Intermediates in the selective reduction of NO by propene over Cu-Al2O3 catalysts: Transient in-situ FTIR study

Article abstract of DOI:10.1021/jp9930705

Systematic in-situ FTIR studies of the formation and the reaction of adsorbed species in C₃H₆-SCR on Cu-Al₂O₃ catalyst were conducted, particularly to clarify the role of adsorbed acetate. The reaction started with the formation of adsorbed nitrates via the NO oxidation by O₂ and the formation of the acetate via the partial oxidation of propylene with O₂ and nitrates. The acetate (the predominant adsorbed species in the steady-state condition) was the key intermediate of C₃H₆-SCR on a series of Cu-Al₂O₃. On the surface, acetate acted as an active reactant and reduced nitrates to N₂, which could be the rate-determining step over a wide temperature range. In this step, Al-NCO, Cu-NCO, and CN species were produced. Cu-NCO and possibly CN species were reactive toward nitrates and NO to yield N₂ and CO₂. Thus, they could be seen as final intermediates while Al-NCO species could be relatively inactive on the surface. The proposed reaction mechanism showed that the role of O₂ in promoting SCR was activating both NO and propylene into the reactive species in the reaction's preliminary steps. Cu²⁺ was the main active component in Cu-Al₂O₃ catalysts, playing a vital part in all important stages: the activation of hydrocarbon to the acetate; oxidation of NO to nitrates; the reaction of the acetate with nitrates; and the subsequent formation of the final intermediate (Cu-NCO) resulting in N₂ production.

Full text of DOI:10.1021/jp9930705

J. Phys. Chem. B 2000, 104, 2885-2893
2885
Intermediates in the Selective Reduction of NO by Propene over Cu-Al2O3 Catalysts:
Transient in-Situ FTIR Study
Ken-ichi Shimizu,* Hisaya Kawabata, Hajime Maeshima, Atsushi Satsuma,* and
Tadashi Hattori
Department of Applied Chemistry, Graduate School of Engineering, Nagoya UniVersity, Chikusa-ku,
Nagoya 464-8603, Japan
ReceiVed: August 30, 1999; In Final Form: NoVember 23, 1999
The mechanism of the selective catalytic reduction (SCR) of NO by C
3
H
6
on Cu-Al
consist of highly dispersed Cu2 ions in the surface aluminate phase, are investigated by in-situ FTIR
spectroscopy. During NO + C + O reaction, the acetate is produced via the partial oxidation of C
and becomes the predominant adspecies in the steady-state condition at 473-623 K. The acetate, which is
stable in NO, is quite reactive with NO + O , leading to the formation of isocyanate species (Cu-NCO) on
the surface and N and CO in the gas phase. The rate of acetate reaction in NO + O is close to the steady-
2 3
O catalysts, which
+
H
3 6
2
3 6
H
2
2
2
2
state rate of NO reduction over wide range of temperature, indicating that the acetate is an intermediate in the
SCR and takes part in the rate-determining stage. A mechanism is proposed; the acetate and nitrates, formed
by NO + O
and CO
. This mechanism explains the role of oxygen in facilitating SCR. Cu2+ ion is the principal active
component in Cu-Al
the acetate with nitrates.
2
, react to generate the Cu-NCO species, then Cu-NCO reacts with nitrates or NO to produce
N
2
2
O
2 3
catalysts; it plays crucial roles in all the important steps, including the reaction of
intermediacy of the adsorbed species.1
1-13
As for the mechanistic
Introduction
study of HC-SCR, very few attempts were devoted to this aspect,
and hence the reaction mechanism of HC-SCR has not yet been
fully clarified. In addition, most of the detailed mechanistic
studies have been focused on the metal-exchanged zeolites,1
while rather few have focused on the mechanism over the metal
Selective catalytic reduction of nitrogen oxides by hydrocar-
bons (HC-SCR) has attracted much attention, because it has a
potential ability to remove NOx from diesel and other oxygen
4-20
1,2
rich exhausts. Although the majority of the HC-SCR catalysts
1
,2
reported are metal-containing zeolites, such as Cu-ZSM-5,
2
1-23
oxide based catalysts.
we have recently reported that Cu-aluminate catalysts, contain-
ing highly dispersed Cu2+ ions in the aluminate phase, showed
high de-NOx activity comparable to Cu-ZSM-5 and higher
The role of transition metal and the support in the reaction is
also a matter of controversy. For example, Hamada et al.24
studied SCR by C3H6 on Co/Al2O3, and suggested that the
transition metal and the support play a different role; NO is
oxidized to NO2 on impregnated cobalt species, and NO2 reacts
with C3H6 to form N2 on Al2O3. On the other hand, we have
recently reported a reaction study of HC-SCR on the transition
metal aluminate catalysts, and suggested that a primary role of
Al2O3 support is to atomically disperse the transition metal
cations, which are responsible for the activity of HC-SCR.3
3
hydrothermal stability than Cu-ZSM-5.
Extensive studies have been devoted for understanding the
mechanism of HC-SCR. In earlier studies, the reaction mech-
anism was proposed on the basis of some empirical observations.
For example, the observation that reaction does not take place
without oxygen has been discussed in terms of the role of
oxygen and attributed to different causes. One is the possible
formation of NO2 as a key intermediate.3,4 It has been also
suggested that the partial oxidation of hydrocarbon molecules
In this study, we have carried out systematic in-situ FTIR
studies of the formation and the reaction of adsorbed species
in C3H6-SCR on Cu-Al2O3 catalyst. It was of particular interest
to clarify the role of adsorbed acetate; the transient reactions of
acetate with NO + O2 were followed by IR and GC analysis,
and the reaction rates of surface and gas-phase molecules were
compared. Further, the transient reaction rate is compared with
the steady-state rate of NO reduction to prove that the acetate
is an intermediate of this reaction. By comparison with the
results on Al2O3, the role of dispersed Cu ions and the Al2O3
support in each reaction steps is discussed.
to generate HxCyOz species could be the active reductant for
3
,5,6
NOx.
