Effects of O2 on the reduction of NO over prereduced CaO surfaces: A mechanistic understanding

Article abstract of DOI:10.1021/jp983408f

The effect of O₂ on the reduction of NO over prereduced CaO surfaces is investigated. The experimental results suggest the existence of at least three different reaction channels, of which two are related to the high-temperature reduction of the CaO surfaces and involve the use of extra electrons in breaking the NO bond. The third reaction channel does not employ extra electrons for bond breaking, but the activity is affected by the amount of adsorbed surface oxygens. The difference between the former two reaction channels is found in the temperature needed for an observable activity. The reaction channel which is already active at low temperatures is described by a model based on F-centers, whereas the one which needs elevated temperatures involves a hole transport through the bulk. The activation energy for this transport is determined experimentally using a temperature-programmed reaction technique as well as theoretically by means of ab initio quantum chemistry calculations. Room-temperature exposure to O₂ is suggested to result in a poisoning of the F-centers, but has only a minor effect on the reaction channel proposed for high temperatures. Effects on the reduction of NO of time as well as temperature for the O₂ exposure step are also investigated and found to be consistent with an understanding based on the coexistence of different reaction channels. ? 1999 American Chemical Society.

Full text of DOI:10.1021/jp983408f

J. Phys. Chem. B 1999, 103, 2195-2201
2195
Effects of O2 on the Reduction of NO over Prereduced CaO Surfaces: A Mechanistic
Understanding
Filip Acke* and Itai Panas
Department of Inorganic Chemistry, Chalmers UniVersity of Technology and Go¨teborg UniVersity,
S-412 96 Go¨teborg, Sweden
ReceiVed: August 18, 1998; In Final Form: January 20, 1999
The effect of O2 on the reduction of NO over prereduced CaO surfaces is investigated. The experimental
results suggest the existence of at least three different reaction channels, of which two are related to the
high-temperature reduction of the CaO surfaces and involve the use of extra electrons in breaking the NO
bond. The third reaction channel does not employ extra electrons for bond breaking, but the activity is affected
by the amount of adsorbed surface oxygens. The difference between the former two reaction channels is
found in the temperature needed for an observable activity. The reaction channel which is already active at
low temperatures is described by a model based on F-centers, whereas the one which needs elevated
temperatures involves a hole transport through the bulk. The activation energy for this transport is determined
experimentally using a temperature-programmed reaction technique as well as theoretically by means of ab
initio quantum chemistry calculations. Room-temperature exposure to O2 is suggested to result in a poisoning
of the F-centers, but has only a minor effect on the reaction channel proposed for high temperatures. Effects
on the reduction of NO of time as well as temperature for the O2 exposure step are also investigated and
found to be consistent with an understanding based on the coexistence of different reaction channels.
1. Introduction
and Fe) of the CaO material.12 The role of the impurities is to
provide electron sinks during the surface oxygen abstraction
Nitrogen oxides, produced during combustion of fossil fuels,
are harmful to the environment. Catalytic materials can be used
to effectively reduce them. Several catalytic systems have been
commercially available for some years. The three-way catalyst
for stoichiometric combustion is one of them. This technique
uses supported and promoted Pt to decrease both the NOx, CO,
and hydrocarbon levels in car exhausts. For stationary power
sources selective catalytic reduction by NH3 over a V2O5-based
catalyst is commonly used. At present, substantial efforts are
being made to find an operative system for the reduction of
NOx by reducing species present in the flue gas under oxygen
excess. Supported noble-metal and several zeolite systems have
been shown to be efficient in the reduction of NOx by
hydrocarbons.1-3 However, the most promising approach is the
direct catalytic decomposition of NO to N2 and O2.4 Several
materials have been shown to increase the rate of this reaction,
i.e., Cu-ZSM-5,5,6 Sr-Fe oxides,7 perovskite-type com-
pounds,8,9 and cobalt oxide modified with Ag. One feature,
shown by all these materials, is the ability to release oxygen at
moderate temperatures.10
Although a thorough mechanistic understanding is wanted,
it is clear that the detailed activity of each catalyst is system
specific. Still, common properties are sought by focusing on
the necessary prerequisites for the reduction of NO. This article
is part of a series that tries to elucidate the nitrogen chemistry
during reduction of NO over alkaline earth oxide surfaces.11-13
CaO is used as a model substrate to study basic reactions, both
by experiment11-13 and theory.14-19 In a previous work, cor-
relation was demonstrated between catalytic activity toward the
reduction of NO by CO or H2, and the impurity content (Na
by the reducing agent. Two models describing this step are
currently under debate, i.e., a semiconductor model based on
O- states in the band gap12,20 and a model based on F-centers.21
The oxygen abstraction step creates sites for breaking the N-O
n-
bond,11,12 and a mechanism involving an N2O2 intermediate
was suggested.11,19 Furthermore, the effect of the oxygen
abstraction step on the subsequent reduction of NO was shown
to be independent of the reducing agent, i.e., H2 or CO.11
Reduction of NO upon NO adsorption at room temperature and
during subsequent heating ramps in Ar was even observed for
fully oxidized CaO surfaces.22 The effect of preadsorbed surface
oxygens was examined.
The effect of excess oxygen on the activity of reduced oxide
surfaces is investigated. This is done by exposing a prereduced
CaO surface to O2 at different temperatures and measuring the
effect on the reduction of NO. The results are compared with
the reduction of NO over fully oxidized surfaces. The effects
of sequential and simultaneous exposure to O2 and NO,
combined with the discovery of sites accessible only at elevated
temperature, result in a comprehensive understanding of the
reduction of NO and in the proposition of reaction mechanisms
involving both semiconductor properties and F-centers. The
former sites are discussed to be the ones involved in a reaction
channel operative at high temperature, whereas the latter are
involved in a reaction channel which is already active at low
temperature. Ab initio quantum chemistry calculations are used
to further develop the understanding of the semiconductor model
in terms of a hole hopping mechanism for charge transport.
2. Experimental Section
A fixed bed reactor connected to a quadropole mass spec-
trometer (QMG 421 of Balzers) was used. The quartz reactor
* Corresponding author. Telephone: +46 (0)31 772 28 86. Fax: +46
(0)31 772 28 53. E-mail: filip@inoc.chalmers.se.
10.1021/jp983408f CCC: $18.00 © 1999 American Chemical Society
Published on Web 03/05/1999

