Kinetics and mechanism of steady-state catalytic NO decomposition reactions on Cu-ZSM5

Article abstract of DOI:10.1006/jcat.2002.3622

A sequence of elementary steps for the formation of N₂O, NO₂, N₂, and O₂ during NO decomposition on Cu-ZSM5 was inferred from steady-state and transient rate data combined with previous spectroscopic evidence for adsorbed species and Cu structures. The adsorbed species and Cu structures involved in these elementary steps were consistent with previous spectroscopic evidence. The redox nature of the NO and N₂O decomposition pathways was confirmed by transient kinetic measurements, which showed the net incorporation of oxygen and a concomitant decrease in reaction rates as decomposition proceeded on Cu-ZSM5 samples containing reduced Cu⁺ dimers. The low activation energy for this step, the significant heat of adsorption of NO on Cu⁺ dimers, and the transition in prevalent Cu structures from {Cu²⁺-O^2--Cu²⁺}²⁺ to {Cu⁺-□-Cu⁺}²⁺ led to the observed decrease in NO decomposition rates > 750 K. Rate equations for NO decomposition and for product formation and the temperature dependence of their rate parameters were consistent with the published kinetic data and with the effects of temperature on the relative abundance of adsorbed species and of Cu structures. These elementary steps suggested that redox steps restricted to Cu²⁺ and Cu⁺ cycles, the presence of vicinal Cu atoms to accommodate two adsorbed NO molecules and a two-electron reduction without the formation and agglomeration of Cu metal, and the balanced equilibrium between {Cu²⁺-O^2--Cu²⁺2+}²⁺ and {Cu⁺-□-Cu⁺}²⁺ species account for the high NO decomposition rates achieved on Cu-ZSM5 catalysts.

Full text of DOI:10.1006/jcat.2002.3622

Journal of Catalysis 209, 75–86 (2002)
doi:10.1006/jcat.2002.3622
Kinetics and Mechanism of Steady-State Catalytic NO Decomposition
Reactions on Cu–ZSM5
Bj o¨ rn Mod e´ n, Patrick Da Costa, Benjamin Fonf e´ , Deuk Ki Lee, and Enrique Iglesia2
1
Department of Chemical Engineering, University of California at Berkeley, Berkeley, California 94720
Received October 11, 2001; revised March 25, 2002; accepted March 28, 2002
acid rain. Catalytic reduction processes using NH3 or traces
A sequence of elementary steps for the formation of N2O, NO2, of residual hydrocarbons are currently used in order to
N2, and O2 during NO decomposition on Cu–ZSM5 was inferred reduce N O, NO, and NO concentrations in combustion-
2
2
from steady-state and transient rate data combined with previ-
ous spectroscopic evidence for adsorbed species and Cu structures.
Transient product evolution rates confirmed the redox nature of
this sequence and the role of Cu dimers with labile oxygen atoms.
N2O formed near ambient temperature as the initial decomposition
effluent streams (1, 2). Such processes become more chal-
lenging as lean-burn strategies are used to increase energy
efficiency. The catalytic decomposition of NO is favored by
thermodynamics at typical effluent conditions and this re-
action does not require a reductant (3). Reported catalytic
NO decomposition rates, however, remain too low for prac-
tical applications.
+
product after quasiequilibrated adsorption of NO on reduced Cu
dimers. The low activation energy for this step, the significant heat
+
+
of adsorption of NO on Cu dimers, and the transition in preva-
2
2−
2+ 2+
+
+
2+
Cu-exchanged pentasil zeolites (Cu–ZSM5) are among
lent Cu structures from {Cu –O –Cu
}
to {Cu –ꢀ–Cu }
lead to the observed decrease in NO decomposition rates above the most active NO decomposition catalysts reported (4–8).
2
+
2−
2+ 2+
7
50 K. Quasiequilibrium between O2 and {Cu –O –Cu
}
is Cu dimers bridging two exchange sites have been proposed
mediated by a set of steps involving NO2 formation and decompo- as active sites for NO decomposition (8–11, 16); the bridg-
sition, which lead to equilibrium NO2 concentrations during NO
ing oxygens in such dimers provide redox sites required for
catalytic turnovers. Liu and Robota (12) reported a paral-
decomposition at 650–850 K on Cu–ZSM5. These pathways are
2
+
faster than recombinative desorption steps requiring vicinal Cu
lel increase in NO decomposition rates and in the number
dimers with labile oxygen atoms to form O2. In these steps, NO
acts as a regenerable oxygen carrier that allows kinetic communi-
cation among remote adsorbed oxygen atoms via diffusion in the gas
phase. Rate equations for the decomposition of NO and for product
formation and the temperature dependence of their rate parameters
are consistent with the kinetic data reported and with the effects
+
of Cu species detected by X-ray absorption spectroscopy
during reaction at temperatures up to 773 K. These data
+
suggested that turnovers involve cycling between Cu and
2
+
+
Cu states and that NO adsorption occurs on Cu cations,
which oxidize by abstracting oxygen from adsorbed NO.
of temperature on the relative abundance of adsorbed species and Infrared spectroscopy has been used to probe interactions
+
2+
+
of Cu structures. These elementary steps suggest that redox steps between NO and Cu ions (Cu and Cu ) in ZSM5 (8, 9,
restricted to Cu2 and Cu cycles, the presence of vicinal Cu atoms 13–23). The detection of a band for Cu (NO) confirmed
+
+
+
to accommodate two adsorbed NO molecules and a two-electron
reduction without the formation and agglomeration of Cu metal,
the proposal that Cu adsorbs NO (17, 20, 21) as well as
+
2+
the kinetic relevance of Cu /Cu cycles (17). Various ad-
sorbed NOx species have been proposed as intermediates
leading to the formation of O2, among others a bridging
2
+
2−
2+ 2+
and the balanced equilibrium between {Cu –O –Cu
}
and
+
+
2+
{
Cu –ꢀ–Cu } species account for the high NO decomposition
rates achieved on Cu–ZSM5 catalysts.
ꢀc 2002 Elsevier Science (USA)
2+
−
2+
nitrate species {Cu –(NO )–Cu } (20).
3
Key Words: NO decomposition; kinetics; Cu–ZSM5; redox
mechanism; copper dimers; temperature effect.
+
2+
Cu dimers that cycle between Cu and Cu during each
NO decomposition turnover (8, 11) coexist with isolated
2
+
Cu sites anchored at two exchange sites and with Cu
oxides (Cu2O, CuO) (9, 24). In the experiments reported
in the current study, Cu–ZSM5 samples contain predomi-
INTRODUCTION
2
+
2−
2+ 2+
2+
Nitrogen oxides (NO and NO2) form in combustion pro- nately{Cu –O –Cu
}
andCu species, eachinteract-
cesses and lead to local high ozone concentrations and to ing with two exchange sites (10). No CuO structures were
evident from the kinetics and stoichiometry of their reduc-
tion in H2 or CO, which were used in order to probe the
speciation of Cu species in these samples (10).
NO decomposition rate equations have been reported on
several metal and metal oxide surfaces. NO concentration
1
Permanent Address: Division of Civil and Environmental Engineer-
2
+
ing, Kwangju University, Kwangju 503-703, Korea.
2
To whom correspondence should be addressed. Fax: (510) 642-4778.
E-mail: iglesia@cchem.berkeley.edu.
75
0
021-9517/02 $35.00
c 2002 Elsevier Science (USA)
All rights reserved.
ꢀ