Afterward, various surface species, such as NOx
7
,8
7,9
7,9,10
adspecies,
NCO,
and CN
species, formed during
adsorption of reagents were identified by using IR, and they
were suggested to be possible intermediates. As Matyshak et
al. noted,11 most of the early mechanistic studies were performed
under conditions that were not close to those of catalytic
reaction. Recently some studies have succeeded to observe
surface complexes in situ (under reaction conditions). However,
observed species are not guaranteed to be directly involved in
the reaction pathway, unless dynamic response of them is
studied. Further, the kinetics of the reaction of the adsorbed
species should be tested and compared with the rate of the
steady-state reaction to obtain quantitative evidence on the
Experimental Section
Catalysts. Al2O3 (gamma type, a reference catalyst JRC-
2
5
ALO-1A ) was supplied from the Catalysis Society of Japan.
*
Address to whom correspondence should be addressed.
Supported catalysts, Cu-Al2O3, were prepared by impregnating
1
0.1021/jp9930705 CCC: $19.00 © 2000 American Chemical Society
Published on Web 03/09/2000

2886 J. Phys. Chem. B, Vol. 104, No. 13, 2000
Shimizu et al.
TABLE 1: Physical Characteristics of the Catalysts
bulk Cu
mol %
surface Cu
mol %
surface area
m g
catalystsa
Al
b
2
-1
O
2 3
0
0
160
98
82
Cu(2)-Al
Cu(4)-Al
Cu(8)-Al
2
O
2
O
2
O
3
3
3
1.6
3.2
6.5
4.2
6.5
10.0
76
a
Cu content (wt %). b Determined from Al 2s and Cu 2p3/2 peaks
in XPS.
Al2O3 with an aqueous solution of copper acetate followed by
evaporation to dryness at 395 K and by calcination in air at
1
073 K for 12 h. The calcination condition of Cu-Al2O3, which
26,27
is similar to that employed by Hierl et al.
would result in
the formation of the copper aluminate phase on Al2O3 surface
as will be confirmed later. The specific surface area of the
sample was determined using nitrogen adsorption (BET method).
X-ray diffraction patterns of the powdered catalysts were taken
in a Rigaku RINT 1100. UV-vis spectra were measured using
UV-vis spectrometer (JASCO V-750) in a diffuse reflectance
mode. XPS measurements were performed with AXIS-HSi
photoelectron spectrometer using Al KR radiation (1485.9 eV).
In-Situ FTIR. The samples were pressed into 0.04-0.07 g
of self-supporting wafers and mounted into a quartz IR cell with
CaF2 windows. In situ IR spectra were recorded on a JASCO
FT/IR-620 equipped with the IR cell connected to a conventional
flow reaction system, which was used in our previous stud-
2 3
Figure 1. UV-vis diffuse reflectance spectra of Cu-Al O catalysts
of various Cu loading.
18,22,23
were two absorption bands at 1450 nm and 750 nm, which have
ies.
All the spectra were measured at the reaction
2
+
-1
been assigned to the electric transitions in Cu in tetrahedral
temperatures accumulating 100 scans at a resolution of 2 cm
.
27
and octahedral sites of spinel lattice, respectively. These results
A reference spectrum of the catalyst in He was subtracted from
each spectrum. Prior to each experiment the catalyst was (1)
heated in 6.7% O2/He at 773 K for 1 h, (2) cooled to the desired
temperature and purged for 30 min with He, (3) and then various
2
6,27
are similar to those observed by Hierl et al.
for Al2O3-
supported Cu catalyst, whose calculation condition is similar
to that employed in this study. According to their proposal, the
most probable picture of the Cu species in Cu-Al2O3 catalysts
3
-1
gas mixtures were fed at a flow rate of 42 cm min . The
concentration of NO, C3H6, and O2 in the gas mixtures were
2
+
is the surface spinel, in which Cu ions are incorporated in
2
6,27
the alumina surface.
Therefore, it was shown that Cu-Al2O3
1
000 ppm, 1000 ppm, and 6.7% with He balance, respectively.
catalysts employed in this study consist of the highly dispersed
Cu species surrounded by Al2O3.
The extinction coefficient of the acetate and nitrates on the
2
+
catalyst were determined in the same fashion as described in
23
Formation of Adsorbed Species during SCR Reaction.
Figure 2 shows the IR spectra of adsorbed species during NO
our previous study. The extinction coefficient of the acetate
-
17
-17
-1
2
and nitrates were 1.7 × 10
and 1.3 × 10
cm ‚cm /
+
C3H6 + O2 reaction at various temperatures (473-623 K) in
molecule, respectively. They were independent of the temper-
ature of IR measurement and the Cu loading on the catalyst
within an experimental error ((7%).
the steady-states. Figure 2a includes the spectra recorded at 523
K during C3H6 + O2 reaction and after an introduction of the
acetic acid (120 µmol/g-catalyst) for comparison. In flowing
NO + C3H6 + O2, two strong bands at around 1580-1590 and
Catalytic Test. To determine the steady-state and transient
rate of the gaseous product formation, separate experiments were
performed in a conventional flow reactor with 0.05-2.0 g of a
-1
1452 cm were observed. These bands were in good agreement
with the bands observed in flowing C3H6 + O2, and those of
acetic acid adsorbed on the same catalyst. These bands are very
close to those assigned to νas(COO) and νs(COO) of adsorbed
acetate, which were observed during C3H6-SCR on Al2O3 in
3
-1
catalyst at a flow rate of 100 cm min under the condition
where both NO and C3H6 conversions were below 30%. Product
analysis was performed by a NOx analyzer (Best BCL-100 uH)
and a gas chromatograph with 13X molecular sieve and Porapak
Q columns. The pretreatment of the sample and the concentra-
tions of reaction mixtures were the same as in the corresponding
in-situ IR experiments. It should be noted that no formation of
CO or N2O were observed in all the reaction experiments.
2
2,23
our previous studies.
Thus, the bands at around 1580-1590
-
1
and 1452 cm were assigned to νas(COO) and νs(COO) of the
adsorbed acetate, respectively. Weak bands at 1720, 1642, 1392,
-
1
1
376, 1311, and 1278 cm observed in NO + C3H6 + O2 were
in good agreement with the bands observed in C3H6 + O2. These
bands are therefore the nitrogen-free species and can be assigned
Results
-
1
18
-1 22,23
to carbonyl (1720 cm ), formate (1392 and 1376 cm ),
-
1 28-30
Structure of the Catalysts. Table 1 lists the bulk and surface
concentration of Cu and the BET surface area of the catalysts
employed. For all the Cu-Al2O3 catalysts, the surface concen-
tration of Cu atom as determined by XPS was higher than the
bulk Cu concentration, indicating a surface enhancement of Cu.
