Kinetic and in situ infrared studies on SCR of NO with propane by silver-alumina catalyst: Role of H2 on O2 activation and retardation of nitrate poisoning

Article abstract of DOI:10.1016/j.jcat.2006.02.011

The mechanistic cause of enhanced C₃H₈-SCR activity of Ag/Al₂O₃ by H₂ addition was investigated by kinetic and in situ infrared (IR) studies. Kinetic studies found that hydrogen addition results in de

Full text of DOI:10.1016/j.jcat.2006.02.011

Journal of Catalysis 239 (2006) 402–409
www.elsevier.com/locate/jcat
Kinetic and in situ infrared studies on SCR of NO with propane
by silver–alumina catalyst:
Role of H2 on O2 activation and retardation of nitrate poisoning
Ken-ichi Shimizu ∗, Junji Shibata, Atsushi Satsuma
Department of Applied Chemistry, Graduate School of Engineering, Nagoya University, Chikusa-ku, Nagoya 464-8603, Japan
Received 25 October 2005; revised 8 February 2006; accepted 18 February 2006
Available online 23 March 2006
Abstract
The mechanistic cause of enhanced C H -SCR activity of Ag/Al O by H addition was investigated by kinetic and in situ infrared (IR) studies.
3
8
2
3
2
−1
Kinetic studies found that hydrogen addition results in decreased activation energy for NO reduction (from 224 to 61 kJ mol ), decreased reaction
orders with respect to O (from 1.38 to 0.01) and C H (from 1.91 to 0.76), and increased order with respect to NO (from −2.53 to 0.49). The
2
3 8
essential role of H is in the reductive activation of molecular oxygen into reactive oxygen species involved in the oxidative activation of C H .
2
3 8
Addition of hydrogen can also retard nitrate poisoning by reducing nitrates and thus decreasing nitrate coverage. N formation via the unsteady-
2
state reaction of nitrates with hydrogen and NH formation via the NO + H reaction also occur, demonstrating an additional effect of hydrogen
3
2
on the promotion of NO reduction under the unsteady-state and unselective conditions.
 2006 Elsevier Inc. All rights reserved.
Keywords: Nitrogen oxides; Selective reduction; Silver alumina
1. Introduction
or by NH3 [19] on Ag/Al2O3 and SCR by C3H8 on Ag zeolites
[20–22]. Recently, several authors have investigated mechanis-
tic causes of the hydrogen effect and have presented different
proposals. Burch posited that hydrogen destroys nitrate species
that are strongly adsorbed, thus poisoning the silver sites [4].
Brosius et al. [17] showed that hydrogen can promote the re-
moval of nitrates from Ag as well as the formation of nitrates
on the alumina support. We reported a series of spectroscopic
studies on the hydrogen effect and proposed several possible
mechanisms of the hydrogen effect on C3H8-SCR [12,13,20–
24]. First, we proposed that hydrogen acts not as a reducing
agent of NO, but rather as a promoter of the C3H8-SCR [12].
Second, we showed that hydrogen promotes the following re-
dox reactions over Ag/Al2O3 [12,13]: (1) oxidation of NO to
nitrates (adsorbed NO2 on the catalyst surface); (2) oxidation
of NO to NO2 in the gas phase; (3) partial oxidation of C3H8
to acetate; (4) oxidation of C3H8 to COx; (5) reaction of ni-
trates with a C3H8 + O2 mixture and subsequent formation of
acetate; and (6) reaction of acetate with a NO + O2 mixture.
Similar conclusions were reported by Sazama et al. for the hy-
drogen effect in decane-SCR [18]. Thus, it can be speculated
The selective catalytic reduction of NO by hydrocarbons
(HC-SCR) in the presence of excess oxygen is a potential
method of removing NOx from lean-burn and diesel exhausts
[1–4]. It is now well established that Ag/Al2O3 is among the
most active and selective catalyst for SCR by alcohol [5] or
by higher hydrocarbons [6–9] under lean-burn exhaust condi-
tions [3]. Ag/Al2O3 has moderate tolerance to water and SO2
[5–7,10]. One drawback of Ag/Al2O3 is its significantly low
activity at low temperatures (below ca. 700 K) for SCR by
light hydrocarbons, such as C3H8, that can be the main com-
ponent of unburned hydrocarbons in lean-burn and diesel ex-
hausts. Recently, Satokawa discovered a dramatic positive ef-
fect of H2 addition on C3H8-SCR activity of Ag/Al2O3 at low
temperatures [11]. Subsequently, several research groups have
investigated the “hydrogen effect” for different catalytic sys-
tems [12–24], including SCR by higher hydrocarbons [16–18]
*
Corresponding author. Fax: +81 52 789 3193.
E-mail address: kshimizu@apchem.nagoya-u.ac.jp (K.-i. Shimizu).