´
MODEN ET AL.
7
6
orders are between 1 and 2 and reaction orders in O2 vary
between −1 and −0.5 (25–28). A first-order NO depen-
dence on Pt led to the proposal that the initial adsorption of
NO on Pt is the kinetically relevant step (26–28) and mobile
TABLE 1
Composition of Cu–ZSM5 Catalysts and N2 Formation Rates
Obtained for the Steady-State NO Decomposition Reaction
chemisorbed oxygen atoms were invoked in order to ex- Catalyst
Cu (wt%)
0.76
13.3
0.12
2.24
14.6
0.38
2.25
13.1
0.36
3.31
14.3
0.58
3.88
13.2
0.60
plain the −1 order in O2. NO decomposition rate studies on
Si/Al (atomic ratio)
Cu/Al (atomic ratio)
N2 Formation Ratea
Ba/MgO (29), La0.8Sr0.2CoO3 and La0.4Sr0.6Mn0.8Ni0.2O3
(
30), and Mn2O3 and Mn3O4 (31) have shown a second-
order NO dependence. On metal oxides, oxygen vacancies
are thought to provide the active sites for the adsorption
of NO, in the first step of redox NO decomposition cycles
−1 −1
µmol N2 (g catalyst)
s
0.067
0.56
1.66
4.71
1.41
3.98
3.35
6.42
4.08
6.64
3
−1 −1
1
0 mol N2 (mol Cu)
s
a
3
−1
Measured at 773 K, 0.05 g catalyst, 1.0 cm s flow rate, and 1 kPa
NO. Extrapolated to zero conversion (1 kPa NO, 0 kPa O2) using Eq. [13].
(
32).
Several NO decomposition rate studies on Cu–ZSM5
have been reported (4, 5, 33–37). Li and Hall proposed
a power-law equation with a first-order NO dependence
and a −0.5 order in O2 (34). These orders varied with
temperature; the O2 dependence became weaker and the
NO dependence increased slightly with increasing temper-
ature. NO decomposition rates increased with temperature
up to 773 K and then decreased at higher temperatures
was repeated in order to achieve higher Cu exchange levels.
The samples were then dried in ambient air at 383 K for 24 h,
pelleted, crushed, and sieved to retain agglomerates with
125- to 250-µm diameters. The Si, Al, and Cu concentra-
tions were measured by ion-coupled plasma emission spec-
troscopy (Galbraith Laboratories) (Table 1). All Cu–ZSM5
samples are labeled Cu(x), where x is the atomic Cu/Al
ratio. More-detailed characterization of these catalysts is
presented elsewhere (10). A Pt/Al2O3 catalyst (5 wt%;
Alfa–Aesar) was used as a comparison with Cu–ZSM5. The
titration of adsorbed oxygen by H2 at 373 K was used to
measure a Pt dispersion of 29%. All catalysts were treated
in He at 773 K for 2 h before NO decomposition kinetic
experiments.
(
4, 5, 34). Li and Hall (34) suggested that NO adsorption
becomes unfavorable as temperature increases, leading to
a decrease in the concentration of these intermediates at
higher temperatures. In contrast, NO decomposition rates
on Pt catalysts showed a normal Arrhenius dependence and
increased monotonically with increasing temperature (26;
see also Fig. 9).
This study aims to combine recent spectroscopic findings
about the nature of reactive intermediates with transient ki-
netic measurements (11) into a mechanism consistent with
measured rates for NO decomposition and for NO2, N2O,
O2, and N2 formation. The elementary steps proposed in-
clude specific pathways for the removal of adsorbed oxy-
gen; these steps avoid the previously proposed recombina-
tive desorption steps, which are difficult to reconcile with
the isolated nature of oxygen atoms in Cu dimers. Finally,
the steps proposed incorporate the formation of N2O and
NO2 into an overall mechanism of NO reactions in the pres-
ence of O2. This set of elementary steps is consistent with
steady-state rate data and with available spectroscopic and
transient data reported elsewhere.
Steady-State NO Decomposition Rate Measurements
NO, He, O2, NO2, and a calibration gas mixture were
used in NO decomposition kinetic measurements. The NO
mixture was supplied by Matheson as 1.00% NO, 9.98%
Ar, and 89.02% He (<10 ppm other gases). Helium (UHP
grade, Bay Airgas) was used for catalyst pretreatment and
for dilution of the 1.00% NO mixture to lower NO concen-
trations. The O2 mixture contained 4.99% O2 and 95.01%
He (Praxair). The NO2 mixture contained 0.05% NO2, 10%
Ar, and 89.95% He (Praxair). The calibration gas contained
0.05% N2, 0.05% O2, 0.05% N2O, 10% Ar, and 89.85% He
(
Praxair).
The reactant and effluent streams were analyzed using an
EXPERIMENTAL SECTION
online mass spectrometer (MKS Instruments, Orion Com-
pact Residual Gas Analyzer) differentially pumped in or-
der to sample streams at ambient pressure without mass
Catalyst Synthesis and Characterization
Cu–ZSM5 samples were prepared by ion exchange of discrimination (Pfeiffer, vaccum system TCP015). The Ar
Cu2 ions into Na–ZSM5 zeolite (Zeochem, 10 g; two present in the NO reactant mixture was used as an inter-
batches with Si/Al = 14.7 and 13.3) using 0.01–0.1 M cop- nal standard in measuring the concentrations of NO, N2,
per acetate solutions (1 liter, Cu(CH3COO)2 H2O, Sigma N2O, O2, and NO2. Calibration factors for N2 (28 amu),
Aldrich, ACS reagent) in doubly deionized water. Ion ex- N2O (44 amu), and O2 (32 amu) relative to Ar (40 amu)
+
change was carried out for 20 h at ambient temperature. were obtained from ion yields measured using the calibra-
3
These samples were then rinsed six times using 300-cm
tionmixtures. NOandNO2 calibrationswereobtainedfrom
batches of doubly deionized water with filtering between NO (30 amu) and NO2 (46 amu) ion yields using the NO
rinsing steps. For some samples, this exchange procedure and NO2 reactant mixtures.