The XRD analysis of Cu-Al2O3 samples showed diffraction
and carbonate species (1642, 1311, and 1278 cm ).
A
-
1
small band at 1300 cm observed in NO + C3H6 + O2 may
2
3
be assigned to the unidentate nitrate. In the region of 2300-
-
1
2100 cm (Figure 1b), small bands at 2236, 2198, and 2150
-
1
-1
cm were observed. The band at 2198 cm , which is similar
to that assigned by Kung et al. to NCO species on Cu-ZrO2
(2190 cm ) and that assigned by Solymosi et al. to NCO
1
0
3
-1
31
lines due to a spinel phase, but no lines due to CuO. Figure 1
-
1
shows UV-vis diffuse reflectance spectrum of samples. There
species bound to the Cu ions (2180-2201 cm ), is assigned

Intermediates in Selective Reduction of NO by Propene
J. Phys. Chem. B, Vol. 104, No. 13, 2000 2887
Figure 4. Time dependence of the acetate concentration in (O) NO +
C
C
3
H
H
6
+O
+ O
2
and in (b) C
3
H
6
+ O
2
2 3
on Cu(8)-Al O and in (4) NO +
3
6
2
on Al at 523 K.
2 3
O
3 6
Figure 2. IR spectra of adsorbed species in a flow of NO + C H +
2
our previous study of C3H6-SCR on Al2O3,23 this phenomena
has been explained that the specific OH groups are reactive and
removed or exchanged upon formation of acetate, which results
in a coordination of acetate on the coordinatively unsaturated
M-O site (MCUS-O site). For Cu-Al2O3, the negative bands
O
2 3
for 200 min over Cu(8)-Al O at various temperatures. Figure 2b
includes the spectrum of adsorbed acetic acid and that in flowing C
3 6
H
+
O
2
for 200 min at 523 K.
-
1
appeared at around 3740 ( 4 cm , while two bands appeared
-
1
at 3764 and 3720 cm for Al2O3. This suggests that the
adsorption sites of acetate on Cu-Al2O3 and Al2O3 are different
in nature from each other. Together with the result that position
of the acetate band for Al2O3 was higher than that for Cu-
Al2O3, it is suggested that the acetate on Cu-Al2O3 catalysts
is not associated with the AlCUS-O site but the CuCUS-O site.
Figure 4 shows the time course of the acetate concentration
in a flow of NO + C3H6 + O2 on Cu(8)-Al2O3 at 523 K. The
acetate concentration was estimated by using integrated intensi-
-
1
ties of the acetate band (1452 cm ) and the extinction
coefficient of acetate. During C3H6-SCR reaction, the acetate
concentration increased with time-on-stream and became con-
stant. It should be noted that intensity of the negative band at
Figure 3. IR spectra of adsorbed species on Al
with various Cu loading in a flow of NO + C + O
23 K. (The spectra were normalized per gram and the surface area of
the catalysts.)
O
2 3
and Cu-Al
2
2 3
O
-
1
H
3 6
for 200 min at
3740 cm increased with time-on-stream in a flow of NO +
C3H6 + O2 (results not shown). This trend is likely to coincide
with an increase of the acetate concentration during NO + C3H6
+ O2 reaction, suggesting again the coordination of the acetate
on CuCUS-O sites. As shown in Figure 2, the acetate was also
formed in C3H6 + O2. Figure 4 includes the change in the acetate
band during the reaction of C3H6 + O2 at 523 K. The steady-
state concentration of acetate in NO + C3H6 + O2 was almost
the same as in C3H6 + O2. The initial rate of acetate formation
in NO + C3H6 + O2, determined from the slope of the curve
5
to NCO species bound to the Cu ions (Cu-NCO). The band at
-
1
2
236 cm can be assigned to NCO species bound to Al (Al-
NCO), since the band has been observed on Al2O3 by Ukisu et
3
2
-1
al. The band at 2150 cm can be assigned to CN spe-
1
0,32,33
cies.
Figure 3 shows the IR spectra recorded during NO + C3H6
O2 reaction in a steady-state at 523 K on Cu-Al2O3 with
+
2
various Cu loading and on Al2O3. In the case of Cu-Al2O3,
was 2.0 nmol/m s, and was 2 times higher than that in C3H6 +
-
1
2
two main bands due to the acetate (1580-1590 cm and 1452
O2 (1.0 nmol/m s). Change in the acetate concentration during
-1
cm ) were observed, together with small bands due to carbonyl
NO + C3H6 + O2 reaction on Al2O3 is also shown in Figure 4.
-
1
-1
-1
2
(
1720 cm ), binyl (1642 cm ), formate (1392 and 1376 cm ),
The initial rate of acetate formation on Al2O3 (0.15 nmol/m s)
-
1
-1
carbonate (1311 and 1278 cm ), Al-NCO (2236 cm ), Cu-
NCO (2198 cm ), and CN species (2150 cm ). A shoulder at
around 2260 cm can be assigned to Al-NCO.
was about 13 times lower than that on Cu(8)-Al2O3.
Formation and Reaction of Nitrates. In a flow of NO +
O2 over Cu(8)-Al2O3 catalyst (Figure 5A, spectrum c), couples
-
1
-1
-
1
31,32,34
These
-
1
results indicate that the acetate should be the main adspecies
during the reaction on a series of the catalysts tested. In the
case of Al2O3, νs(COO) of the acetate band is centered at 1465
of bands in the region of 1650-1500 cm and 1170-1300
-
1
cm were observed. These bands are generally assigned to the
ν3 split stretching vibration of the nitrates on the metal
-
1
22,23,35-37
cm , which is higher than those on Cu-Al2O3 catalysts (1452
oxides.
