
24
CAPTAIN AND AMIRIDIS
at lower temperatures (20, 21). When 2% O2 was added to centered at approximately 1530 and 1430 cm 1. Similar
�
�
1
the flowingNO mixture, a newdoublet wasobserved at 1630 peaks at approximately 1550 and 1440 cm have been pre-
�
1
and 1595 cm (Fig. 1b). These two peaks correspond to the viously observed during propylene oxidation over various
asymmetric stretching vibrations of NO2 (22–24). However, metal oxides and have been assigned to surface carboxy-
when the sample was flushed with He (Fig. 1c), these two late species, and more specifically, the asymmetric and sym-
�
peaks disappeared immediately from the spectrum, indi- metric –COO stretching vibrations (19, 25–30). These re-
cating that the corresponding NO2 species was either very sults indicate that during exposure of the oxidized Pt/SiO2
�
weakly adsorbed or in the gas phase.
catalyst to propylene at 250 C partial oxidation products
are formed. The necessary oxygen for these reactions is
present on the catalyst surface as a result of the catalyst
oxidation pretreatment. Both peaks did not increase in
intensity during our experiments after the first 6 min on
stream, indicating that the surface oxygen needed for their
formation had been depleted at that point. The intensi-
ties of both peaks corresponding to the surface carboxy-
late species decreased with increased time of flushing in He
C3H6 Adsorption
SiO2. The only peaks present in the spectrum of the
silica support exposed to a mixture of 2000 ppm C3H6 in
�
He at 250 C are those corresponding to gas phase propy-
lene. Flushing the cell with He resulted in an immediate
disappearance of these peaks, confirming that the observed
propylene is either in the gas phase, or at most, very weakly
adsorbed on the silica support.
(
Fig. 2c), indicating that either some type of surface rear-
rangement resulting in decreasing absorption cross section
Pt/SiO2. The spectrum of the Pt/SiO2 catalyst collected takes place, or that these species desorb from the surface
�
at 250 C following exposure to propylene (Fig. 2a), is sig- under these conditions. In all probability conversion to the
nificantly different from that of the SiO2 support. This time, acid form takes place before desorption occurs. The cell
in addition to the gas phase peaks, a very broad peak cen- was also flushed with 2% O2/He and 2000 ppm NO/He
�
1
tered at approximately 1550 cm is observed. Flushing the mixtures in order to determine if the surface carboxylates
cell with He for 2 min (Fig. 2b), removes the gas phase are reactive toward these gases. The observed decrease in
propylene and resolves the broad peak into two peaks intensity in these cases was similar to when the cell was
�
FIG. 2. In situ infrared spectra at 250 C of an initially oxidized Pt/SiO2 catalyst: (a) 2000 ppm C3H6 in He for 30 min; (b) flushing of (a) with He
for 2 min; (c) flushing of (a) with He for 20 min; (d) 2000 ppm C3H6, 2% O2 in He for 30 min; (e) flushing of (d) with He for 2 min.

NO REDUCTION BY PROPYLENE OVER Pt/SiO2
225
flushed with only He, indicating that the surface carboxy- at 2100 cm 1, with a calculated 1 N shift of 28 cm , to
�
15
� 1
lates are not reactive in O2/NO environments under these an adsorbed cyanide (–CN) species. Additionally, previous
conditions.
studies on HCN adsorption over a polycrystalline Pt foil
�
1
When 2% O2 was added to the propylene stream also assigned a similar peak at 2086 cm to an adsorbed
Fig. 2d), an increase in intensity of the IR peaks, and hence, cyanide species (32). Consequently, we can also assign the
(
�
1
the surface concentration of the carboxylate species was ob- observed peak at 2090 cm to an adsorbed cyanide species.
served. Removal of the gas phase and flushing of the cell Whether this species is on Pt or on the SiO2 support is un-
with He (Fig. 2e) reveals that the carboxylate peaks are clear, given the proximity of the cyanide assignments in the
broader and more intense in this case, with the peak at aforementioned previous studies. Upon removal of the re-
�
1
� 1
1
530 cm being shifted to approximately 1570 cm .
acting gas mixture and flushing of the cell with He (Fig. 3c)
�
1
the peak at 2090 cm quickly disappears from the spec-
trum, indicating that the –CN species is weakly adsorbed
Reaction Studies
The FTIR spectrum collected during the exposure of the
on the surface.