2196 J. Phys. Chem. B, Vol. 103, No. 12, 1999
Acke and Panas
(i.d. 22 mm, length 500 mm) had an asymmetric construction
to avoid heating the upper metal fitting and the vacuum-tight
Viton O-ring. Gas sampling was done by a quartz capillary 2
mm under the bed supporting a sintered quartz filter. The
temperature was measured with a K-type thermocouple 3 mm
under the tip of the capillary, avoiding any interference in the
results due to its catalytic activity. The heating ramps were
controlled with a second thermocouple in contact with the
heating coil.
The bed material investigated consisted of 0.750 g of CaO
(Fisher Scientific) mixed with 1 g of quartz sand, pro analysi
(Merck), and was placed on a sintered quartz filter in the reactor.
The quartz sand was added to reduce the pressure drop. The
gases used were 5000 ppm NO, 3000 ppm isotope-labeled
15NO, 5000 ppm H2, and 10% O2, all in Ar. The flows were
controlled by mass flow controllers (Brooks, type 5850E) to a
flow rate of about 20 mL/min (293 K, 1 bar), unless stated
otherwise. The bed material had been heat treated overnight at
800 °C in an Ar flow and was then exposed to H2 at this
temperature for 15 min. This reducing step was repeated after
every experiment.
Figure 1. Room-temperature N2O formation during exposure of a
prereduced CaO surface to NO: effect of intermediate O2 exposure at
room temperature.
In addition, quantum chemistry calculations were performed
in order to clarify basic features of the semiconductor model.
An embedded cluster model was employed, where three atoms
are described explicitly and the rest by point charges or
projection operator dressed point charges. The latter are used
to mimic the effect of Pauli repulsion toward the inert
Ca2+ ion cores. The all-electron descriptions of the linear
[O-Ca-O]-1 unit, which is taken to accommodate one hole,
uses the 4s 3p 1d and 5s 4p atomic natural orbitals descriptions23
for oxygen and calcium, respectively. The calculations were
performed by means of a Coulomb hole augmented restricted
open-shell Hartree-Fock method.24,25 The overall reliability of
similar embedded cluster schemes has been shown in previous
studies.16,19
3. Results
3.1. O2 Treatment at Room TemperaturesEffects on the
Reduction of NO. In a first series of experiments, the room-
temperature stability of the sites induced by the prereduction
step were investigated by exposing the bed material to a gas
flow of 5.2% O2 in Ar. The effect of the intermediate exposure
to O2 on the activity of the substrate toward the reduction of
NO is examined after the reactor was flushed by Ar to evacuate
remaining O2. The bed material was therefore exposed to a gas
flow of NO (2400 ppm) for 15 min while the reactant, NO, as
well as possible reaction products, N2, O2, and N2O, were
monitored by the mass spectrometer. Oxygen exposure times
between 0 (no intermediate O2 exposure) and 120 min were
investigated. The results are shown in Figure 1. In the absence
of any O2 treatment, a transient formation of N2, followed by a
transient formation of N2O, is observed. The N2 peak is due to
a reaction channel for complete reduction of NO. The reduction
capacity of the CaO substrate diminishes after about 180 s and
the further reduction of NO becomes incomplete, i.e., a partial
reduction of NO to N2O. Intermediate exposure of the reduced
substrate to O2 causes the formation of N2 to disappear. Only
the reaction channel for partial reduction remains. Increased O2
exposure time results in a decreased formation of N2O. No O2
formation was observed in any of the experiments.
Figure 2. Formation of N2O (a) and N2 (b), as well as NO desorption
(c) during the successive heating ramps in Ar: effect of intermediate
O2 exposure at room temperature.
The bed material was then exposed to a heating ramp up to
800 °C. This was done in a flow of Ar, using a heating rate of
0.5 K/s. In Figure 2a, the formation of N2O is plotted versus
ramp temperature for the different O2 exposure times. A decrease
in N2O formation is observed with increasing O2 exposure time.
The same trend is observed for the formation of N2 (Figure
2b). The results show the existence of reaction channels which