As shown in the figure, these bands are close
-1
cm ). In the O-H stretching region of the spectra (Figure 3c),
to those assignable to nitrates on Al2O3 (spectrum b) but slightly
-
1
37
bands with negative absorbance appeared above 3700 cm . In
differ in their peak positions. According to the literature, the

2888 J. Phys. Chem. B, Vol. 104, No. 13, 2000
Shimizu et al.
the catalyst was exposed to a flow of NO + O2, while recording
IR spectra as a function of time. Figure 6A shows the changes
in the integrated intensities of the bands due to acetate, Cu-
NCO, and nitrates with time on stream. The intensity of acetate
band steeply dropped by exposure to a flow of NO + O2.
Simultaneously, the band due to Cu-NCO appeared, and its
intensity increased with time up to 9 min and then decreased.
The nitrate band appeared after an induction period of 15 min,
and its intensity increased gradually with time. Figure 6B shows
the results of another set of experiments to determine the rate
of gaseous product formation in the reaction of adsorbed acetate
with NO + O2. When the catalyst, pretreated by the introduction
of acetic acid (120 µmol/g) at 523 K, was exposed to a flow of
NO + O2, the formation of N2 and CO2 was observed. The
formation rates of N2 and CO2 first increased with time and
passed through a maximum at 17 min, and then decreased. From
a comparison with the results in Figure 6A, it is likely that N2
and CO2 are produced through a consecutive reaction of the
acetate with NO + O2 (or possibly nitrates) via Cu-NCO
species. On the other hand, when the adsorbed acetate on the
catalyst was exposed to a flow of NO, the acetate band hardly
decreased and N2 and CO2 were hardly produced (results not
shown). From these results, it was shown that the acetate is
active as a reductant and reacts with NO + O2 (possibly with
nitrates) to produce N2 and CO2 via the formation and reaction
of Cu-NCO species. It should be noted that the reaction of the
formate, formed by introducing the formic acid (120 µmol/g)
on the same catalyst, with NO + O2 at 523 K did not produce
N2, but resulted in a rapid desorption and/or oxidation of the
formate to CO2.
Figure 5. (A) IR spectra of adsorbed species on Cu(8)-Al
and on Al (b) at 523 K. The spectra were measured (a) in a flow of
NO for 60 min, (b, c) in a flow of NO + O for 180 min, (d) He purge
for 10 min after (c), (e) in a flow of C for 10 min after (c), and (d)
in a flow of C + O for 10 min after (c). (B) Time dependence of
nitrate concentration on (O) Cu(8)-Al and (4) Al
2 3
O (a, c-f)
2 3
O
2
3
H
6
3
H
6
2
2
O
3
2 3
O .
-
1
-1
bands at 1544 and 1296 cm , the bands at 1574 and 1256 cm
and the bands at 1600 and 1208 cm-1 observed in this study
are assigned to unidentate, bidentate, and bridging nitrates,
respectively. In a flow of NO (spectrum a), bands assignable to
nitrates were observed at around 1600-1574 cm , though their
intensity was much less than in NO + O2. This indicates that
the formation of nitrates is greatly promoted by the presence
of O2.
-1
The reactivity of nitrates toward hydrocarbon reductant was
evaluated by the transient response of the IR spectra at 523 K.
-
1
The nitrate bands (in the region 1170-1300 cm ) formed in a
flow of NO + O2, which was relatively stable in He (spectrum
d), disappeared by exposing to a flow of C3H6 (spectrum e) or
C3H6 + O2 (spectrum f) for 10 min, and the bands due to acetate
Transient Reaction. The reactivity of adsorbed species
formed during C3H6-SCR reaction toward NOx species can be
evaluated by the transient response of IR spectra. The wafer of
Cu(8)-Al2O3 catalyst was first exposed to a flow of NO +
C3H6 + O2 for 200 min at 523 K, and then the flowing gas was
switched to He, NO, O2, or NO + O2. Figure 7 shows the change
in the IR spectra as a function of time after the feed was
switched to NO + O2. As NO + O2 was passed over the catalyst,
-
1
-1
(
1
1452 cm ), formate (1392, 1376 cm ), and carbonates (1311,
-1
276 cm ) simultaneously appeared. This indicates that nitrates
can chemically interact with hydrocarbon species and take part
in hydrocarbon oxidation to the carboxylates and carbonates.
-1
In addition, the bands due to Al-NCO (2236 cm ), Cu-NCO
-1
-1
(
2198 cm ), and CN (2150 cm ) also appeared by the reaction
of nitrates with C3H6 or C3H6 + O2, which indicates the
N-insertion from the nitrates to hydrocarbon or partially oxidized
hydrocarbon species.
-1
there was a gradual decrease in acetate band (1452 cm ), rather
-
1
steep decrease in Cu-NCO (2198 cm ) and CN bands (2150
-1
cm ), and in contrast, a slow decrease in Al-NCO band (2235
Figure 5B shows the time course of the nitrate concentration
during the nitrate formation in NO + O2 and the nitrate reaction
with C3H6 mentioned above. The reaction rate of nitrate in a
flow of C3H6 estimated from the slope of the curve (2.0 nmol/
-1
cm ). Meanwhile, nitrate band was progressively formed after
2 min. Figure 8 shows the change in the integrated intensity
1
of acetate bands. The intensity hardly decreased in He or in
NO, indicating that the acetate is stable in an inert gas and in
NO. The acetate band fairly decreased in O2, indicating that
the acetate fairly reacts with O2. The band significantly
decreased in NO + O2, indicating that the acetate has high
reactivity with NO + O2. It should be noted that Cu-NCO band
2
m s) was higher than that of nitrate formation in NO + O2
2
(
0.87 nmol/m s). Change in the nitrate concentration during
the same experiment on Al2O3 is also shown in Figure 5B. The
2
rate of nitrate formation on Al2O3 (0.44 nmol/m s) was about
2
times lower than that on Cu(8)-Al2O3. The rate of nitrate
-
1
2
(2198 cm ) also decreases in a flow of NO, O2, or He with
reaction with C3H6 over Al2O3 (0.13 nmol/m s) was about 15
times lower than that over Cu(8)-Al2O3.
rather lower rates than in NO + O2.