The remaining two peaks in Fig. 3a, at 2236 and
�
1
Pt/SiO2 catalyst to a mixture of 2000 ppm NO, 2000 ppm 2204 cm , can be assigned to nitrous oxide (23, 24, 33).
�
� 1
C3H6, and 2% O2 in He at 250 C is shown in Fig. 3a. In addi- Their isotopic shifts to 2140 and 2170 cm , respectively,
tion to the peaks corresponding to gas phase NO and propy- are consistent with the proposed assignment, which was
lene and the adsorbed carboxylates, this spectrum also con- further confirmed by the introduction of gas phase N2O to
tains three new peaks at 2236, 2204, and 2090 cm . When the Pt/SiO2 catalyst. Kinetic results have shown that this
isotopically labeled nitric oxide ( NO) was used (Fig. 3b), catalyst is active under the conditions of the FTIR experi-
the peak at 2090 cm shifts by 30 cm to 2060 cm indicat- ments and that N2O is produced in significant amounts (5).
ing the presence of nitrogen in the corresponding species. Once again upon removal of the reacting gas mixture and
Previous literature reports focusing on the adsorption of flushing of the cell with He for 4 min (Fig. 3c), the two
HCN on silica (31), have assigned a similar FTIR peak peaks corresponding to nitrous oxide disappear from the
�
1
15
�
1
� 1
� 1
�
FIG. 3. In situ infrared spectra of an initially oxidized Pt/SiO2 catalyst: (a) 2000 ppm NO, 2000 ppm C3H6, 2% O2 in He for 1 h (250 C); (b) 2000 ppm
1
5
�
�
NO, 2000 ppm C3H6, 2% O2 in He for 1 h (250 C); (c) flushing of (a) with He for 4 min (250 C); (d) 2000 ppm NO, 2000 ppm C3H6, 2% O2 in He for
�
4
0 min (350 C).

2
26
CAPTAIN AND AMIRIDIS
�
FIG. 4.1. In situ infrared spectra at 250 C of an initially oxidized Pt/SiO2 catalyst: (a) 2000 ppm NO, 2000 ppm C3H6, 2% O2 in He for
4
min; (b) flushing of (a) with He for 2 min; (c) flushing of (a) with He for 4 min; (d) flushing of (a) with He for 6 min. (4.2) In situ infrared spectra at
�
2
50 C of an initially oxidized Pt/SiO2 catalyst: (a) 2000 ppm NO, 2000 ppm C3H6, 2% O2 in He for 4 min; (b) flushing of (a) with 2% O2 in He for 2 min;
(
c) flushing of (a) with 2% O2 in He for 4 min; (d) flushing of (a) with 2% O2 in He for 6 min.
spectrum indicating that the N2O species responsible for species. Spectra of the Pt/SiO2 catalyst exposed to a mix-
�
them is either weakly adsorbed or in the gas phase.
ture of NO/C3H6/O2/He at 250 C for 4 min are shown as
The spectrum of the Pt/SiO2 catalyst under reaction con- Figs. 4.1a through 4.5a. The catalyst was first flushed with
�
ditions at 350 C is shown in Fig. 3d. Our kinetic experiments He for a period of 6 min (Fig. 4.1). In this case, the inten-
(
�
1
5) have shown that this temperature is above the temper- sity of the peak at 2090 cm decreased significantly after
ature of maximum NO reduction, and propylene oxidation 2 min, and the peak completely disappeared after 4 min,
is reaching 100% under these conditions. In comparison to consistent with the results shown in Fig. 3c. This indicates
�
the spectrum collected at 250 C, several differences can be that the cyanide species is weakly adsorbed on the catalyst
�
1
observed. The carboxylate peaks in the 1300–1700 cm re- surface. During the same experiment, the N2O doublet at
�
1
gion are no longer present, consistent with the complete 2236 and 2204 cm was also significantly lower in intensity
oxidation of propylene. In addition, significant amounts of after 2 min of He flushing and completely disappeared af-
gas phase CO2 and H2O (the products of C3H6 oxidation) ter 4 min, indicating that the observed N2O species is either
are observed, as evidenced by the strong peaks in the 2360– weakly adsorbed or in the gas phase.