0021-9517/$ – see front matter  2006 Elsevier Inc. All rights reserved.
doi:10.1016/j.jcat.2006.02.011

K.-i. Shimizu et al. / Journal of Catalysis 239 (2006) 402–409
403
that H2 should play an essential role in the activation of mole-
cular oxygen, as suggested by Richter et al. [15], though there
currently is no evidence confirming that the activation of mole-
cular oxygen with hydrogen is responsible for the promotion
of the steady-state NO reduction. Third, we proposed that hy-
drogen promotes the partial reduction of isolated Ag+ cations
reaction system, as in our previous studies [7–9,12,13,25,26].
The sample was pressed into a 0.07 g of self-supporting wafer
and mounted into the quartz IR cell with CaF2
windows. Spec-
tra were measured accumulating 20–100 scans at a resolution of
2 cm−1. A reference spectrum of the catalyst wafer in He taken
at measurement temperature was subtracted from each spec-
trum. Before each experiment, the catalyst disk was heated in
10% O2/He at 773 K for 1 h, followed by cooling to the desired
temperature and purging for 30 min in He. Then a flow of var-
ious gas mixtures was fed at a rate of 100 cm3 min−1. Typical
feed gas compositions were the same as those in the reaction ex-
periments. The time-resolved change in the outlet concentration
of reactants or products was monitored by the mass quadru-
pole apparatus. To prevent changes in the operating pressure of
the IR cell, the excess flow from the cell was sent to vent. The
quantification of the product concentration was carried out by
calibrating the mass response using a calibrated mixture in He
of each component.
δ+
to Agn clusters, which should act as more efficient catalytic
sites than Ag+ cations [14,22–24]. There are two different opin-
ions concerning the role of Agn clusters: (1) Hydrogen reduces
Ag+ to Agn clusters, which activate oxygen, as speculated by
Richter et al. [15], and (2) Agn clusters formed during H2–HC-
SCR do not represent the unique active sites, but hydrogen itself
functions in enhancement of HC-SCR [18]. The latter opinion
was supported by the finding that CO cofeeding in HC-SCR
caused Agn cluster formation but no positive effect on de-NOx
activity [24].
Although the effect of hydrogen addition on the nature of
the surface adsorbed species has been well investigated by cat-
alytic and in situ infrared (IR) studies performed under transient
conditions and under simple model reaction conditions, such
as NO + O2 and HC + O2 reactions, the mechanistic cause of
hydrogen effect needs to be discussed, taking into account the
steady-state HC-SCR kinetics, though such an investigation has
not yet been attempted. In addition, information on the hydro-
gen effect under unsteady-state condition is also useful, because
the automotive catalyst should work even under unsteady-state
conditions such as A/F (air–fuel ratio) switching conditions. In
this paper, we report kinetic and in situ IR studies of C3H8-SCR
on Ag/Al2O3 conducted to provide experimental evidence of
the mechanistic causes of the hydrogen effect. We demonstrate
that hydrogen enhances steady-state C3H8-SCR by promoting
the activation of oxygen and by retarding nitrate poisoning. We
also investigate N2 formation via the reaction of nitrates with
H2 and NH3 formation via a NO + H2 reaction to demonstrate
the additional role of hydrogen in reduction under unsteady-
state conditions.
3. Results and discussion
3.1. Effect of hydrogen on O2 activation
Fig. 1 shows an Arrhenius plot for C3H8-SCR over Ag/
Al2O3. In the absence of H2, the rates of NO reduction to N2
and C3H8 oxidation by NOx and O2 to COx gave fairly good
straight lines (in a temperature range of 623–698 K), and the
apparent activation energies were 224 and 229 kJ mol−1, re-
spectively (Table 1). The apparent activation energies decreased
from the addition of 0.5% H2 in the reaction gas mixture; the
apparent activation energies for NO reduction and C3H8 oxi-
dation were 61 and 63 kJ mol−1, respectively (in a tempera-
ture range of 448–673 K). This indicates that the addition of
hydrogen results in decreased activation energy of the rate-
determining step or a change in the rate-determining step. The
deuterium kinetic isotope effect was examined in a temperature
range of 448–673 K (Table 1). The apparent activation energies
2. Experimental
in the D2–C3H8-SCR reaction for NO reduction (60 kJ mol−1
)
The Ag/Al2O3 catalyst (Ag, 2 wt%; surface area, 251 m2
g
−1) was prepared by impregnating γ -AlOOH with an aque-
ous solution of silver nitrate, followed by evaporation to dry-
ness at 393 K and calcination in air at 873 K for 4 h. The
catalytic test was performed in a fixed-bed flow reactor with
0.01–0.1 g of catalyst at a flow rate of 100 cm3 min−1. Typical
compositions of feed gas in the absence and the in the pres-
ence of H2 were NO/C3H8/O2/He = 0.1%/0.2%/10%/balance
and NO/C3H8/O2/H2/He = 0.1%/0.1%/10%/0.5%/balance, re-
spectively. Kinetic studies (e.g., determination of reaction order
and apparent activation energy) were made under the conditions
of NO and C3H8 conversions <30% by varying the amount of
catalyst. The effluent gas was analysed by gas chromatography
(GC) and NOx analysis (Best BCL-100uH). The reaction rates
of NO and C3H8 were calculated using the amounts of N2 and
COx (CO + CO2) determined by GC.