By using the flow reactor, the rate of NO2 formation in NO
Figure 9 shows the results of another set of experiments to
determine the rate of gaseous products formed during the same
series of transient reactions. The Cu(8)-Al2O3 catalyst was first
pretreated at 523 K for 200 min in a flow of NO + C3H6 + O2
to obtain the steady-state activity. The catalyst was then exposed
to a stream of He, O2, NO, or NO + O2 at 523 K. When the
pretreated catalyst was exposed to a flow of NO + O2, the
formation of N2 and CO2 was observed with a gradual decrease
in their formation rates. N2 was hardly produced in a flow of
O2, but CO2 was produced with a slightly lower formation rate
than in NO + O2. In a flow of He, the formation of N2 and
CO2 was hardly observed. In a flow of NO, the formation of
N2 and CO2 was observed, though their formation rates were
+
O2 reaction over Cu(8)-Al2O3 catalyst was measured at 523
K. The rate is listed in Table 2, together with the rates of other
reactions on Cu(8)-Al2O3 catalyst at 523 K determined by in-
situ IR and flow reaction experiments. The rate of NO2
2
-
formation (0.35 nmol/m s) was lower that the rate of NO3
2
formation (0.87 nmol/m s).
Reaction of Adsorbed Acetate with NO + O2. The
reactivity of the acetate toward NOx species was examined by
the in-situ IR experiment at 523 K. Acetic acid (120 µmol/g)
was first introduced to the IR disc of Cu(8)-Al2O3 as a pulse
in the He carrier at 523 K, which resulted in the formation of
the adsorbed acetate (the spectrum shown in Figure 2). Then,

Intermediates in Selective Reduction of NO by Propene
J. Phys. Chem. B, Vol. 104, No. 13, 2000 2889
TABLE 2: Rates in Steady-State and Transient Reactions at 523 K over Cu(8)-Al
2 3
O
2
rate nmol/m s
reaction conditions
species
formation consumption
comments
C
C
3
H
H
6
+ NO + O
+ O
2
2
2
2
acetate
acetate
acetate
nitrate
2.0a
1.0
Figure 4
Figure 4
Figure 10
Figure 5B
a
3
6
2
NO + O
NO + O
NO + O
flow after steady-state in C
3
H
6
+ NO + O
2
0.26a
2.0a
0.87a
0.35b
NO
nitrate
2
C
3
H
6
flow after NO + O
Steady-state in C + NO + O
2
Figure 5B
Figure 9A
Figure 9B
Figure 6B
3
H
6
2
N
2
0.14b
b
CO
2
1.9
c
NO + O
acid adsorption
2
flow after acetic
N
2
0.19
CO
2
1.9c
a
Initial rates in transient conditions. b Rates in steady-state conditions. c Rates at t ) 17 min in Figure 6B.
Figure 7. Dynamic changes in the IR spectra as a function of time in
at 523 K. Before the measurement,
flowing NO + O
2
2 3
on Cu(8)-Al O
the catalyst was pre-exposed to a flow of NO + C
3
H
6
2
+ O for 200
min at 523 K.
the steady-state rate constant of NO reduction, which indicates
that the steady-state N2 formation occurs by the reaction of the
adsorbed intermediates with NO + O2. On the other hand, the
-
rate of acetate consumption, d[CH3COO ]/dt, during the
-
transient reaction of CH3COO with NO + O2, which was
determined by using the extinction coefficient of the acetate
band and by numerically differentiating the IR data in Figure
8
, was plotted in Figure 10. Clearly, the rate of acetate
consumption was very close to the transient rate of NO reduction
to N2 over a wide range of reaction time. In addition, the rate
of acetate consumption was the same order of magnitude as
the rates of CO2 formation during the same transient reaction.
The first-order rate constant of acetate consumption was
-
4 -1
estimated to be 4.6 × 10 s , which was close to those of N2
Figure 6. Dynamic changes in the (A) surface and (B) gas-phase
-4 -1
-4 -1
(
2.5 × 10 s ) and CO2 (2.8 × 10 s ) formation. These
molecular as a function of time in a flow of NO + O
Cu(8)-Al , on which the acetic acid (120 µmol/g) was pre-
adsorbed: intensity of the IR bands due to (b) acetate, (1) Cu-NCO,
2
at 523 K over
results confirm that the NO reduction to N2 occurs by the
reaction of the adsorbed acetate with NO + O2.
2
O
3
-
and (]) NO
3
, rates of (O) N
2
and (4) CO
2
formation.
Comparison of Steady-State and Transient Reaction
Rates. As shown in Figure 2, the acetate was present as a
predominant adsorbed species in the steady-state condition over
a wide temperature range (473-623 K). For each temperature,
transient rates of acetate consumption in NO + O2 were
determined from the transient IR experiments in the same
manner as mentioned above. Figure 11 shows the initial rates
of acetate consumption for each temperature as an Arrenius plot.
Figure 11 includes the steady-state-rates of NO reduction to N2
and C3H6 oxidation to CO2 in NO + C3H6 + O2 reaction over
Cu(8)-Al2O3. It is shown that the initial rates of acetate reaction
were close to the steady-state rates of N2 and CO2 formation
over a wide range of temperatures. Furthermore, Arrenius plots
for these rates gave good straight lines, and the apparent
much less than in NO + O2. As mentioned above, the acetate
band hardly decreased, but Cu-NCO band decreased in a flow
of NO. Thus, the results should indicate that the reaction of
Cu-NCO species with NO resulted in N2 formation. This
3
8
interpretation is supported by the results of Iwamoto et al., in
which the NCO species on Cu-ZSM-5 catalyst gave N2 by the
reaction with NO.
Figure 10 shows the first-order plots of NO reduction to N2
and CO2 formation during the transient reaction of surface
adspecies with NO + O2. The result gave fairly good straight
lines. It should be noted that the extrapolation of the first-order
plot for NO reduction rate to the intercept was consistent with

2890 J. Phys. Chem. B, Vol. 104, No. 13, 2000
Shimizu et al.
Figure 8. Time dependence of the relative intensities of the acetate
band normalized by the initial intensity in (O) NO + O , (2) O , (4)
NO, and (b) He. Before the measurements, the Cu(8)-Al catalyst
2
2
Figure 10. First-order plots of acetate consumption (dotted line), (O)
2
O
3
NO reduction to N
23 K over Cu(8)-Al
+ O for 200 min at 523 K.