�
1
� 1
2
320 cm and 1800–1400 cm regions, respectively. Fi-
Treating the catalyst sample with a 2% O2 in He mix-
nally, while the peak corresponding to adsorbed cyanide at ture (Fig. 4.2) resulted in a complete disappearance of the
�
�
1
2
2
090 cm is no longer present, the N2O peaks at 2236 and cyanide peak within the first two minutes of treatment, in-
1
204 cm are still visible.
dicating the reaction of cyanide with oxygen. Surprisingly,
treating the catalyst sample with a 2000 ppm NO in He
mixture (Fig. 4.3) resulted in an apparent delay in the dis-
appearance of the cyanide peak. In this case, the intensity of
Reactivity and Origin of the Cyanide Species
To further investigate the reactivity of the surface cyanide the cyanide peak was unchanged after 2 min of treatment,
species observed, experiments were conducted with the was reduced to half its original value after 4 min, and even-
SCR reactants being flowed over surfaces containing this tually disappeared only after 6 min. The N O doublet was
2

NO REDUCTION BY PROPYLENE OVER Pt/SiO2
227
�
FIG. 4.3. In situ infrared spectra at 250 C of an initially oxidized Pt/SiO2 catalyst: (a) 2000 ppm NO, 2000 ppm C3H6, 2% O2 in He for 4 min;
(
b) flushing of (a) with 2000 ppm NO in He for 2 min; (c) flushing of (a) with 2000 ppm NO in He for 4 min; (d) flushing of (a) with 2000 ppm NO in
�
He for 6 min. (4.4) In situ infrared spectra at 250 C of an initially oxidized Pt/SiO2 catalyst: (a) 2000 ppm NO, 2000 ppm C3H6, 2% O2 in He for 4 min;
(
b) flushing of (a) with 2000 ppm NO2 in He for 2 min; (c) flushing of (a) with 2000 ppm NO2 in He for 4 min; (d) flushing of (a) with 2000 ppm NO2 in
He for 6 min.
also present throughout the first six minutes, and in fact, ever, demonstrate that during the treatment with NO two
did not completely disappear even after 6 min. In contrast, parallel reactions that result in both the formation and con-
treatment of the catalyst sample with a 2000 ppm NO2 in He sumption of the surface cyanide species take place. It is also
mixture (Fig. 4.4) yielded results similar to those observed clear that N2O is a product of the CN–NO reaction, similar
during the treatment with the O2/He mixture (Fig. 4.2), in to the CN–NO2 case. The above results are consistent with
that the cyanide peak almost completely disappeared af- the findings of Bell et al. (34, 35) who reported that surface
ter 2 min of treatment, indicating that NO2 reacts with this cyanide species adsorbed on Mn and Co-ZSM-5 catalysts
species at a rate comparable to that of O2. N2O appears to are reactive in NO, O2 and NO2 environments.
be one of the products of this reaction, since the intensity
of the N2O doublet after 2 min of treatment is higher in the cyanide species and NO, O2, and NO2 were determined
case of NO2 (Fig. 4.4b) than in the case of O2 (Fig. 4.3b). during transient reactor experiments conducted in paral-
The products of the reactions between the surface
The unexpected nature of the results obtained during the lel to the FTIR studies. The surface cyanide species was
treatment with the NO in He mixture prompted us to fur- formed in this case by passing a mixture of 2000 ppm NO/
ther investigate this case utilizing isotopically labeled 15NO. 2000 ppm C3H6/1% O2/He through a fixed-bed reactor con-
�
1
� 1
The results of this treatment are shown in Fig. 4.5. In this taining the Pt/SiO2 sample, at a rate of 12,000 ml h g .