Fig. 1. Arrhenius plot for the reactions of (!, ") H –C H -SCR and (P, Q)
2
3 8
In situ IR spectra were recorded on a JASCO FT/IR-620
equipped with a quartz IR cell connected to a conventional flow
C H -SCR on Ag/Al O : (open symbols) rates of NO reduction to N ; (closed
3
8
2
3
2
symbols) rates of C H oxidation by NO and O to CO .
x
x
3
8
2

404
K.-i. Shimizu et al. / Journal of Catalysis 239 (2006) 402–409
Table 1
Kinetic parameters for selective reduction of NO over Ag/Al O
2
3
C H -SCR
H –C H -SCR
3
8
2
3 8
NO
C H
3
NO
C H
3 8
8
a
−1
E
(kJ mol
)
224 5.6
229 0.5
61 2.5
(60 0.9)
0.54 0.02
0.76 0.01
0.01 0.008 0.49 0.01
0.49 0.01
−0.45 0.01 −0.69 0.01
63 0.4
(63 0.3)
0.74 0.02
0.72 0.02
a
c
c
b
Order in H
Order in C H
3
Order in O
Order in NO
–
–
2
b
1.91 0.19 1.79 0.32
1.38 0.12 1.85 0.17
−2.53 0.1 −2.83 0.02
8
b
2
b
d
d
e
e
f
0.88 0.02
f
a
Estimated from the result in Fig. 1.
Estimated from the result in Figs. 2, 3, and 5.
NO (0.1%)/C H (0.1%)/O (10%)/D (0.5%) reaction in a temperature
b
c
3
8
2
2
range of 448–673 K.
d
NO concentration range is 200–1000 ppm.
NO concentration range is 130–550 ppm.
NO concentration range is 550–1000 ppm.
e
f
Fig. 3. Effect of O concentration on the rates of (!) NO reduction to N and
2
2
(") C H oxidation to CO for (a) C H -SCR at 673 K and (b) H –C H -SCR
x
3
8
3
8
2
3 8
at 573 K on Ag/Al O .
2
3
O2 concentration. The empirical reaction orders for NO and
C3H8 conversions with respect to O2 in the range of 1.5–10%
O2 were 0.01 and 0.49, respectively. Taking into account the
results in Fig. 1 showing that the addition of hydrogen leads to
decreased activation energy of the rate-determining step or to a
change in the rate-determining step, the results shown in Fig. 3
can be understood based on the assumptions that the conver-
sion of molecular oxygen into reactive oxygen species is the
rate determining-step in the absence of hydrogen, and that the
activation energy for this step is decreased by the addition of
hydrogen. The zero-order rate dependence on O2 concentration
(Fig. 3b) and the absence of deuterium kinetic isotope effect
(Table 1) indicate that the O2 activation by hydrogen is rela-
tively rapid and is not involved in the rate-determining step in
the H2–C3H8-SCR reaction.
Fig. 4 compares the effect of C3H8 concentration on reac-
tion rates in the absence and presence of H2. In the absence of
H2, the rates of NO reduction and C3H8 oxidation depended
strongly on C3H8 concentration (Fig. 4a), and these rates in-
creased sharply at higher concentrations. The reaction orders
for NO reduction and C3H8 oxidation with respect to C3H8 in
the range of 200–2100 ppm C3H8 were 1.91 and 1.79, respec-
tively (Table 1), suggesting that involvement of C3H8 activation
in the important step in NO reduction. The dependence of the
reaction rates on C3H8 concentration was significantly different
in the presence of hydrogen (Fig. 4b) than in the absence of hy-
drogen, and the empirical reaction orders with respect to C3H8
for the NO reduction (0.76) and C3H8 oxidation (0.72) were
Fig. 2. Effect H concentration on the rates of (!) NO reduction to N and
2
2
(") C H oxidation to CO for H –C H -SCR over Ag/Al O at 573 K.
x
3
8
2
3
8
2 3
and C3H8 oxidation (63 kJ mol−1) were close to those in the
H2–C3H8-SCR reaction, indicating that hydrogen was not in-
volved in the rate-determining step of the H2–C3H8-SCR re-
action. Fig. 2 shows the effect of H2 concentration on the NO
reduction and C3H8 oxidation rates for C3H8-SCR. The rates
increased with increasing H2 concentration. As given in Ta-
ble 1, the empirical reaction orders for the NO reduction and
C3H8 oxidation (by NOx and O2) with respect to H2 were 0.54
and 0.74, respectively.