2
and (4) CO
2
formation in flowing NO + O
2
at
was pretreated in NO + C
3
H
6
+ O
2
for 200 min at 523 K.
5
C
2 3
O
, which was pre-exposed to a flow of NO +
H
3 6
2
Figure 9. Formation rates of N
time in flowing (O) NO + O
8)-Al . Before the measurements, the catalyst was pre-exposed to
a flow of NO + C + O for 200 min at 523 K.
2
2
(A) and CO (B) as a function of
2
2
, (2) O , (4) NO, and (b) He over Cu-
(
2 3
O
H
3 6
2
activation energy for the acetate reaction rate (76 kJ/mol) were
close to that of the steady-state rates of N2 formation (81 kJ/
mol).
For Cu-Al2O3 catalysts with various Cu loading, the initial
rates of acetate consumption in NO + O2 at 523 K were
determined from the transient IR experiments. These rates were
plotted in Figure 12 together with the steady-state NO reduction
and C3H6 conversion rates. Again, the rate of acetate reaction
with NO + O2 and NO reduction rate were very close to each
other for each Cu-Al2O3 samples. In addition, these two rates
increased in a similar manner with an increase in Cu loading.
Taking into account the result in Table 1 that the surface
concentration of Cu2 ion increases with an increase in Cu
Figure 11. Arrhenius plots of (b) the initial rate of acetate consumption
, and the steady-state rates of (O) NO reduction to N and
conversion to CO for NO + C + O reaction over Cu-
in NO + O
2
2
(
4) C
3
H
6
2
H
3 6
2
(8)-Al
2
O .
3
Discussion
Role of Acetate and Nitrates. As shown above, the formation
of nitrates is greatly promoted by the presence of O2 (Figure
2
3
5
A). In our previous study, the same result was obtained in
+
the case of the Al2O3 catalyst, and it was suggested that the
formation of nitrates proceeds by NO oxidation and the
subsequent adsorption of NO2 on basic oxygen sites. The nitrates
were quite reactive toward C3H6 at 523 K, leading to the
2+
loading, the above results indicate that Cu ion is the principal
active component, which plays an important role in the reaction
of the adsorbed acetate with NO + O2.

Intermediates in Selective Reduction of NO by Propene
J. Phys. Chem. B, Vol. 104, No. 13, 2000 2891
Figure 13. Proposed mechanism of C
catalysts.
3
H
6
2 3
-SCR over Cu-Al O
the result that the rate of acetate formation in NO + C3H6 +
2
O2 (2.0 nmol/m s) was about 8 times higher than that of the
2
acetate reaction in NO + O2 (0.26 nmol/m s), the above results
lead to the conclusion that the acetate is an intermediate of this
reaction on a series of Cu-Al2O3 catalysts and possibly takes
part in the rate-determining stage over a wide temperature range.
As shown in Figure 5A, the reaction of nitrates with C3H6 or
C3H6 + O2 results in the formation of N-containing organic
species, Al-NCO, Cu-NCO, and CN species, on the surface.
This indicates the N insertion from the nitrates to hydrocarbon
or carboxylate species. Combined with the above-mentioned
discussions and the result that nitrates were immediately formed
in NO + O2 (Figure 5B), it can be concluded that the acetate
react with nitrates to produce N2 and CO2 via Cu-NCO species,
though the presence of nitrates was not clearly confirmed during
SCR reaction (Figure 2). The absence of nitrate bands during
SCR reaction can be explained by the result shown in Figure
Figure 12. Effect of Cu loading on (b) the initial rate of acetate
consumption in NO + O
to N , and (4) C conversion to CO
at 523 K.
2
, the steady-state rates of (O) NO reduction
2
H
3 6
2
for NO + C + O reaction
3
H
6
2
5
B that the reaction rate of nitrate with hydrocarbon species is
formation of the acetate (Figure 5A). As shown in Figure 2,
the acetate was formed in NO + C3H6 + O2 and in C3H6 + O2,
and the rate of acetate formation was 2 times higher in the
former reaction. This result should indicate that during NO +
C3H6 + O2 reaction both oxygen and nitrates take part in the
C3H6 oxidation to the acetate.
higher than that of nitrates formation in NO + O2. The
intermediacy of nitrates is in agreement with the findings in
our previous studies on C3H6-SCR mechanism over Na-H-
1
8
23
mordenite and Al2O3, where it was also shown that the
nitrates are involved in the reaction with organic species on
the surface, which finally results in N2 formation. There have
been suggestions in the literature that gaseous NO2 formed from
the oxidation of NO is a possible intermediate of SCR reaction,
which subsequently reacts with hydrocarbons or surface organic
The acetate was the dominant adspecies in NO + C3H6 +
O2 reaction on Cu-Al2O3 at 473-623 K (Figure 2). As shown
in Figure 6, the acetate formed by introducing the acetic acid
is an active reductant, which reacts preferentially with NO +
O2 to produce N2 and CO2 via the formation and reaction of
Cu-NCO species. On the other hand, the surface species formed
during NO + C3H6 + O2 reaction in a steady-state reacted
preferentially with NO + O2 to produce N2 and CO2 (Figure
3
,4
species. However, considering the results that the NO2
2
formation rate in NO + O2 reaction (0.35 nmol/m s) was lower
2
that the rate of nitrates formation (0.87 nmol/m s) and nitrates
2
reaction in C3H6 (2.0 nmol/m s), and that the nitrates were
fairly stable in He purge, the reactive NOx species in the present
system should not be the gaseous NO2 but the nitrates (adsorbed
NO2) on the catalyst surface.
9). It should be noted that the N2 formation rate in this transient
reaction (Figure 9) is very close to that in the reaction of
adsorbed acetic acid with NO + O2 (Figure 6). Therefore, it is
the acetate formed during SCR reaction that acts as the dominant
reductant, which reacts with NO + O2 to produce N2 and CO2.
This conclusion is quantitatively confirmed by the result in
Figure 10, in which the transient rate of acetate consumption
in NO + O2 was the same order of magnitude as the transient
rates of NO reduction and CO2 formation over a wide range of
reaction time.