�
1
case, the cyanide peak is shifted to 2080 cm after 2 min The NO/C3H6/O2/He mixture was first flowed through the
and to 2060 cm , the position corresponding to –C N, af- sample at 250 C for a period of 10 min. The reactor was
�
1
15
�
ter 4 min of treatment. At the same time, the N2O doublet subsequently flushed with He for a period of 2 min to purge
�
1
is shifted to 2140 and 2170 cm within the first two min- the gas phase. This treatment was followed by a second
utes of treatment and remains visible at this position even pulse of either O2, NO, NO2, or He for 0.5 min. The IR
after 6 min. These results are consistent in terms of the results indicate that the adsorbed cyanide is still present on
overall intensities of the observed peaks with the results the catalyst surface following the 2-min He purge. When
obtained with 14NO. The observed shifts in position, how- the sample was treated with He for a period of 2.5 min, a

2
28
CAPTAIN AND AMIRIDIS
in this case to facilitate on-line product analysis) for a pe-
riod of 0.5 min following the 2-min He purge, resulted in
the formation of N2 (161 ppm) and a small amount of N2O
(
14 ppm). Treatment of the sample with 2000 ppm NO
for the same period also resulted in the formation of N2
322 ppm) and N2O (108 ppm). The amounts of N2 and
(
N2O however, are much greater in this case. Almost identi-
cal amounts of N2 (326 ppm) and N2O (111 ppm) were ob-
served in the exit stream when the sample was treated with
2
000 ppm NO2. The above results indicate that N2 and CO2
are formed during the reaction of –CN and O2, whereas
N2O is also formed during the reaction of –CN with NO
and NO2. Furthermore, the similar results obtained with
NO and NO2 show that the presence of NO2 does not sig-
nificantly enhance the CN –NOx reaction.
To further investigate the origin of the cyanide species,
studies were conducted with NO/He, NO2/He, and NO/
O2/He mixtures being flowed over a surface that had been
preexposed to C3H6/He or C3H6/O2/He mixtures. The spec-
tra collected during the flow of NO/He over a Pt/SiO2 sam-
ple preexposed to C3H6 are shown in Fig. 5.1. Figure 5.1a
is the spectrum of a surface exposed to a flow of C3H6
for 5 min and subsequently flushed with He for 2 min, in
order to remove weakly adsorbed and gas phase propy-
lene. NO was then flowed over this surface, and spectra
were recorded at different time intervals as can be seen in
Figs. 5.1b–5.1g. The results indicate that the flow of NO re-
sults in the formation of both the surface cyanide and the
nitrous oxide product after a period of 4 min. Both species
are present in very small amounts and remain on the sur-
face even after 20 min of NO flow. The spectra of Fig. 5.1
demonstrate that the cyanide species can be formed by the
reaction between an adsorbed hydrocarbon derivative and
NO. When NO2 was flowed over a similarly pretreated sur-
face, no cyanide formation was observed, indicating that
NO alone is responsible for this reaction.
�
FIG. 4.5. In situ infrared spectra at 250 C of an initially oxidized
Pt/SiO2 catalyst: (a) 2000 ppm NO, 2000 ppm C3H6, 2% O2 in He for
1
5
4
min; (b) flushing of (a) with 2000 ppm NO in He for 2 min; (c) flushing of
a) with 2000 ppm NO in He for 4 min; (d) flushing of (a) with 2000 ppm
1
5
(
1
5
NO in He for 6 min.
significant decrease in the amount of N2 (from 423 to
4 ppm) and N2O (from 276 to 0 ppm) was observed in
the exit stream (Table 1). Treatment of the sample with
6
2
000 ppm O2 (a concentration lower than the 1% was used
Next, the same experiment was conducted, with oxygen
being co-fed to the catalyst together with C3H6. In this case,
when the catalyst was subsequently exposed to NO, the
TABLE 1
Product Analysis during Transient Reactor Studies
� 1
observed intensity of the surface cyanide peak at 2090 cm
was significantly higher (Figs. 5.2b–5.2f). This result clearly
demonstrates that the presence of O2 during the adsorption
of C3H6 eventually promotes the formation of the cyanide
species, through what probably is an activation mechanism
of the hydrocarbon. This result is consistent with our kinetic
studies, which show that under similar conditions the rate
of NO reduction by C3H6 is promoted by O2 (5).