STEADY-STATE CATALYTIC NO DECOMPOSITION REACTIONS ON Cu–ZSM5
77
Steady-state NO decomposition rates were measured us-
ing a quartz microreactor (18-mm inner diameter) with the
catalystsamplessupportedonaporousquartzfrit. Thecata-
lyst bed contained 50-mg samples with a bed height of <5
particle diameters; this shallow bed allows reaction rate de-
scriptionsusingtheformalismapplicabletowell-mixed flow
reactors (38). All experiments were carried out at ambient
total pressure. Bed temperatures were measured using a
type K thermocouple affixed to the outer reactor surface.
The temperature was maintained using an electronic con-
troller (Watlow, Series 988) and a resistively heated furnace
(
National Element, Type FA 120, 360 W). Each of the gas
mixtures was metered using calibrated electronic mass flow
controllers (Porter Instruments, Model 201).
The overall NO molar decomposition rate is twice the
combined molar formation rates of N2O and N2. Through-
out this manuscript, we use the sum of the formation rates
of N2 and N2O as a measure of the rate of the reaction.
Steady-state N2 + N2O formation rates [mol (N2 + N2O)
−1
−1
(
mol Cu)
s ] were compared with mechanism-derived
FIG. 1. Comparison of N2 and N2O conversion with varying residence
rate expressions using least-squares methods (39, 40). The time, 673 K, 1 kPa NO, Cu(0.60). Pretreated in He at 773 K for 2 h prior
regression analysis minimized the sum of the square of to the first NO exposure.
the relative deviations for all data points. Rate data were
obtained over a range of NO pressures (0.2–1 kPa), O2
pressures (0–1 kPa), and contact times (10–60 ms) at three
reaction temperatures (723, 773, and 823 K).
not exclude an intermediate role of N2O at these higher
temperatures.
On Cu(0.36), the apparent NO reaction order is 1.1 at
6
73 K and 1.5 at 873 K (Fig. 3). These noninteger values and
Temperature-Programmed and Isothermal
Transient Experiments
their variation with temperature reflect a complex kinetic
behavior inappropriate for power-law analysis. The reac-
tion order in NO approaches a value of 2 as temperature
increases. The expected decrease in the surface coverage
In temperature-programmed experiments, samples were
treated in 50% O2/He at 773 K for 2 h. Then, the tem-
perature was decreased to 298 K and He was introduced in
order to remove weakly adsorbed O2. The temperature was
then increased in either He or 0.464% NO/He streams in
order to probe O2 formation pathways in the presence and
absence of NO. During steady-state NO decomposition ex-
periments, rapid isothermal transients were performed be-
tween NO-containing streams and He. The details of these
temperature-programmed and isothermal transient exper-
iments are reported elsewhere (11).
RESULTS AND DISCUSSION
NO conversion to N2O initially increased with contact
time, but ultimately decreased, as N2O decomposed fur-
ther to form N2 and O2 (Fig. 1). This behavior is typical
of sequential reactions and it indicates that N2O acts as an
intermediate in the decomposition of NO to N2 and O2 on
Cu–ZSM5. Thus, N2 formation proceeds via the sequence
NO → N2O → N2. Below 723 K, N2O is detected at all con-
tact times, but N2 becomes the only detectable nitrogen-
containing product at higher temperatures (Fig. 2). This im-
plies that N2O decomposition rate increases more strongly
FIG. 2. N2 formation and N2O formation rates vs temperature for
the catalyst Cu(0.58), with 1 kPa NO in the feed at contact time 0.031 s.
with temperature than its formation rate from NO; it does Pretreated in He at 773 K for 2 h prior to the first NO exposure.

´
MODEN ET AL.
7
8
amounts above 723 K, indicating the substantial absence of
N2O and NO2 and the excellent N and O balances in these
kinetic measurements.
The redox nature of NO decomposition reactions was
probed by switching between pure He and NO-containing
streams at a constant temperature; these experiments are
described in detail and for a wider range of materials else-
where (11). N2 and O2 formation rates during isothermal
transients from pure He to 0.464% NO in He are shown
in Fig. 5 for Cu(0.36) at 773 K, the temperature where the
NO decomposition rates are the highest. Initial N2 forma-
tion rates were very high, but they ultimately decreased
to steady-state rates significantly lower than the maximum
values reached during the early stages of the transient. O2
formation rates were initially much lower than N2 forma-
tionrates, butthenincreasedtotheirequimolarsteady-state
values. The nitrogen surplus in the effluent indicates the net
retention of some oxygen by Cu–ZSM5 during the initial
stages of the transient. The resulting increase in oxygen
surface concentration with time leads in turn to a smaller
+
+
number of the reduced {Cu –ꢀ–Cu } centers required for
NO adsorption and reaction. These data appear to confirm
that NO decomposition on Cu–ZSM5 proceeds via a redox
mechanism, in which N2 and N2O formation steps require
FIG. 3. Apparent order in NO (nNO) measured on Cu(0.36) for 673 K
ꢁ) and 873 K (ꢂ). Pretreated in He at 773 K for 2 h prior to the first NO
exposure.
(
+
+
reduced centers {Cu –ꢀ–Cu }.
Figure 6 shows net oxygen desorption rates from samples
pretreated in 50% O2/He at 773 K for 2 h and then heated
from ambient temperature to 1000 K using He or 0.464%
of all intermediates as temperature increases leads us to
suggest that this behavior reflects bimolecular reactions of
adsorbed NO on surfaces mostly free of adsorbed interme-
diates at high temperatures. The adsorbed NO molecules
involved in these bimolecular pathways are likely to reside
+
on a single site, such as two vicinal Cu cations in sufficient
proximity to form an oxygen bridge upon the decomposi-
tion of the two adsorbed NO and the desorption of N2O (8,
9
, 11).
NO2 was detected in concentrations near those predicted
by the thermodynamics for the O + 1/2O2 ꢀ NO2 reac-
tion on all Cu–ZSM5 catalysts at all reaction temperatures
(
Fig. 4). Thus, this reaction reaches quasiequilibrium within
the catalyst bed. The extent of reaction is unaffected by any
re-equilibration via homogeneous pathways within cooler
heated transfer lines and analytical equipment after the re-
actor. NO2 concentrations decreased sharply as tempera-
tureincreasedbecauseofunfavorablethermodynamics;the
equilibrium constant for this reaction decreases from 3.5 at
6
73 K, to 0.90 at 773 K, and to 0.32 at 873 K (3). The strong
temperature dependence of the thermodynamic proper-
ties accounts for the slightly higher than equilibrium NO2
concentrationsdetectedinsomecases. Inallcases, however,
the approach to equilibrium was independent of reactor
residence time. In the absence of added O2, NO2 was only
detected in substantial amounts below 723 K, and N2 and
O2 were the predominant NO decomposition products at
FIG. 4. Comparison of measured NO2 concentration for Cu(0.58)
and NO2 concentration calculated from the equilibrium constant for
1
NO + 2 O2 ꢀ NO2 (evaluated at each temperature) and the NO and O2
concentrations. ꢀ, 673 K, with 0.6 kPa NO and 0.215 kPa O2; ꢁ, 723 K,
with 0.6 kPa NO and 0.215 kPa O2; ꢃ, 773 K, with 0.6 kPa NO and 1.0 kPa
O2; ꢂ, 823 K, with 0.6 kPa NO and 1.0 kPa O2, with varying residence
higher temperatures. N2 and O2 were detected in equimolar time.