Fig. 3 compares the effect of oxygen concentration on the
reaction rates in the absence and presence of H2. In the absence
of H2 (Fig. 3a), the rates of NO reduction and C3H8 oxidation
were very low in regions of low O2 concentration. At higher O2
concentrations, the rates increased sharply with O2 concentra-
tion. The empirical reaction orders estimated from the kinetic
studies are listed in Table 1. The reaction orders for NO re-
duction and C3H8 oxidation with respect to O2 in the range of
1.5–10% O2 were 1.38 and 1.85, respectively, suggesting that
O2 activation was involved in several important steps of the NO
reduction with C3H8. As shown in Fig. 3b, the dependence of
the reaction rates on O2 concentration in the presence of hydro-
gen differs significantly from that in the absence of hydrogen.
The rate of NO reduction increased sharply from the addition
of 1.5% O2 and remained almost unchanged with increasing

K.-i. Shimizu et al. / Journal of Catalysis 239 (2006) 402–409
405
Fig. 5. (!, P) NO conversion to N and (", Q) C H conversion to CO for
x
2
3 8
C H -SCR on Ag/Al O at 573 K. NO concentration is (!, ") 200 ppm or
3
8
2 3
(P, Q) 1000 ppm.
lite are reduced by hydrogen to generate metallic or partially
reduced clusters [20–22,33–35]. We reported Ag cluster forma-
tion over Ag/Al2O3, with in situ UV-vis results demonstrating
that Ag+ ions as the main Ag species on as-calcined Ag/Al2O3
were rapidly converted in a flow of O2 (10%) + H2 (0.5%)
mixture to partially reduced Agnδ+ clusters [14]. Spectroscopic
evid+ence on the interaction of hydrogen with partially reduced
Ag3 cluster in Ag–A zeolite have been well investigated by
1
Baba et al. using H-MAS NMR [35]. They reported that hy-
drogen dissociates over Ag3 sites to generate acidic proton
Fig. 4. Effect of C H concentration on the rates of (!) NO reduction to N and
3
8
2
+
(") C H oxidation to CO for (a) C H -SCR at 673 K and (b) H –C H -SCR
x
3
8
3
8
2
3 8
and Ag3–H. Based on this observation, it is reasonable to as-
sume that the addition of hydrogen results in the formation of
hydride on silver clusters, which may react with oxygen to form
a reactive oxidant, such as hydroperoxy radical (HO2), perox-
ide (O22−), or superoxide ions (O2−), as speculated by Sazma
et al. [18]. The hypothetical role of hydrogen in the formation
of reactive oxygen species is currently under investigation and
will be reported in the near future.
at 573 K on Ag/Al O .
2
3
lower than those in the absence of hydrogen. This indicates
that the hydrogen addition promotes the C3H8-SCR reaction
through the activation of reductant, C3H8. Previously we in-
vestigated the hydrogen effect on HC-SCR over Ag/Al2O3 by
in situ IR spectroscopy and showed that the addition of H2 in-
creased the formation rate of hydrocarbon-derived oxygenate
(acetate) by a factor of 190 [13]. Summarizing the above ex-
perimental results, one of hydrogen’s most important roles in
the promotion of C3H8-SCR can be described as follows. In
the absence of hydrogen, activation of molecular oxygen on the
surface into reactive oxygen species and subsequent oxidative
C3H8 activation to surface oxygenates occur slowly, and either
of these steps can be the rate-determining step in C3H8-SCR.
The addition of hydrogen decreases the activation energy for
the reaction of molecular oxygen into reactive oxygen species
involved in the oxidative activation of C3H8.
3.2. Poisoning effect of nitrates
Fig. 5 shows the effect of NO concentration on the conver-
sion of NO and C3H8 for C3H8-SCR over Ag/Al2O3 in the ab-
sence of hydrogen. Clearly, the activity increases with decreas-
ing NO concentration; the temperature at which NO conversion
starts is about 100 K lower for 200-ppm NO than 1000-ppm
NO. Note that this feature is suitable for the practical diesel de-
NOx system, in which very small concentration of NOx must
be reduced.