Exit stream concentrations
Treatment
N2 (ppm)
CO2 (ppm)
N2O (ppm)
NO/C3H6/O2/He (10 min)
423
64
2411
102
276
0
NO/C3H6/O2/He (10 min)
He (2.5 min)
NO/C3H6/O2/He (10 min)
He (2 min)
O2/He (0.5 min)
161
322
326
355
244
630
14
108
111
The same experiments were repeated with the catalyst
treated with C H in the absence of O , followed by a short
3
6
2
NO/C3H6/O2/He (10 min)
He (2 min)
NO/He (0.5 min)
(1 min) pulse of O . The cell was then once again flushed
2
with He, and NO was subsequently introduced. The spec-
tra collected in this case (Figs. 5.3b–5.3g) show a behavior
similar to the one observed when O2 was present during the
treatment of the catalyst with C3H6 (Fig. 5.2). In both cases,
there is a significantly higher amount of surface cyanide
NO/C3H6/O2/He (10 min)
He (2 min)
NO2/He (0.5 min)

NO REDUCTION BY PROPYLENE OVER Pt/SiO2
229
�
FIG. 5.1. In situ infrared spectra at 250 C of an initially oxidized Pt/SiO2 catalyst, preexposed to 2000 ppm C3H6 in He for 5 min: (a) flushing with
�
He for 2 min; followed by flushing with 2000 ppm NO in He for (b) 2, (c) 4, (d) 6, (e) 10, (f) 12, and (g) 20 min. (5.2) In situ infrared spectra at 250 C of
an initially oxidized Pt/SiO2 catalyst, preexposed to 2000 ppm C3H6, 2% O2 in He for 5 min: (a) flushing with He for 2 min; followed by flushing with
2
000 ppm NO in He for (b) 2, (c) 4, (d) 6, (e) 10, (f) 12, and (g) 20 min.
formed, indicating the promoting effect of O2. The results rotational vibration peaks of gas phase H2O (Fig. 6b). Our
of Fig. 5.3 further demonstrate that oxygen reacts with an kinetic results indicate that the performance of the Pt/SiO2
adsorbed hydrocarbon species to form the activated species catalyst is not affected by the presence of H2O, which is
that eventually reacts with NO toward CN. It is important consistent with the spectra shown in Fig. 6.
to note that when a longer pulse of oxygen was used (i.e.,
5
min) no evidence was found in the IR spectra for the for-
Mechanistic Implications
mation of either the surface cyanide or the nitrous oxide
product. On the contrary, experiments with a prolonged
He flushing of 5 min over a surface preexposed to C3H6,
yielded results identical to those of Fig. 5.1, indicating that
desorption of the active hydrocarbon species is not a prob-
lem. The results with the prolonged O2 pulse suggest that
oxygen can also react further and consume the activated
hydrocarbon species, which is the precursor to –CN. These
observations are once again consistent with the results of
our kinetic studies (5), which demonstrated the dual role of
oxygen in the reduction of NO: as an activating agent (at
low concentrations) and as a competing oxidizing agent for
the activated hydrocarbon (at higher concentrations).
Finally, the reduction of NO by propylene was conducted
Based on the results presented in this paper and our ear-
lier kinetic studies (5), the following reaction pathway is
proposed for the selective catalytic reduction of NO by
propylene over Pt/SiO2 at temperatures below the temper-
ature of maximum NO conversion:
�
C3H6 + � → Cx H
[1]
[2]
[3]
[4]
[5]
[6]
[7]
y
�
�
�
C H + O /O → I
x
2
y
�
�
I + O2/O → CO2 + H2O
�
0�
Cx H + NO → I
y
�
0�
�
�
�
I /I + NO/NO → CN
in the presence of approximately 1% H2O at 250 C, and
the spectrum collected at steady state is shown in Fig. 6b.
The in situ spectra collected in the absence and presence of
water are very similar with the exception of the presence of
�
�
CN + O2/O → CO2 + N2
�
�
CN + NO /NO → N O + N + CO .
x
2
2
2
x

2
30
CAPTAIN AND AMIRIDIS
The steps of this pathway are not intended to represent
formal stoichiometrically balanced chemical reactions. In-
stead, they are simply indicating the origin and reactivity of
the various species that are, in our opinion, involved in the
reaction.