STEADY-STATE CATALYTIC NO DECOMPOSITION REACTIONS ON Cu–ZSM5
79
used (Fig. 6). This suggests the involvement of NO-derived
species or of gas-phase NO, and not just adsorbed oxy-
gen, in the steps required for removing the oxygen species
formed in NO decomposition as O2. The equilibrium NO2
concentrations achieved at all reaction conditions (Fig. 4)
require that the reaction of NO with adsorbed oxygen and
its reverse reaction—NO2 decomposition—proceed more
rapidly than NO decomposition. The equilibrated forma-
tion of NO2 guarantees the presence and the rapid nature
of a set of elementary steps that ensures kinetic commu-
nication among oxygens residing at distant dimer sites via
the formation, desorption, and gas-phase diffusion of NO2.
These distant oxygen atoms would otherwise require oxy-
gen atom migration for recombinative desorption steps to
occur at steady-state rates commensurate with the rate of
formation of these oxygen species during NO decomposi-
tion.
Infrared studies during NO decomposition have de-
tected a nitrate species bridging two Cu atoms as the most
FIG. 5. Isothermal transient to follow the NO decomposition on Cu– abundant nitrogen-containing species chemisorbed on Cu–
ZSM5, Cu(0.36), switch from He to 0.464 kPa NO/He at 773 K. Pretreated
in He at 773 K for 2 h prior to the first NO exposure.
ZSM5 at low temperatures (<673 K) (20–22). These species
have been proposed as intermediates in O2 formation dur-
ing NO decomposition; in temperature-programmed des-
NO/He as the carrier gas (11). Net oxygen desorption rates orption studies of preadsorbed NO, these nitrate species
in the presence of NO were measured from an oxygen atom decompose to NO and O2 (11, 41), thus re-forming the NO
balance in the effluent gas (Eq. [1]):
molecule used as the mobile oxygen shuttle that allows ef-
ficient kinetic communication among Cu dimer sites.
Several adsorbed species coexist during NO decomposi-
tion (8, 17, 20). The most abundant among them are shown
in Scheme 1.
rO,ads = rN,tot − rO,tot = 2rN + rN O − 2rO − rNO . [1]
2
2
2
2
Preadsorbedoxygendesorbsatlowertemperatureswhen
NO is present in the gas phase than when pure He is
∗
∗
∗
For some of these species, such as NO , NO , and O ,
2
3
2
structures other than those shown are possible for the
stoichiometry shown. We have assigned a formal 2-charge
to adsorbed NO2 and NO3 species in order to maintain
FIG. 6. Net oxygen desorption rate from Cu(0.58), for samples
treated in 50% O2/He at 773 K for 2 h. The temperature was ramped
−1
from 298 to 900 K at 0.167 K s using either He or 0.464 kPa NO/He as
the carrier gas.
SCHEME 1

´
MODEN ET AL.
8
0
TABLE 2
diffuses in the gas phase, and either decomposes back to
∗
NO and O at a reduced center (∗) or reacts with another
Mechanism for NO Decomposition on Cu–ZSM5
∗
∗
O at a remote dimer to form (NO ). These nitrate species
3
NO activation and N2O and N2 formation
∗
(
NO ) have been detected by infrared studies during NO
∗
3
1
2
3
4
5
.
.
.
.
.
∗ + NO
ꢁ
ꢄ NO
(quasiequilibrated)
(quasiequilibrated)
decomposition (20).
NO acts in steps 6 and 7 (Table 2) as a reducing agent,
in a manner reminiscent of the role of other reductants,
∗
∗
NO + NO
ꢁ
ꢄ (NO)
2
∗
∗
(NO) → O + N2O
2
∗
∗ + N2O
ꢁ
ꢄ N2O
(quasiequilibrated)
∗
∗
N2O → O + N2
such as CO and NH , in NO reduction reactions (1, 2).
3
O2 formation
In NO reduction by CO, the thermodynamic stability of
∗
∗
6
7
8
9
.
.
.
.
O + NO
ꢁ
ꢄ NO2
(quasiequilibrated)
(quasiequilibrated)
(quasiequilibrated)
(quasiequilibrated)
(quasiequilibrated)
the CO formed causes the oxygen atoms formed in NO
2
∗
NO
ꢁ
ꢄ ∗ + NO2
2
decomposition to leave the reaction zone as CO . In NO
2
∗
∗
O + NO2
ꢁ
ꢄ NO
3
∗
∗
O
2
∗
decomposition, however, the oxygen carrier (NO2) is un-
NO
ꢁ
ꢄ O + NO
3
2
stable, and it decomposes to NO and O within the reactor;
2
1
0.
ꢁ
ꢄ ∗ + O2
as a result, at least one NO molecule undergoes a reversible
∗
∗
Note. Steps 1–5 : 2 + 2NO → 2O + N2 (N2 formation).
oxidation–reduction cycle during each NO decomposition
∗
∗
Steps 6–10 : 2O
ꢁ
ꢄ 2 + O2 (O2 formation).
∗
turnover. The steps involved in the decomposition of NO3
Steps 1–10 : 2NO → N2 + O2 (overall reaction).
into ∗, NO, and O2 occur concurrently (steps 9 and 10),
because NO and O2 evolve simultaneously and in close to
equimolar amounts at 660 K during the decomposition of
preadsorbed NO on Cu–ZSM5 (11). During NO decompo-
charge neutrality for all exchanged Cu2 species, but the
validity of this assignment or the actual electron distribu-
tion between the adsorbed species and the Cu atoms do not
influence the kinetic treatment or the mechanistic conclu-
sions of this study, for which only stoichiometry, and not
molecular structure or charge, is relevant.
+
∗
sition, the O concentration is determined by the thermody-
∗
∗
namics for the overall 2O ꢀ O2 + 2 reaction, irrespective
of the mechanism by which this net reaction proceeds. As
a result, the direct recombinative desorption and the pro-
∗
posed pathways (steps 6–10, Table 2) lead to identical O
The results presented above led us to propose an NO de-
composition mechanism (Table 2) that is consistent with the
adsorbed species detected by infrared (8, 17, 20) and with
transient reduction and reaction studies (10, 11) on Cu–
ZSM5. The equilibrium between NO, O2, and NO2 guaran-
tees the quasiequilibrated nature of oxygen removal steps
via steps 6–10 (Table 2) during catalytic NO decomposition.
coverages, but the NO-mediated O2 formation pathways
are significantly faster and provide the kinetic route to this
equilibrium during NO decomposition on Cu–ZSM5.
NO Decomposition Kinetic Rate Equations
The sequence of elementary steps shown in Table 2 and
The proposed steps, many of them elementary in nature, are in Fig. 7 can be used to obtain a rate expression by using the
listed in Table 2 and shown as a catalytic cycle in Fig. 7.
pseudo-steady-state approximation for all adsorbed species
Steady-state rate measurements cannot discern the rela- and quasiequilibrium assumptions for some of the reaction
tive kinetic relevance of steps 2 and 3, because the resulting
rate equations (but not the meaning of the lumped appar-
ent rate constant) are identical when either step 2 or 3 is
quasiequilibrated. In both cases, NO concentration appears
2
as [NO] in the numerator of the rate expression. In the se-
quence shown in Table 2, the initial irreversible formation
of N2O determines overall NO decomposition rates. The
subsequent N2O decomposition steps influence N2 selec-
tivity, but not NO decomposition rates. Thus, the overall
NO molar decomposition rate is given by twice the com-
bined molar formation rates of N2O and N2.
NOactsintheproposedmechanismasanoxygentransfer
agent, which allows efficient kinetic communication among
distant oxygen atoms and the formation of O2 without re-
quiring surface or gas-phase diffusion of oxygen atoms at
distant Cu dimers. The equilibrated nature of the catalytic
oxidation of NO to NO2 using O2 ensures that this O2
formation pathway is not only available during NO de-
composition, but also much faster than NO decomposition
∗
turnovers. NO reacts with O to form NO2, which desorbs,
FIG. 7. Catalytic cycle for NO decomposition on Cu–ZSM5 catalysts.