The promotional effect of hydrogen in redox catalysis has
also been reported from studies on the partial oxidation of hy-
drocarbons, including CH4 [27–29], ethane [30], C3H6 [31],
and benzene [32]. Wang et al. [28,30] reported that the addi-
tion of hydrogen enhanced the selective oxidation of methane
or ethane by oxygen over an iron phosphate catalyst. Observ-
ing an IR band due to peroxide (O22−) on the catalyst from
adding hydrogen to the reaction mixture, these authors pro-
posed that hydrogen, as the electron donor as well as the proton
donor, reductively activates molecular oxygen [28,30]. On the
other hand, it is well known that isolated Ag+ ions in zeo-
The reaction rates of NO and C3H8 for C3H8-SCR over
Ag/Al2O3 at 623 K were measured as a function of NO con-
centration (Fig. 6a). As the NO concentration decreased from
1000 ppm to 200 ppm, the reaction rates of NO and C3H8 in-
creased by a factor of 101–102. Above 200 ppm NO, empirical
reaction orders for NO reduction and C3H8 oxidation with re-
spect to NO showed negative values (between −2.5 and −2.8),
suggesting that certain NOx adspecies inhibit the reduction of
NO by C3H8. Fig. 7 shows IR spectra of adsorbed species
during C3H8-SCR over Ag/Al2O3 at steady state. During the
reaction at 573 K, strong bands due to unidentate (1554 and

406
K.-i. Shimizu et al. / Journal of Catalysis 239 (2006) 402–409
Fig. 8. IR spectra of adsorbed species during C H -SCR on Ag/Al O in
3
8
2 3
steady states at 573 K.
tivation of the reactants, resulting in decreased activity of the
Ag/Al2O3 catalyst for C3H8-SCR. The IR results in Fig. 7 show
that decreasing NO concentration leads to increasing intensity
of the bands due to acetate (1460 cm−1) [13,25,26]. This in-
dicates that the nitrates inhibit the activation of hydrocarbon
molecule. Nitrate coverage also strongly depends on tempera-
ture. Increasing reaction temperature lead to decreased intensity
of the bands due to nitrates and increased intensity of the bands
due to acetate (1460 cm−1). As shown in Fig. 5, NO conver-
sions increased sharply above 673 K under the condition of
1000 ppm NO. This can be understood, taking into account
the result in Fig. 7, as follows. As the nitrate coverage de-
creases with temperature, the number of surface sites available
for C3H8 activation increases, and thus SCR activity increases.
From these results, it can be concluded that the nitrates, espe-
cially the bidentate nitrate, deactivate the Ag/Al2O3 catalyst for
C3H8-SCR by strongly adsorbing on the reaction site and con-
sequently inhibiting the adsorption and subsequent activation of
C3H8.
Fig. 6. Effect of NO concentration on the rates of (!) NO reduction to N and
2
(") C H oxidation to CO for (a) C H -SCR at 623 K and (b) H –C H -
x
3
8
3
8
2
3 8
SCR at 573 K on Ag/Al O .
2
3
3.3. Effect of hydrogen on nitrates removal
Fig. 6b plots the reaction rates for C3H8-SCR in the pres-
ence of hydrogen as a function of NO concentration. The NO
reduction and C3H8 oxidation rates increased with increasing
NO concentration up to 550 ppm, then decreased at higher NO
concentrations. Below 550 ppm NO, empirical reaction orders
for NO reduction and C3H8 oxidation with respect to NO were
0.49 and 0.88, respectively, whereas those at higher NO con-
centration (>550 ppm) exhibited negative values (−0.45 and
−0.69). These results suggest that certain NOx species were
involved in the SCR reaction and that they inhibited the reac-
tion at high NO concentration. The reaction orders with respect
to NO were increased by the addition of hydrogen (Table 1),
indicating that hydrogen retarded the poisoning effect of NOx
adspecies. As shown in Fig. 8, in situ IR results indicate that the
addition of hydrogen to the C3H8-SCR reaction caused a signif-
icant decrease in the band intensity of bidentate nitrate (1580,
Fig. 7. IR spectra of adsorbed species during C H -SCR on Ag/Al O in the
steady states.
3
8
2 3
1292 cm−1) and bidentate (1580 and 1246 cm−1) nitrates [13,
25,26,36–38] were observed. The nitrate coverage depended
strongly on the NO concentration in the gas phase; as the NO
concentration decreased from 1000 to 200 ppm, the intensity of
the bands due to nitrates (especially the bidentate nitrate) de-
creased. Taking into account the result that the reaction rates
for NO reduction and C3H8 oxidation increased with decreas-
ing NO concentration, it is evident that the nitrates adsorbed
on the reaction site inhibited the adsorption and subsequent ac-

K.-i. Shimizu et al. / Journal of Catalysis 239 (2006) 402–409
407
Fig. 9. Changes in the surface concentration of nitrates estimated using the
−1
integrated areas of IR band in a range 1190–1338 cm as a function of time
of exposure of Ag/Al O to (!) H (0.5%)/He, (") H (0.5%)/O (10%)/He
or (P) He at 523 K. Before the measurement, the catalyst was pre-exposed to
Fig. 10. Changes in (!) the rate of nitrates consumption and mass intensities for
(—) NO, (–·–) N , (· · · · ·) H O and (- - -) H as a function of time of exposure
2
3
2
2
2
2
2
2
NO (0.1%)/O (10%)/H (0.5%)/He for 0.5 h, followed by purging with He for
of Ag/Al O to H (0.5%)/He at 523 K. Before the measurement, the catalyst
2
2
2
3
2
0.5 h.
was pre-exposed to NO (0.1%)/O (10%)/H (0.5%)/He for 0.5 h, followed by
2
2
purging with He for 0.5 h.