Step 1 indicates that an absorbed hydrocarbon species
is involved in the reaction, consistent with the results pre-
sented in Fig. 5. Steps 2 and 3 demonstrate the dual role
of oxygen as an activating agent for propylene (step 2) and
as a competing oxidizing agent (step 3) and are consistent
with the results shown in Figs. 5.2 and 5.3 and the previous
kinetic results. The exact nature of the activated hydrocar-
�
bon intermediate (I ) is not currently known, although it
is clear from the results of Figs 5.2 and 5.3 that this species
is adsorbed on the catalyst. Our experiments cannot differ-
entiate and identify the form of oxygen involved in this, as
well as subsequent steps (i.e., gas phase or adsorbed, molec-
ular, or atomic), and hence the two general forms are shown
in the proposed pathway. Step 4 suggests that the reaction
toward the formation of an active intermediate, and even-
tually –CN, can also proceed in the absence of oxygen as
indicated by the results of Fig. 5.1. The rate of this reaction
and/or the subsequent reaction for the formation of –CN is
FIG. 5.3. In situ infrared spectra at 250 C of an initially oxidized significantly lower than the rate of step 2. Furthermore, it is
�
Pt/SiO2 catalyst, preexposed to 2000 ppm C3H6 in He for 5 min and sub-
sequently flushed with 2% O2 in He for 1 min: (a) flushing with He for
min, followed by flushing with 2000 ppm NO in He for (b) 2, (c) 4, (d) 6,
not clear if the same activated intermediate is formed both
in the presence and absence of oxygen. Step 5 representsthe
formation of the surface cyanide species. Evidence for this
2
(
e) 10, (f) 12, and (g) 20 min.
�
FIG. 6. In situ infrared spectra at 250 C of an initially oxidized Pt/SiO2 catalyst: (a) 2000 ppm NO, 2000 ppm C3H6, 2% O2 in He for 1 h;
(
b) 2000 ppm NO, 2000 ppm C3H6, 2% O2, 1% H2O in He for 1 h.

NO REDUCTION BY PROPYLENE OVER Pt/SiO2
231
step is abundant in the spectra of Figs. 4.5 and 5.1–5.3. The reaction intermediate in the reduction of NO by C3H6. This
exact form of the nitrogen oxide involved (i.e., gas phase or species is held relatively weakly on the catalyst surface and
adsorbed) is debatable, although our kinetic and spectro- is reacting with O2, NO, and NO2. The formation of N2,
scopic results appear to suggest that gas phase or weakly N2O, and CO2 was observed as a result of the CN–NO and
adsorbed nitric oxide (NO) is primarily involved. Step 6 CN–NO2 reactionswhile N2 and CO2 are the productsofthe
represents the oxidation of the surface cyanide intermedi- CN–O2 reaction. Additionally, the surface concentration of
ate by oxygen. Evidence for this step is shown in Fig. 4.2. the cyanide was not affected by the presence of water in the
The products of this reaction, as shown in Table 1, are pri- reacting gas mixture. The conclusions of these FTIR studies
marily N2 and CO2. Finally, step 7 represents the reaction of are in excellent agreement with our previous kinetic work
–
CN with nitrogen oxides as evident from the results shown with the same catalyst.
in Figs. 4.4 and 4.5. N2, N2O, and CO2 are the products of
We have recently reported the results of similar FTIR
this reaction (Table 1). The mechanisms of N2 and N2O for- studies of the same reaction over a Pt/Al2O3 catalyst with
mation are not currently understood clearly. One possible the same Pt content (19). In that case, and in agreement
pathway includes the formation of surface atomic nitrogen with other published reports (36–41), we concluded that a
(
N), which can desorb as N2 or combine with NO to form surface isocyanate species (–NCO) is a potential reaction
N2O.