Reaction Mechanism of C H -SCR over Cu-Al O . On
3
6
2
3
the basis of the above results, a reaction mechanism for the
C H -SCR on Cu-Al O catalysts is presented as shown in
3
6
2
3
Figure 13. The reaction begins with the formation of adsorbed
nitrates via the NO oxidation by O and the formation of acetate
2
via the partial oxidation of propene with O and possibly with
2
nitrates. The acetate, which acts as a surface reductant, reduces
nitrates to N . In the course of this step, N-containing hydro-
2
As shown in Figure 10, the extrapolation of the first-order
plot for NO reduction rate to the intercept was consistent with
the steady-state rate constant of NO reduction, which indicates
that the steady-state N2 formation occurs by the reaction of the
adsorbed intermediates with NO + O2. As shown in Figure 11,
the initial rates of acetate reaction with NO + O2 were close to
the steady-state rates of NO reduction over a wide range of
temperature, and the apparent activation energy for the former
reaction was also close to that of the latter reaction. Furthermore,
for Cu-Al2O3 samples with various Cu loading (0 - 8 wt %),
the rates of acetate reaction with NO + O2 and the steady-state
NO reduction were very close to each other. Taking into account
carbon species, Al-NCO, Cu-NCO, and CN adspecies, are
formed. Cu-NCO and possibly CN species rapidly react with
nitrates and NO to produce N2 and CO2, and thus may be
regarded as the final intermediates. This is in agreement with
10
the results by Kung et al. in which NCO and CN species were
observed as possible intermediates for C3H6-SCR over Cu-
ZrO2 catalyst. In contrast, Al-NCO species is rather stable in
NO + O2, and thus can be a relatively inert spectator on the
9,32
surface. This conclusion is contrary to the earlier finding that
-
1
Al-NCO species (2270-2230 cm ) was proposed as the
intermediate for HC-SCR over Cu-based catalysts. The reaction
pathway in Figure 13 is similar to that reported in our previous

2892 J. Phys. Chem. B, Vol. 104, No. 13, 2000
Shimizu et al.
study18 where the consecutive reaction of surface organic
species, including N-containing hydrocarbon species, with the
nitrate was proposed. This scheme also includes the route of
the competitive oxidation of the C3H6 with O2 as follows. The
formation of formate and carbonates adspecies also occurs via
the partial oxidation of propene (Figure 2 and 5A). These species
do not act as the surface reductant, but disrobe or react with O2
to produce CO2. Acetate can also react with O2 to produce CO2
by O2 and the formation of the acetate via the partial oxidation
of C3H6 with O2 and nitrates. The acetate, which is the
predominant adsorbed species in the steady-state condition, is
the important intermediate of C3H6-SCR on a series of Cu-
Al2O3; it acts as an active reductant on the surface and reduces
nitrates to N2, which could be the rate-determining stage over
a wide temperature range. In the course of this step, Al-NCO,
Cu-NCO, and CN species are produced. Cu-NCO and possibly
CN species are reactive toward nitrates and NO to produce N2
and CO2 and thus can be regarded as the final intermediates,
while Al-NCO species can be a relatively inert spectator on
the surface. The mechanism presented in Figure 13 has been
proposed, which explains the role of oxygen in facilitating SCR;
oxygen can activate both NO and C3H6 into the reactive species
in the initial steps of the reaction.
(Figures 5B and 6B). These pathways contribute the unselective
oxidation of the C3H6 with O2, and compete with the SCR
reaction for the consumption of C3H6 reductant. As shown in
Figure 8, the rate of acetate oxidation by NO + O2 (nitrates)
was larger than that by O2. This result, along with the result
that nitrates were immediately formed in NO + O2, explains
why, in the presence of great excess of oxygen, hydrocarbons
interact selectively with NOx, instead of oxygen. Dissociative
2
+
Cu ion is the principal active component in Cu-Al O3
2
2+
adsorption of oxygen on isolated Cu ions may be hindered
as postulated by Kung et al.,39 whereas NO forms nitrates rather
easily, which could lead to the selective oxidation of the acetate
intermediate by nitrates.
catalysts. It plays an important role in all the important steps:
(1) the activation of hydrocarbon to the acetate; (2) oxidation
of NO to nitrates; (3) the reaction of the acetate with nitrates;
(4) subsequent formation of the final intermediate (Cu-NCO)
that results in N2 production.
It is generally accepted that the role of oxygen is primary of
importance to discuss the reaction mechanism of SCR. The
proposed mechanism explains the role of oxygen in facilitating
the reduction of NO by C3H6; oxygen activates both NO and
C3H6 into the reactive species in the initial steps of the reaction.
In addition, as Marquez-Alvarez et al. clarified, another role of
oxygen should be to maintain the copper atoms in an appropriate
Acknowledgment. K.S. acknowledges support by the Fel-
lowship of JSPD for Japanese Junior Scientists.
References and Notes
(
1) Iwamoto, M.; Yahiro, H. Catal. Today 1994, 22, 5.
(2) Shelef, M. Chem. ReV. 1995, 95, 209.
3) Shimizu, K.; Maeshima, H.; Satsuma, A.; Hattori, T. Appl. Catal.
B 1998, 18, 163.
2
+ 40
oxidation state (Cu ).
(
Role of Copper Ions in C3H6-SCR. The role of Cu ions in
the SCR reaction has been of another interest, and a number of
possibilities have been proposed. Among them are activation
of hydrocarbon, oxidation of NO to NO2, and the reaction of
activated hydrocarbon with NOx species and subsequent forma-
tion of N-containing reaction intermediates that results in N2
production. The mechanism shown in Figure 13 is basically
consistent with that proposed for Al2O3,23 which implies that
the presence of Cu ions does not markedly change the reaction
mechanism. However, the rates of acetate formation (Figure 4),
nitrate reaction (Figure 5B), and acetate reaction with NO +
(4) Hamada, H.; Kintaichi, Y.; Sasaki, M.; Ito, T.; Tabata, M. Appl.
Catal. 1991, 70, L15.
(5) Bennett, C. J.; Bennett, P. S.; Golunski, S. E.; Hayes, J. W.; Walker,
A. P. Appl. Catal. A 1992, 86, L1.
(
6) Jewell, L.; Sokolovskii, V. D.; Coville, N. J.; Glasser, D.; Hilderadt,
D. Chem. Commun. 1996, 2081.
7) Bamwenda, G. R.; Ogata, A.; Obuchi, A.; Oi, J.; Mizuno, K.;
Skrzypek, J. Appl. Catal. B 1995, 6, 311.