1246 cm−1), a relatively small decrease in the band intensity
of unidentate nitrate (1554, 1292 cm−1), and an increase in the
band intensity of acetate. The coverage of nitrates during the
H2–C3H8-SCR reaction depended on the NO concentration in
the gas phase. As the NO concentration decreased from 1000 to
200 ppm, the band intensity of nitrates, especially the bidentate
nitrate, decreased. From these results, it can be concluded that
the addition of hydrogen is effective in decreasing nitrate cov-
erage on the surface and thus retarding the poisoning effect by
nitrates, and that the bidentate nitrate is mainly responsible for
the poisoning effect.
The dynamics of nitrate removal by hydrogen were evalu-
ated based on the transient response of the IR spectra at 523 K
(Fig. 9). The concentration of nitrates was estimated from the
intensity of the bands in a range of 1180–1335 cm−1 due
to nitrates (including unidentate and bidentate nitrates) and the
extinction coefficient of nitrates (1.3 × 10−17 cm−1 cm2 mole-
cule−1) [25]. The catalyst was first exposed to a flow of NO +
O2 + H2 for 0.5 h to produce nitrates on the catalyst surface,
followed by purging with He for 0.5 h to remove NOx, O2,
and H2 in gas-phase and weakly adsorbed species. The ad-
sorbed nitrates were stable in a flow of He. At this time, the
nitrate concentration was about 0.21 mmol g−1, close to the Ag
content of the sample (Ag = 0.185 mmol g−1), suggesting that
nitrate is adsorbed mainly on Ag+ sites. The same conclusion
was recently reached by Brosius et al. [17] for nitrate adsorp-
tion on an Ag/Al2O3 catalyst. The catalyst was exposed to a
flow of H2 + O2 or H2 while IR spectra were recorded as a
function of time. The bands due to nitrates (in a range 1190–
1338 cm−1) decreased rapidly on exposure to H2. The rate of
nitrate consumption, d[NO3−]/dt, during the transient reaction
of nitrates with hydrogen was determined by numerically dif-
ferentiating the data in Fig. 9 and are plotted in Fig. 10. Mass
spectroscopy directly connected to the in situ IR flow cell was
used to provide information on the gas-phase products in the
reaction of adsorbed nitrates with hydrogen. Fig. 10 includes
the time dependence of mass intensities of gaseous products.
Exposing the pretreated catalyst to H2 produced NO and N2.
A similar experiment using a NOx analyzer for the quantifica-
tion of released NOx produced a NO desorption curve similar
to that determined by mass spectroscopy shown in Fig. 10 (not
shown). Note that N2O (m/e = 44) formation was negligible
and that no NO2 (m/e = 46) formation was confirmed by mass
spectroscopy or NOx analysis. These results indicate that the ni-
trates adsorbed on the catalyst surface were reduced to mostly
NO and N2 through the reaction with H2 as a reductant. The
amounts of NO and N2 estimated from the peak area in Fig. 10
were 0.13 and 0.056 mmol g−1, respectively. The amount of ni-
trates consumed for 1000 s (0.20 mmol g−1) was fairly close to
that of the product, NO + 2N2 (0.24 mmol g−1). These results
quantitatively confirm that most of the nitrates on the surface
of Ag/Al2O3 were reduced by H2 to NO and N2. As Burch
assumed, nitrite might be produced on the surface by the re-
duction of nitrates. However, no band assignable to adsorbed
NO species, nitro or nitrite, was observed during this transient
reaction (results not shown). We speculate that such NO ad-
species may be formed but quickly desorb as NO or react with
hydrogen.
3.4. NH3 formation by NO + H2 reaction
Fig. 11 shows IR spectra of adsorbed species during the
NO + H2 reaction over Ag/Al2O3. During the reaction at 373–
673 K, bands at around 2900 and 3100 cm−1, assignable to
+
NH stretching vibration of adsorbed NH3 (NH4 or coordi-
nated NH3) [39,40], were observed, with the highest inten-
sity at 473 K. In t+he NH deformation vibration region, bands
assignable to NH4 (1400, 1465, and 1690 cm−1) [39,40] and
those assignable to coordinated NH3 (1215–1292, 1620 cm−1
)
[39,40] were observed. As shown in Fig. 11 (spectra g and h),
these bands were also observed when Ag/Al2O3 was exposed
to a flow of NH3 (0.1%)/He, confirming the foregoing band
assignments. Bentrup et al. recently reported IR spectrum of
adsorbed species on Ag/Al2O3 in a flow of NO2 + H2 at 573 K.