intermediate. The presence of a surface cyanide species was
The proposed mechanism cannot answer every question also identified on the Pt/Al2O3 catalyst, but this species was
about this reaction; in fact, it generates several questions very stable and did not react with either O2 and/or NOx. We
as pointed out in the previous paragraph. It is not clear yet have not been able to detect the formation of any surface
if this pathway is the only one in operation, for example. isocyanates on the Pt/SiO2 catalyst under the conditions
What is clear however, is that the proposed scheme is a studied. The isocyanate species is known to adsorb on SiO2
�
1
major reaction pathway for the reduction of NO under the with a characteristic IR peak at 2310 cm (31). Since this
conditions of this study. Strong experimental evidence from is in the region of the IR absorption of gas phase CO2 (a
both our kinetic and FTIR studies supports this statement. product of the SCR reaction), it is possible that a weak peak
in this region is covered in our spectra by the strong CO2
�
1
doublet between 2360 and 2320 cm
.
CONCLUSIONS
A reaction mechanism was proposed for the selective
catalytic reduction of NO by C3H6 over a Pt/SiO2 catalyst.
While it is conceivable that other pathways may also be
in operation, experimental evidence from our infrared and
kinetic work strongly supports the notion that the proposed
scheme is a major route for the reduction of NO under the
conditions studied.
�
Introduction of NO at 250 C to an oxidized Pt/SiO2 sam-
ple did not result in the formation ofanysignificant amounts
of adsorbed NOx-type species, associated with either the
Pt catalyst or the silica support. On the contrary, the ad-
sorption of C3H6 under the same conditions resulted in the
formation of carboxylate species on the catalyst surface. Pt
catalyzes the formation of these species, although it is possi-
ble, that following their formation they migrate to the silica
support. The necessary oxygen for these reactions is present
ACKNOWLEDGMENTS
The authors express their gratitude to the National Science Foundation
on the catalyst surface as a result of the oxidation pretreat- (CTS-9624433) and the Department of Energy (DE-FG02-96ER14666)
for financial support of this work.
ment. As expected, when oxygen is co-fed with propylene,
the intensity of all peaks associated with these partial oxi-
dation products is significantly higher.
REFERENCES
Under reaction conditions the formation of two new
species is observed with characteristic IR peaks in the 2000–
1
2
3
. Iwamoto, M., and Hamada, H., Catal. Today 10, 57 (1991).
. Held, W., Koening, A., Richter, T., and Puppe, L., SAE Paper 900496.
. Amiridis, M. D., Zhang, T., and Farrauto, R. J., Appl. Catal. B 10, 203
(1996).
�
1
2
300 cm region. These have been assigned to a surface
�
1
cyanide (peak at 2090cm ) and a gaseous/weaklyadsorbed
�
1
4
5
6
7
. Roberts, K. L., and Amiridis, M. D., Ind. Eng. Chem. Res. 36, 3529
nitrous oxide (doublet at 2236 and 2204 cm ) species. Pt
catalyzes the formation of both. Experiments conducted in
an attempt to probe the origin of the surface cyanide species
indicate that this species is formed as a result of the reaction
between NO and a surface hydrocarbon derivative. Small
amounts of oxygen assisted the activation of the adsorbed
hydrocarbon and enhanced the formation of the surface
cyanide.
(
1997).
. Captain, D. K., Roberts, K. L., and Amiridis, M. D., Catal. Today 42,
3 (1998).
. Sasaki, M., Hamada, H., Kintaichi, Y., Ito, Y., and Tabata, M., Catal.
Lett. 15, 297 (1992).
9
. Hirabayashi, H., Yahiro, H., Mizuno, N., and Iwamoto, M., Chem. Lett.
2235 (1992).
8
9
. Burch, R., and Millington, P. J., Catal. Today 29, 37 (1996).
. Burch, R., Millington, P. J., and Walker, A. P., Appl. Catal. B 4, 65
The reactivity of the surface cyanide was also studied,
(
1994).
and the results show that this species could be a potential 10. Burch, R., and Ottery, D., Appl. Catal. B 9, L19 (1996).

2
32
CAPTAIN AND AMIRIDIS
1
1
1
1
1
1. Yahiro, H., Hirabayashi, H., Shin, H. K., Mizuno, N., and Iwamoto,
M., Trans. Mater. Res. Soc. Jpn. A 18, 409 (1994).