8) Hadjiivanov, K.; Klissurski, D.; Ramis, G.; Busca, G. Appl. Catal.
B 1996, 7, 251.
(
(
(9) Ukisu, Y.; Sato, S.; Muramatsu, G.; Yoshida, K. Catal. Lett. 1991,
11, 177.
1
O2 (Figure 12) were significantly promoted (by a factor of 10 -
(10) Li, C.; Bethke, K. A.; Kung, H. H.; Kung, M. C. J. Chem. Soc.,
Chem. Commun. 1995, 813.
11) Matyshak, V. A.; Il’ichev, A. N.; Ukharsky, A. A.; Korchak, V.
2
1
0 ) by the presence of Cu ions. The rate of nitrate formation
(
on Cu(8)-Al2O3 was also higher (by a factor of 2) than Al2O3
Figure 5B). The results in Figure 3 suggested that the adsorption
N. J. Catal. 1997, 171, 245.
(12) Tamaru, K. Dynamic Heterogeneous Catalysis; Academic Press:
London, 1978.
(
site of the acetate intermediate for Cu-Al2O3 catalysts is not
associated with AlCUS-O site but CuCUS-O site. Further, Cu
ions were found to be the adsorption site of the most probable
N-containing reaction intermediates (Cu-NCO) that can results
in N2 production. From these results, the surface Cu ions should
take part in all the important steps; (1) the activation of
hydrocarbon to the acetate; (2) oxidation of NO to nitrates; (3)
the reaction of the acetate with nitrates; (4) subsequent formation
of the final intermediates (Cu-NCO) that results in N2
production. This could be due to the superior redox property of
Cu ions. Thus, the C3H6-SCR reaction proceeds primarily on
(13) Matyshak, V. A.; Krylov, O. V. Catal. Today 1995, 25, 1.
(
14) Adelman, B. J.; Beutel, T.; Lei, G.-D.; Sachtler, W. M. H. J. Catal.
1
996, 158, 327.
(15) Aylor, A. W.; Lobree, L. J.; Reimer, J. A.; Bell, A. T. Stud. Surf.
Sci. Catal. 1996, 101, 661.
16) Lobree, L. J.; Aylor, A. W.; Reimer, J. A.; Bell, A. T. J. Catal.
997, 169, 188.
17) Xin, M.; Hwang, I. C.; Woo, S. I. J. Phys. Chem. B 1997, 101,
9005.
(18) Satsuma, A.; Enjoji, T.; Shimizu, K.; Sato, K.; Yoshida, H.; Hattori,
T. J. Chem. Soc., Farady Trans. 1998, 94, 301.
19) Chen, H.-Y.; Voskoboinikov, T.; Sachtler, W. M. H. J. Catal. 1998,
(
1
(
(
1
80, 171.
3
the surface Cu ions. As pointed out in our previous study, the
(20) Lobree, L. J.; Aylor, A. W.; Reimer, J. A.; Bell, A. T. J. Catal.
1999, 181, 189.
2
+
primary role of the Al2O3 support should be to disperse Cu
ions in an atomic level.
(
(
21) Long, R. Q.; Yang, R. T. J. Phys. Chem. B 1999, 103, 2232.
22) Shimizu, K.; Kawabata, H.; Satsuma, A.; Hattori, T. Appl. Catal.
B 1998, 19, L87.
Conclusions
(23) Shimizu, K.; Kawabata, H.; Satsuma, A.; Hattori, T. J. Phys. Chem.
B 1999, 103, 1542.
The reaction mechanism of C3H6-SCR on Cu-Al2O3 cata-
lysts, which consists of highly dispersed Cu2+ ions in the surface
aluminate phase, was investigated by in-situ FTIR spectroscopy,
and the following conclusions can be drawn. The reaction begins
with the formation of adsorbed nitrates via the NO oxidation
(24) Hamada, H. Catal. Today 1994, 22, 21.
(25) Murakami, Y. Stud. Surf. Sci. Catal. 1983, 16, 775.
(
26) Ertl, G.; Hierl, R.; Knozinger, H.; Thiele, N.; Urbach, H.-P. Appl.
Surf. Sci. 1980, 5, 49.
27) Hierl, R.; Knozinger, H.; Urbach, H.-P. J. Catal. 1981, 69, 475.
(28) Morterra, C.; Magnacca, G. Catal. Today 1996, 27, 497.
(

Intermediates in Selective Reduction of NO by Propene
J. Phys. Chem. B, Vol. 104, No. 13, 2000 2893
(
29) van Neer, F. J. R.; van der Linden, B.; Bliek, A. Catal. Today 1997,
8, 115.
30) Bando, K. K.; Sayama, K.; Kusama, H.; Okabe, K., Arakawa, H.
Appl. Catal. A 1997, 165, 391.
(35) Davydov, A. A. Infrared Spectroscopy of Adsorbed Species on the
Surface of Transition Metal Oxides; Wiley: New York, 1990; p 53.
3
(
(
36) Ramis, G.; Busca, G.; Lorenzelli, V.; Forzatti, P. Appl. Catal. 1990,
6
4, 243.
31) Solymosi, F.; Bansagi, T. J. Catal. 1995, 156, 75.
(
(
(
37) Parkyns, N. D. Proc. 5th Int. Congr. Catal. 1972, 1, 255.
38) Takeda, H.; Iwamoto, M. Catal. Lett. 1996, 38, 21.
39) Kung, H. H.; Kung, M. C. Catal. Today 1996, 30, 5.
1
6, 11.
(33) Captain, D. K.; Amiridis, M. D. J. Catal. 1999, 184, 377.
34) Sumiya, S.; He, H.; Abe, A.; Takezawa, N.; Yoshida, K. J. Chem.
(
(40) Marquez-Alvarez, C.; Rodriguez-Ramos, I.; Guerrero-Ruiz, A.;
Haller, G. L.; Fernandez-Garcia, M. J. Am. Chem. 1997, 119, 2905.
Soc., Faraday Trans. 1998, 94, 2217.