The spectrum exhibited bands at 1396, 1463, and 1689 cm−1
,

408
K.-i. Shimizu et al. / Journal of Catalysis 239 (2006) 402–409
Fig. 11. IR spectra of adsorbed species in flowing (a) NO (0.4%) at 673 K, in flowing NO (0.4%)/H (2%) at (b) 773 K, (c) 673 K, (d) 573 K, (e) 473 K, (f) 373 K
2
and in flowing NH (0.1%) at (g) 773 K and (h) 473 K on Ag/Al O after flowing each gas mixture for 0.5 h.
3
2 3
which these authors assigned to nitro species, monodentate ni-
tro, and adsorbed NO2, respectively. Although the position of
+
these bands are close to those of the NH4 bands (1400, 1465,
1690 cm−1) observed on Ag/Al2O3 during the NO + H2 re-
action (spectra d and e) or in flowing NH3 (spectrum g), their
assignments are different from ours. If these bands were as-
signed to nitro or NO2 species, then the bands should not shift
even in a flow of NO + D2. But no bands were observed in
the NH deformation vibration region (1200–1700 cm−1) dur-
ing the NO (0.4%) + D2 (2%) reaction over Ag/Al2O3 at 473
and 773 K, demonstrating the low energy shift of all bands (re-
sult not shown). These results indicate that bands in the NH
deformation vibration region (1200–1700 cm−1) observed dur-
ing NO + H2 reaction (spectra b–g) are not assigned to nitro
+
Fig. 12. Changes in the band intensities due to (!) NH
−1
(in a range 1426–
4
+
−1
or NO2 species. Above 473 K, the band intensity for NH4
1520 cm ), (P) coordinated NH (in a range 1190–1338 cm ) and (") the
3
decreased and that for coordinated NH3 increased with temper-
ature.
mass intensity of outlet NH as a function of time of exposure of Ag/Al O to
3
2 3
the NO (0.4%)/H (2%) mixture at 473 K.
2
Fig. 12 shows the time course of the integrated intensity
+
of the bands due to NH4 and coordinated NH3 during the
N2 formation by the surface reaction of nitrates with H2 can
be explained as follows: Adsorbed NH3 (NH4 or coordinated
NO+H2 reaction at 473 K. Formation of gas-phase NH3 during
the above reaction was also observed with mass spectroscopy
(Fig. 12). When the flowing gas was switched to He, the band
+
NH3) are produced by NO + H2 reaction on Ag/Al2O3 and the
NH3 species react with nitrate (adsorbed NO2) to produce N2.
Although this mechanism could not be an important NO reduc-
tion pathway in the steady-state HC-SCR system, it may be the
main pathway under oxygen-poor conditions, such as low A/F
condition in the practical automotive system.
+
intensity for NH4 and the mass intensity of gas-phase NH3
decreased, whereas the band intensity for coordinated NH3 in-
creased. These results confirm NH3 formation via the NO + H2
reaction+on Ag/Al2O3. It is established that adsorbed NH3 such
as NH4 can react with NO2 to produce N2 [40]. Therefore,

K.-i. Shimizu et al. / Journal of Catalysis 239 (2006) 402–409
409
4. Conclusion
[10] S. Satokawa, K. Yamaseki, H. Uchida, Appl. Catal. B 34 (2001) 299.
[11] S. Satokawa, Chem. Lett. (2000) 294.
[12] S. Satokawa, J. Shibata, K. Shimizu, A. Satsuma, T. Hattori, Appl. Catal.
B 42 (2003) 179.
[13] J. Shibata, K. Shimizu, S. Satokawa, A. Satsuma, T. Hattori, Phys. Chem.
Chem. Phys. 5 (2003) 2154.
[14] A. Satsuma, J. Shibata, A. Wada, Y. Shinozaki, T. Hattori, Stud. Surf. Sci.
Catal. 145 (2002) 235.
[15] M. Richter, U. Bentrup, E. Eckelt, M. Schneider, M.-M. Pohl, R. Fricke,
Appl. Catal. B 51 (2004) 261.
[16] R. Burch, J.P. Breen, C.J. Hill, B. Krutzsch, K. Konrad, E. Jobson, L. Ci-
der, K. Eränen, F. Klingstedt, L.-E. Lindfors, Top. Catal. 30/31 (2004)
19.
[17] R. Brosius, K. Arve, M.H. Groothaert, J.A. Martens, J. Catal. 231 (2005)
344.
[18] P. Sazama, L. Capek, H. Drobna, Z. Sobalik, J. Dedecek, K. Arve,
B. Wichterlova, J. Catal. 232 (2005) 302.
[19] M. Richter, R. Fricke, R. Eckelt, Catal. Lett. 94 (2004) 115.