28. Hoost, T. E., Laframboise, K. A., and Otto, K., Catal. Lett. 37, 153
(1996).
2. Efthimiadis, E. A., Lionta, G. D., Christoforou, S. C., and Vasalos, 29. Escribano, E. S., Busca, G., and Lorenzelli, V., J. Phys. Chem. 94, 8939
I. A., Catal. Today 40, 15 (1998). (1990).
3. Hamada, H., Kintaichi Y., Sasaki, M., Ito, T., and Tabata, M., Appl. 30. Finocchio, E., Busca, G., Lorenzelli, V., and Willey, R. J., J. Catal. 151,
Catal. 75, L1 (1991). 204 (1995).
4. Tanaka, T., Yokota, K., Isomura, N., Doi, H., and Sugiura, M., Appl. 31. Morrow, B. A., and Cody, I. A., J. Chem. Soc. Faraday Trans. I 71, 1021
Catal. B 16, 199 (1998).
5. Bamwenda, G. R., Obuchi, A., Ogata, A., Oi, J., Kushiyama, S., and
Mizuno, K., J. Mol. Catal. A 126, 151 (1997).
(1975).
32. Levoguer, C. L., and Nix, R. M., J. Chem. Soc. Faraday Trans. 92, 4799
(1996).
1
1
6. Gulari, E., and Seker, E., J. Catal. 179, 339 (1998).
7. Eckhoff, S., Hesse, D., van den Tillaart, J. A. A., Leyrer, J., and Lox,
E. S., Stud. Surf. Sci. Catal. 116, 223 (1998).
33. Richardson, W. S., and Wilson, E. B., J. Chem. Phys. 18, 694
(1950).
34. Aylor, A. W., Lobree, L. J., Reimer, J. A., and Bell, A. T., J. Catal. 170,
390 (1997).
35. Lobree J. J., Aylor, A. W., Reimer, J. A., Bell, A. T., J. Catal. 169, 188
(1997).
36. Bamwenda, G. R., Ogata, A., Obuchi, A., Oi, J., Mizuno, A., and
Skrzypek, J., Appl. Catal. B 6, 311 (1995).
1
8. Monroe, D., DiMaggio, C. L., Beck, D. D., and Matekunas, F. A., SAE
Paper 930737, 1993.
1
2
9. Captain, D. K., and Amiridis, M. D., J. Catal. 184, 377 (1999).
0. Djonev, B., Tsyntsarski, B., Klissurski, D., and Hadjiivanov, K.,
J. Chem. Soc. Faraday Trans. 93, 4055 (1997).
2
1. Ghiotti, F., Garrone, E., and Boccuzzi, F., J. Phys. Chem. 91, 5640 37. Coronado, J. M., and Anderson, J. A., J. Mol. Catal. A 138, 83
(
1987).
(1999).
2
2
2. Wilson, M. K., and Badger, R. M., Phys. Rev. 76, 472 (1949).
3. Nightingale, R. E., Downie, A. R., Rotenberg, D. L., Crawford, B.,
and Ogg, R. A., J. Phys. Chem. 61, 1047 (1954).
38. Tanaka, T., Okuhara, T., and Misono, M., Appl. Catal. B 4, L1
(1994).
39. Okuhara, T., Hasada, Y., and Misono, M., Catal. Today 35, 83
2
2
4. Laane, J., and Ohlsen, J. R., Prog. Inorg. Chem. 27, 465 (1980).
5. Gerei, S. V., Rozhkova, E. V., and Goraokhovatsky, Y. A. B., J. Catal. 40. Obuchi, A., Wogerbauer, C., Koppel, R., and Baiker, A., Appl. Catal.
(1997).
2
8, 341 (1973).
B 19, 9 (1998).
2
2
6. Long, R. Q., and Yang, R. T., J. Phys. Chem. B 103, 2232 (1999).
7. Busca, G., Catal. Today 27, 457 (1996).
41. Haneda, M., Kintaichi, Y., Inaba, M., and Hamada, H., Catal. Today
42, 127 (1998).