STEADY-STATE CATALYTIC NO DECOMPOSITION REACTIONS ON Cu–ZSM5
81
steps. In what follows, equilibrium constants are denoted
as Ki and rate constants as ki . The quasiequilibrated step 1
∗
related [NO ] to [∗] and [NO] (Eq. [2]):
∗
[
NO ] = K1[NO][∗].
[2]
If step 2 is quasiequilibrated, after using Eq. [2] to eliminate
∗
[
NO ] we obtain
∗
∗
2
[
(NO) ] = K2[NO][NO ] = K1 K2[NO] [∗].
[3]
2
Then, the reaction rate (defined as the sum of the N2 and
∗
N2O formation rates) is determined by the rate of (NO)2
decomposition in step 3:
∗
2
r3 = k3[(NO) ] = k3 K1 K2[NO] [∗].
[4]
2
A site balance for all active sites [L] contains terms for [∗],
∗
∗
∗
∗
∗
2
[
O ], [NO ], [O ], [NO ], and [NO ]:
2 3
∗
FIG. 8. Parity plots for the rate expression with ∗ and O terms. The
∗
∗
∗
∗
∗
2
[
L] = [∗] + [NO ] + [O ] + [O ] + [NO ] + [NO ]. [5] rate data were obtained for a range of temperatures from 723 to 823 K for
2
3
Cu(0.58).
∗
∗
[
(NO) ] and [N2O ] were not detectable as a denomina-
2
tor term in any of the acceptable rate equations or as ad-
sorbed species in infrared studies during the reaction (20).
An overall equilibrium constant Kα is defined (Eq. [6]) for
quasiequilibrated steps 6–10. This sequence of steps yields
catalytic sites [L]:
2
k
3
K
1
K
2
[NO] [L]
r =
.
1/2
1
/2
+ K1−01[O
1
+ K
1
[NO]+ K
α
[O
2
]
2
]+ K
β
[NO][O
2
]+ K
6
K
α
[NO][O
2
]
∗
an expression of [O ] in terms of free sites [∗] and [O2]
[
12]
concentrations:
We note that reduced and oxidized Cu dimers are interme-
[6] diates in NO decomposition and that it would be inappro-
−1/2
Kα = (K6 K7 K8 K9 K10)
,
priate to describe one or the other as the catalytic site; they
∗
1/2
[
O ] = Kα[O2] [∗].
[7]
represent oxidation states of the Cu centers when the active
sites are “occupied” or “unoccupied.”
Steady-state NO decomposition rate data as a function of
NO (0.2–0.6 kPa) and O2 (0–1 kPa) at three temperatures
∗
∗
∗
Similarly, [O ], [NO ], and [NO ] can be obtained in terms
of [∗]
2
3
2
are compared with the rate equation obtained from the
elementary steps 1–10 (Eq. [12]) in Fig. 8. This comparison
illustrates the consistency among the proposed mechanism,
the expected most abundant surface species (masi (42)),
∗
−1
[
O ] = K [O2][∗],
[8]
2
10
∗
−1
∗
2
[
NO ] = K [NO][O ] = Kβ[NO][O2][∗]
3
9
ꢀ
ꢁ
Kβ = K9−
1
K
−1
, [9]
10
and the measured kinetic rate data. Rate equations with
second-order NO dependencies were required for accurate
descriptions of experimental NO decomposition rates at all
temperatures. At 723 K and higher temperatures only the
∗
1/2
[
NO ] = K6 Kα[NO][O2] [∗].
[10]
2
Equations [2] and [7]–[10] contain the concentrations of all
abundant species in terms of the concentration of reduced
dimers [∗], which can then be substituted into Eq. [5] to
obtain an expression for [∗] in terms of [L]:
1
/2
[
O2] term in the denominator was required, giving a rate
equation of the form (Eq. [13])
kapp[NO]2
r =
,
[13]
ꢀ
1/2
1/2
−1
1 + K [O ]
α
2
[
L] = [∗] 1 + K1[NO] + Kα[O2] + K1 [O2]
0
1
/2ꢁ
∗
consistent with O and ∗ as masi. Denominator terms corre-
+
Kβ[NO][O2] + K6 Kα[NO][O2]
.
[11]
∗
∗
∗
3
sponding to NO , NO , or NO concentrations were not re-
2
quired in order to describe reaction rates at 723 K or higher
This expression can be solved for [∗] and introduced into
temperatures. Infrared (20) and transient desorption (11)
Eq. [4] in order to obtain a rate equation in terms of re-
∗
actants and product concentrations and of the number of studies have shown that NO decomposes at ∼673 K and
3