The mechanistic cause of the hydrogen promotion effect on
the activity of Ag/Al2O3 for C3H8-SCR was evaluated based on
kinetic and in situ IR studies. Our findings can be summarized
as follows. In the absence of hydrogen, the activation of mole-
cular oxygen into reactive oxygen species and subsequent ox-
idative C3H8 activation to surface oxygenates occur slowly, and
either of these steps can be the rate-determining step of C3H8-
SCR. In the H2–C3H8-SCR reaction, hydrogen is responsible
for the reductive activation of molecular oxygen into reactive
oxygen species involved in the oxidative activation of C3H8. In
the absence of hydrogen, nitrates show a significant poisoning
effect on C3H8-SCR. Hydrogen retards nitrate poisoning by re-
ducing nitrates and thereby decreasing nitrate coverage. In the
absence of gas-phase oxygen, N2 is produced via the reaction of
nitrates with hydrogen, probably through the formation of NH3
on the surface by the NO + H2 reaction. This may be an addi-
tional effect of hydrogen, which enhances NO reduction under
unsteady-state (low A/F) condition. It is well known that the
HC-SCR reaction mechanism depends strongly on the reaction
condition. Consequently, we believe that the proposed role of
hydrogen could explain the hydrogen promotion effect in cer-
tain reaction conditions.
[20] J. Shibata, K. Shimizu, Y. Takada, A. Shichi, H. Yoshida, S. Satokawa,
A. Satsuma, T. Hattori, J. Catal. 222 (2004) 368.
[21] J. Shibata, Y. Takada, A. Shichi, S. Satokawa, A. Satsuma, T. Hattori,
Appl. Catal. B 54 (2004) 137.
[22] J. Shibata, K. Shimizu, Y. Takada, A. Shichi, H. Yoshida, H. Yoshida,
S. Satokawa, A. Satsuma, T. Hattori, J. Catal. 227 (2004) 367.
[23] U. Bentrup, M. Richter, R. Fricke, Appl. Catal. B 55 (2005) 213.
[24] B. Wichterlova, P. Sazama, J.P. Breen, R. Burch, C.J. Hill, L. Capek,
Z. Sobalik, J. Catal. 235 (2005) 195.
[25] K. Shimizu, H. Kawabata, A. Satsuma, T. Hattori, J. Phys. Chem. B 103
(1999) 5240.
[26] K. Shimizu, H. Kawabata, H. Maeshima, A. Satsuma, T. Hattori, J. Phys.
Chem. B 104 (2000) 2885.
Acknowledgment
[27] Y. Wang, K. Otsuka, J. Catal. 155 (1995) 256.
This work was partly supported by a grant-in-aid from the
Ministry of Education, Science and Culture, Japan.
[28] Y. Wang, K. Otsuka, K. Ebitani, Catal. Lett. 35 (1995) 259.
[29] J.-S. Min, H. Ishige, M. Misono, N. Mizuno, J. Catal. 198 (2001) 116.
[30] Y. Wang, K. Otsuka, J. Catal. 171 (1997) 106.
[31] T. Hayashi, K. Tanaka, M. Haruta, J. Catal. 178 (1998) 566.
[32] W. Laufer, W.F. Hoelderich, Chem. Commun. (2002) 1684.
[33] M.D. Baker, G.A. Ozin, J. Godber, J. Phys. Chem. 89 (1985) 305.
[34] Y. Kim, K. Seff, J. Phys. Chem. 91 (1987) 668.
[35] T. Baba, N. Komatsu, H. Sawada, Y. Yamaguchi, T. Takahashi, H. Sugi-
sawa, Y. Ono, Langmuir 15 (1999) 7894.
[36] N.D. Parkyns, in: J.W. Hightower (Ed.), Proc. 5th Int. Congress on Catal-
ysis, Miami Beach, vol. 1, 1972, p. 255.
[37] F. Prinetto, G. Ghiotti, I. Nova, L. Lietti, E. Tronconi, P. Forzatti, J. Phys.
Chem. B 105 (2001) 12732.
References
[1] M. Iwamoto, Catal. Today 29 (1996) 29.
[2] H. Hamada, Catal. Today 22 (1994) 21.
[3] R. Burch, J.P. Breen, F.C. Meunier, Appl. Catal. B 39 (2002) 283.
[4] R. Burch, Catal. Rev. 46 (2004) 271.
[5] T. Miyadera, K. Yoshida, Chem. Lett. (1993) 1483.
[6] T. Nakatsuji, R. Yasukawa, K. Tabata, K. Ueda, M. Niwa, Appl. Catal.
B 17 (1998) 333.
[7] K. Shimizu, A. Satsuma, T. Hattori, Appl. Catal. B 25 (2000) 239.
[8] K. Shimizu, J. Shibata, H. Yoshida, A. Satsuma, T. Hattori, Appl. Catal.
B 30 (2001) 151.
[9] K. Shimizu, J. Shibata, A. Satsuma, T. Hattori, Phys. Chem. Chem. Phys. 3
(2001) 880.
[38] A.L. Goodman, E.T. Bernard, V.H. Grassian, J. Phys. Chem. A 105 (2001)
6443.
[39] H. Knozinger, Adv. Catal. 25 (1976) 184.
[40] J. Eng, C.H. Bartholomew, J. Catal. 171 (1997) 27.