´
MODEN ET AL.
8
2
∗
NO desorbs also at ∼673 K (11), in agreement with the (1–4 kPa vs 0.1–1 kPa). For a given NO conversion, the
∗
0.5
proposal of O and ∗ as masi species at temperatures above O2 concentration is higher and the [O2] denominator
00 K. The excellent agreement between predicted and ex- term becomes increasingly dominant as the NO pressure
perimental rates is illustrated in Fig. 8 for all kinetic data at increases. Thus, as NO pressures increase, the proportional
temperatures between 723 and 823 K. increase in O2 concentration leads to the partial cancel-
7
2
Below 723 K, the regression analysis is less conclusive lation of the [NO] numerator term. In our study, these
about the identity of the masi species. At 673 K, a rate changes in O2 concentrations were explicitly considered in
∗
∗
∗
equation containing denominator terms for NO , O , O , the well-mixed integral reactor analysis used for all kinetic
2
and ∗ gave the lowest value of the relative error, but nei- data. The present study also differs from the previous stud-
∗
∗
ther NO nor O have been detected by infrared studies ies in the mechanism proposed for O2 formation. In our
2
∗
during catalytic NO decomposition at these temperatures. study, the recombinative desorption of O is replaced by
∗
∗
A rate equation with NO , O , and ∗ denominator terms a sequence of steps that does not require atomic contact
3
∗
(
Eq. [14]) gave only a slightly larger error than the equation among O sites. This sequence involves the sequential for-
∗
∗
∗
∗
∗
containing also terms for NO , O , O , and ∗. We consider mation of NO , NO2, and NO ; its quasiequilibrium during
2
2
3
∗
∗
the presence of NO , O , and ∗ as masi a more realistic NO decomposition is assured by the NO/NO2/O2 equilib-
3
representation of the situation and of the rate data:
rium observed at all reaction conditions on Cu–ZSM5.
kapp[NO]2
+ Kα[O2]1/2 + Kβ[NO][O2]
Temperature Dependence of the Rate Constants
and the Equilibrium Constants in Rate Equations
r = 1
.
[14]
Steady-state NO decomposition rates on Cu–ZSM5 show
an unusual temperature dependence, characterized by a
normal Arrhenius behavior up to ∼773 K followed by a
decrease in rates at higher temperatures (4, 5, 34). These
ratechangesarereversibleduringtemperaturecycling;they
represent true kinetic effects and not deactivation or struc-
tural changes of active Cu centers at high temperatures.
Typical Arrhenius plots for NO decomposition rates are
shown in Fig. 9 at various NO and O2 pressures on Cu(0.58),
∗
This assessment is based on the detection of NO as the
3
most abundant infrared-active adsorbed species (20) and
on the lack of experimental or theoretical support for signif-
icant O concentrations during reaction or during oxidative
∗
2
treatments of these samples.
Some of the features in Table 2 resemble those proposed
for NO decomposition on Ba/MgO (29), La0.8Sr0.2CoO3
and La0.4Sr0.6Mn0.8Ni0.2O3 (30), and Mn2O3 and Mn3O4
(
31). These studies also found a second-order NO concen-
tration dependence, which was attributed to a quasiequili-
brated first NO adsorption step. On supported Pt and on
unsupported Ni, Fe, Co, and Zr catalysts, NO decompo-
sition rates were found to be first-order in NO (26–28).
The molecular adsorption of NO was proposed as the ki-
netically relevant step on Pt (26–28), even though molecu-
lar adsorption is typically fast and nonactivated. All these
studies reported inhibition of NO decomposition rates by
O2 through quasiequilibrated adsorption–desorption steps,
but the proposed pathways for oxygen desorption steps dif-
fer among these studies. On metal oxide catalysts, recom-
binative desorption of lattice oxygen atoms was proposed
(
29). In contrast with the similar proposal for Cu–ZSM5,
therearenoreasonableobjectionstothisproposalformetal
oxides, which contain a high density of vicinal lattice oxy-
gens and significant lattice oxygen mobility. The first-order
dependence of O2 in the denominator, found on Pt (26–28)
0
.5
and on perovskite-type catalysts (30), differ from the [O2]
dependence observed on Cu–ZSM5 catalysts and it cannot
be reconciled with quasiequilibrated dissociative oxygen
adsorption–desorption steps.
Li and Hall reported that NO decomposition rates were
1.0
−0.5
proportional to [NO] and [O2]
on Cu–ZSM5 (34).
This NO dependence differs from that reported here; this
difference appears to reflect the power-law analysis and
FIG. 9. Arrhenius plots for various NO and O2 partial pressures. ꢅ,
kPa NO on Pt/Al2O3; ꢂ, 1 kPa NO on Cu(0.58); ꢆ, 0.6 kPa NO on
1
the higher NO partial pressures used in the earlier study Cu(0.58); ꢃ, 0.6 kPa NO and 1 kPa O2 on Cu(0.58).

STEADY-STATE CATALYTIC NO DECOMPOSITION REACTIONS ON Cu–ZSM5
83
together with rates measured on Pt/Al2O3. All Cu–ZSM5
samples in this study showed similar temperature trends,
and NO decomposition rates reached maximum values at
∼
773 K for 1 kPa NO. In contrast, NO decomposition on
Pt/Al2O3 showed normal Arrhenius behavior at 673–873 K,
with an apparent activation energy of 18 kcal/mol, which
agrees with the value of 18.4 kcal/mol reported previously
(
26). On Cu–ZSM5, the temperature at which maximum
NO decomposition rates occur varies with NO and O2 par-
tial pressures. At a given NO partial pressure (0.6 kPa), the
rate maximum shifts from ∼753 to ∼793 K when 1 kPa O2 is
added to a stream containing only NO reactants. The tem-
perature required for maximum NO decomposition rates is
slightly higher for 1 kPa NO (∼773 K) than for 0.6 kPa NO
(
∼753 K).
This unusual temperature effect led us to examine the ef-
fects of temperature on individual rate parameters (kapp and
Kα), which were separated into their elementary step rate
and equilibrium constants. This approach allows us to as-
sign specific temperature dependences for elementary steps
and the effects of temperature and of gas-phase concentra-
tions on the relative coverage of reactive intermediates. The
apparent rate constant kapp reflects the kinetic and thermo-
dynamic behavior of several elementary steps. Without loss
of generality, we consider both NO adsorption steps (steps 1
and 2 in Table 2) to be quasiequilibrated, which makes step
FIG. 10. Arrhenius plot for the parameters kapp and Kα obtained from
the regression for Cu(0.58).
The equilibrium constant for O2 adsorption, Kα is given by
Aa,α
Ad,α
Aa,α
Ad,α
ꢀ
Qa/RT
−ꢀHa/RT
e .
Kα =
e
=
[22]
3
the kinetically relevant step in the overall sequence in
Table 2. All rate and equilibrium constants (K1, K2, and k3)
are of Arrhenius form and they appear in the apparent rate
constants (kapp) as
From the regression analysis presented earlier (Fig. 8 and
Eq. [13]), we obtained kapp and Kα at various temperatures
(
723, 773, and 823 K). An Arrhenius plot of these param-
eters is shown in Fig. 10. Both kapp and Kα decreased with
increasing temperature, with Kα showing the stronger tem-
kapp = Aappe−
kapp = K1 K2k3[L],
Aa,1
Eapp/RT
,
[15]
[16] perature dependence. These data give values of −11 kcal/
mol for Ea,app and −18 kcal/mol for ꢀHα.
Aa,1
Ad,1
ꢀ
Q1/RT
−ꢀH1/RT
−ꢀH2/RT
K1 =
e
=
=
e
e
,
,
[17]
The observed negative apparent activation energy re-
quires that the activation energy for the kinetically relevant
step 3 be smaller than the sum of the heats of adsorption
for the two preequilibrated adsorption steps (1 and 2). The
low activation energy for step 3 is evident from the for-
mation of N2O near room temperature during the initial
transient exposure of Cu–ZSM5 samples to NO (11). NO
adsorption steps are likely to be exothermic, as is the case
for most chemisorption processes, and their thermodynam-
ics become unfavorable with increasing temperature. These
data suggest that NO adsorption equilibrium constants (K1
and K2) decrease more strongly with temperature than k3
increases, leading to an overall decrease in kapp with increas-
ing temperature (Fig. 10). The decrease in Kα with increas-
ing temperature shows that Kα represents an exothermic
net reaction. This is consistent with expected exothermic
nature of the net O2 adsorption process described by the
combination of elementary steps leading to Kα. The ob-
Ad,1
Aa,2
Ad,2
Aa,2
Ad,2
ꢀ
Q2/RT
K2 =
e
[18]
[19]
k3 = A3e−
Ea,3/RT
.
Theapparentactivationenergyisobtainedfromtheslope
in the Arrhenius plot where (ln kapp) is plotted against the
reciprocal temperature (Eq. [20]). This apparent activation
energy contains the adsorption enthalpies for adsorption
steps 1 and 2 and the activation energy for step 3:
Ea,app
ln kapp = ln Aapp −
RT
ꢂ
ꢃ
ꢂ
ꢃ
Aa,1
Aa,2
Ad,2
=
ln
+ ln
+ ln A3 + ln[L]
Ad,1
ꢂꢀ
ꢃ
H1 + ꢀH2 + Ea,3
−
,
[20] served temperature behavior of K is also consistent with
α
RT
the observed increase in reduced Cu dimers with increasing
[21] temperature (10, 11).
Eapp = Ea,3 + ꢀH1 + ꢀH2.

´
MODEN ET AL.
8
4
∗
the most abundant O , but at 723 K and above its concen-
∗
tration is negligible compared to ∗ and O . This adds an
extra denominator term below 723 K, with an equilibrium
constant corresponding to an exothermic reaction; it be-
∗
haves with temperature similarly to the term for O , and it
is not included in order to simplify the arguments without
loss of generality. As the temperature increases, Kα[O2]1
and the other denominator terms decrease, leaving ∗ as the
masi, and Eq. [13] approaches
/2
2
r = kapp[NO]
[24]
as O2 desorption is favored and ∗ becomes the masi species.
∗
At temperatures above 773 K, for which the O term in the
denominator of Eq. [13] becomes smaller with increasing
temperature, kapp becomes the overall rate constant for the
NO decomposition reaction; it then acquires a negative ap-
∗
parent activation energy. This change in masi from O to
∗
as temperature increases causes the observed change in
∗
apparent activation energy and the maximum in NO de-
composition rates at intermediate temperatures (Fig. 9).
The shift in the temperature at which the rate reaches a
maximum value for different gas-phase compositions can
be explained in terms of these changes in the identity of
the masi with increasing temperature. High O2 partial pres-
FIG. 11. Calculated oxygen surface coverage ([O ]/[L]) vs partial
pressure of O2 on Cu(0.58).
From these values of Kα and the gas-phase O2 concen-
tration, fractional [O ] surface coverages can be obtained,
using denominator terms in the rate expression valid at
23 K and above (Eq. [13]). The oxygen fractional coverage
[O ]/[L]) is plotted as a function of O2 partial pressure in
Fig. 11 at 723, 773, and 823 K. At low temperatures and
high O2 partial pressures, the oxygen fractional coverage
∗
∗
sures cause the shift in the masi from O to ∗ to occur at
7
(
higher temperatures; as a result, the temperature required
to reach maximum NO decomposition rate would increase
∗
(
Fig. 9). The temperature required for maximum rates is
indeed higher at 1 kPa O2 (793 K) than when O2 is not
added (753 K). Also, this temperature is higher at 1 kPa
NO (773 K) than at 0.6 kPa NO (753 K), because higher O
concentrations are present at the higher O2 concentrations
prevalent during decomposition of higher concentrations
of NO. These data and their mechanistic implications sug-
gest that unfavorable NO adsorption thermodynamics lead
to the observed decrease in NO decomposition rates at high
temperatures.
∗
approaches unity and O becomes the masi during NO de-
composition. At higher temperature and lower O2 partial
pressures, the oxygen coverage decreases and free sites (∗)
become the masi.
Combining the temperature dependence of these param-
eters and the O surface coverage, we return to the rate
expression (Eq. [13]) in order to describe how NO decom-
positionrateschangewithtemperatureatvariousgas-phase
compositions (shown in Fig. 9):
∗
∗
This study illustrates the mechanistic complexity of the
stoichiometrically simple NO decomposition reaction, as
kapp[NO]2
+ Kα[O2]1/2
r = 1
(kapp = K1 K2k3[L]).
[13] well as the role of NO as both a reductant and an oxi-
dant in catalytic turnovers. The proposed sequence of steps
At low temperatures, the Kα[O2]1/2 term in the denomi-
nator of Eq. [13] becomes larger than 1, and the rate ap-
proaches an expression of the form
(Table 2) is consistent with transient (11) and infrared (17,
0), studies, and it includes pathways for the formation and
reaction of N2O and NO2 as gas-phase intermediates in the
formation of N2 and O2 as the ultimate NO decomposition
products.
2
kapp[NO]2
Kα[O2]1/2
r =
.
[23]
CONCLUSIONS
For this rate expression, the overall rate constant is kapp/Kα,
which increases with increasing temperature. This leads to
Steady-state and transient NO decomposition data were
an increase in the NO decomposition rate with increasing used to infer a set of elementary steps for the formation of
temperature and to a positive apparent activation energy N2O, NO2, N2, and O2 during NO decomposition on Cu–
∗
as long as O remains the masi. At temperatures below ZSM5. The adsorbed species and Cu structures involved
∗
∼
723 K, NO is present in substantial amounts along with in these elementary steps are consistent with previous
3

STEADY-STATE CATALYTIC NO DECOMPOSITION REACTIONS ON Cu–ZSM5
85
spectroscopic evidence. N2O was the initial product of NO nancial support from the Korea Science and Engineering Foundation
KOSEF) for D. K. Lee are all gratefully acknowledged. The authors also
(
decomposition and the adsorption of both NO and N2O re-
+
+ 2+
acknowledge the partial funding of this work by an unrestricted grant from
Catalytica, Inc. The authors also thank Mr. Dario Pinna and Prof. Guido
Buzzi-Ferraris for their assistance in accessing and implementing the re-
gression code and Dr. Laurent D. Kieken (Catalytica) for helpful technical
quired reduced Cu dimers {Cu –ꢀ–Cu } formed by the
2+
removal of a bridging oxygen from vicinal Cu species at
2+
2−
2+ 2+
exchange sites {Cu –O –Cu
}
. The redox nature of
the NO and N2O decomposition pathways was confirmed discussions.
bytransientkineticmeasurements, whichshowedthenetin-
corporation of oxygen and a concomitant decrease in reac-
tion rates as decomposition proceeded on Cu–ZSM5 sam-
REFERENCES
+
ples containing reduced Cu dimers. The reaction of two
1
2
3
. Topsoe, N.-Y., Cattech 1, 125 (1997).
. Nakajima, F., Catal. Today 10, 1 (1991).
. Stull, D. R., Westrum, E. F., Jr., and Sinke, G. C., “The Chemical
Thermodynamics of Organic Compounds.” Krieger, Malabar, FL,
NO molecules adsorbed on vicinal Cu atoms led to N2O
formation at room temperature, suggesting that the activa-
tion energy for this bimolecular reaction is very small. The
+
1
987.
significant heat of adsorption of NO on Cu dimers, and the
2+
2−
2+ 2+
+
+ 2+
4. Iwamoto, M., Furukawa, H., Mine, Y., Uemura, F., Mikuriya, S., and
Kagawa, S., J. Chem. Soc. Chem. Commun. 16, 1272 (1986).
transition from {Cu –O –Cu
}
to {Cu –ꢀ–Cu } as
the predominant Cu species, causes the observed decrease
in NO decomposition rates above temperatures higher than
5
. Iwamoto, M., Yahiro, H., Mine, Y., and Kagawa, S., Chem. Lett. 2, 213
1989).
6. Iwamoto, M., Catal. Today 10, 57 (1991).
. Iwamoto, M., Yahiro, H., Tanda, K., Mizuno, N., Mine, Y., and
Kagawa, S., J. Phys. Chem. 95, 3727 (1991).
. Iwamoto, M., Yahiro, H., Mizuno, N., Zhang, W. X., Mine, Y.,
Furukawa, H., and Kagawa, S., J. Phys. Chem. 96, 9360 (1992).
. Beutel, T., Sarkany, J., Lei, G. D., Yan, J. Y., and Sachtler, W. M. H.,
J. Phys. Chem. 100, 845 (1996).
0. Da Costa, P., Moden, B., Meitzner, G., Lee, D. K., and Iglesia, E.,
submitted for publication.
(
∼
750 K.
7
8
9
The communication between O2 and oxygen adsorbed as
bridged species on Cu dimers is mediated by rapid and re-
versible NO oxidation and NO2 decomposition reactions,
which allow reactions between remote adsorbed oxygen
species via migration of NO2 in the gas-phase instead of ad-
sorbed oxygen on the zeolite surface. In these steps, the de-
composition of an adsorbed nitrate species leads to the evo-
lution of O2 molecules and to the regeneration of the NO
molecules used to carry oxygen via the gas phase as NO2.
1
1
1
1. Moden, B., Da Costa, P., Lee, D. K., and Iglesia, E., submitted for
publication.
2. Liu, D. J., and Robota, H. J., Catal. Lett. 21, 291 (1993).
These pathways were confirmed by the equilibrium NO2 13. Giamello, E., Murphy, D., Magnacca, G., Shioya, Y., Nomura, T., and
Anpo, M., J. Catal. 136, 510 (1992).
concentrations measured at all conditions during NO de-
1
1
4. Valyon, J., and Hall, W. K., J. Phys. Chem. 97, 1204 (1993).
5. Spoto, G., Zecchina, A., Bordiga, S., Ricchiardi, G., Martra, G.,
Leofanti, G., and Petrini, G., Appl. Catal. B 3, 151 (1994).
6. Lei, G. D., Adelman, B. J., Sarkany, J., and Sachtler, W. M. H., Appl.
Catal. B 5, 245 (1995).
composition on Cu–ZSM5. These O2 formation pathways
avoid the need for vicinal oxygen atoms and for recombi-
native desorption steps; they suggest that NO acts both as
an oxidant and as a recyclable reductant during its catalytic
decomposition on Cu–ZSM5.
The mechanism proposed is consistent with the steady-
state kinetic data and leads to reasonable values and tem-
perature dependences for all rate parameters. The relative
abundance of various surface species inferred from this rate
equation is consistent with independent spectroscopic stud-
ies of adsorbed species and of Cu structures. The superior
activity of Cu–ZSM5 samples in NO decomposition reac-
tions reflects the availability of two-electron redox cycles
1
1
1
1
2
7. Aylor, A. W., Larsen, S. C., Reimer, J. A., and Bell, A. T., J. Catal.
1
57, 592 (1995).
8. Jang, H. J., Hall, W. K., and d’Itri, J. L., J. Phys. Chem. 100, 9416
1996).
(
9. Schay, Z., Knozinger, H., Guczi, L., and Pal-Borbely, G., Appl. Catal.
B. 18, 263 (1998).
0. Konduru, M. V., and Chuang, S. S. C., J. Phys. Chem. B 103, 5802
(
1999).
2
2
2
1. Konduru, M. V., and Chuang, S. S. C., J. Catal. 187, 436 (1999).
2. Konduru, M. V., and Chuang, S. S. C., J. Catal. 196, 271 (2000).
3. Park, S.-K., Kurshey, V., Luan, Z., Lee, C. V., and Kevan, L., Micro-
porous Mesoporous Mater. 38, 255 (2000).
2
+
+
without requiring the reduction of Cu beyond Cu by us-
2
+
2
2
4. Sarkany, J., d’Itri, J. L., and Sachtler, W. M. H., Catal. Lett. 16, 241
ing a Cu dimer to accommodate a two-electron reduction
process without the formation (and agglomeration) of Cu
metal. In addition, the labile nature of the bridged oxygen
ligands in oxidized Cu dimers and the balanced equilibrium
(
1992).
5. Hightower, J. W., and van Leirsburg, D. A., in “The Catalytic Chem-
istry of Nitrogen Oxides” (R. L. Klimisch and J. G. Larson, Eds.),
p. 63. Plenum, New York, 1975.
2+
2−
2+ 2+
between their oxidized {Cu –O –Cu
}
and reduced 26. Amirnazmi, A., Benson, J. E., and Boudart, M., J. Catal. 30, 55 (1973).
+
+ 2+
2
2
7. Amirnazmi, A., and Boudart, M., J. Catal. 39, 383 (1975).
{
Cu –ꢀ–Cu } states allow the redox cycles required as
8. Lim, K. J., Loffler, D. G., and Boudart, M., J. Catal. 100, 158
part of each NO decomposition turnover to occur rapidly.
(
1986).
2
3
9. Xie, S. B., Rosynek, M. P., and Lunsford, J. H., J. Catal. 188, 24
(1999).
0. Teraoka, Y., Harada, T., and Kagawa, S., J. Chem. Soc. Faraday Trans.
94, 1887 (1998).
ACKNOWLEDGMENTS
A CF Milj o¨ fonden grant for B. Moden, a “Programme Lavoisier” grant
from the French Foreign Affairs Ministry for P. Da Costa, and the fi- 31. Yamashita, T., and Vannice, A., J. Catal. 163, 158 (1996).

´
8
6
MODEN ET AL.
3
3
3
3
2. Winter, E. R. S., J. Catal. 22, 158 (1971).
38. Froment, G. F., and Bischoff, K. B., “Chemical Reactor Analysis and
3. Li, Y., and Hall, W. K., J. Phys. Chem. 94, 6145 (1990).
4. Li, Y., and Hall, W. K., J. Catal. 129, 202 (1991).
5. Campa, M. C., Indovina, V., Minelli, G., Moretti, G., Pettiti, I., Porta, P.,
and Riccio, A., Catal. Lett. 23, 141 (1994).
Design,” 2nd ed. Wiley, New York, 1990.
39. Donati, G., and Buzzi-Ferraris, G., Chem. Eng. Sci. 29, 1504 (1974).
40. Villa, P. L., Forzatti, P., Buzzi-Ferraris, G., Garone, G., and Pasquon,
I., Ind. Eng. Chem. Prod. Res. Dev. 24, 12 (1985).
3
3
6. Zhang, Y., Lee, K. M., Sarofim, A. F., Hu, Z. C., and Flytzani-
Stephanopoulos, M., Catal. Lett. 31, 75 (1995).
7. Schay, Z., and Guczi, L., Catal. Today 17, 175 (1993).
41. Li, Y., and Armor, J. N., Appl. Catal. 76, L1 (1991).
42. Boudart, M., and Djega-Mariadassou, G., “Kinetics of Heterogeneous
Catalytic Reactions.” Princeton University Press, Princeton, NJ, 1984.