Effects of O2 on the Reduction of NO Prereduced CaO
J. Phys. Chem. B, Vol. 103, No. 12, 1999 2197
Figure 3. N2 formation during exposure to NO at 800 °C: effect of
intermediate O2 exposure at room temperature.
Figure 4. Room-temperature N2O formation during exposure of a
prereduced CaO surface to NO: effect of intermediate exposure to O2
at 30, 170, and 380 °C.
were not easily accessible at room temperature. Formation of
N2O is found at temperatures between 50 and 425 °C with a
maximum at about 150 °C independent of O2 exposure time.
N2 is formed between 60 and 700 °C. The corresponding
desorption of NO is shown in Figure 2c. Two desorption peaks,
a low (135 °C)- and a high-temperature (430°) peak, are
observed. The low-temperature desorption peak is not influenced
by the intermediate exposure to O2. This is not the case for the
high-temperature peak, i.e., an increased O2 exposure time
results in an increased NO desorption. This could be related to
the decreased amount of NO that is reduced to N2.
After the heating ramp experiments, the bed materials were
reexposed to NO at 800 °C (same gas mixture as previously).
The transient formation of N2, seen in Figure 3, demonstrates
the presence of sites that were not reoxidized during the previous
treatment and thus in a reaction channel only accessible at
elevated temperatures. No formation of N2O is observed at 800
°C. This is expected since CaO is known to catalyze the
decomposition of N2O.18 Interesting is the slow decrease in the
formation of N2 with increasing O2 exposure time.
3.2. O2 Treatment at Elevated TemperaturesEffects on
the Reduction of NO. In these experiments the effect of an
intermediate O2 exposure was investigated at 30, 170, and 380
°C, while the O2 exposure time was kept at 15 min. The bed
was first cooled in Ar to 30 °C before exposing it to NO for
the two latter O2 exposure temperatures. N2O was the only
reaction product formed at room temperature. In Figure 4 the
N2O formation is displayed versus time for the three investigated
O2 exposure temperatures. A decrease is observed with increas-
ing temperature. After exposure to NO, the bed materials were
heated under an Ar flow to 800 °C. In Figure 5a the transient
N2O formation is plotted versus ramp temperature. A maximum
Figure 5. Formation of N2O (a) and N2 (b), as well as NO desorption
(c) during the successive heating ramps in Ar: effect of intermediate
exposure to O2 at 30, 170, and 380 °C.
is observed at about 150 °C and the amount decreases with
increasing O2 exposure temperature. The low-temperature
formation of N2 disappears for the exposure temperatures of
170 and 380 °C (Figure 5b). The high-temperature N2 formation
is also affected by the O2 exposure temperature, i.e., the amounts
decrease with increasing exposure temperature. The formation
of N2 for an intermediate O2 exposure at 380 °C differs from
the others, i.e., the peak temperature is shifted to higher
temperatures. This effect is correlated to a third NO desorption
peak as can be observed in Figure 5c. In this figure, the
desorption of NO is displayed versus the ramp temperature. This
third desorption peak is accompanied by an O2 formation and
is ascribed to the desorption of an -NO2 intermediate (see
below). It is interesting to note that increased O2 exposure
temperatures result in a decreased NO desorption. This is the
opposite of what was observed by increasing the time of O2
exposure.
The bed material was then reexposed to NO at 800 °C. A
transient formation of N2 is seen in Figure 6 for all O2 exposure
temperatures, demonstrating the presence of sites that were not
reoxidized during the O2 treatment. The amount of formed N2
decreases with increasing intermediate O2 exposure temperature.
The unexpected robustness of this high-temperature channel to
O2 exposure is intriguing and was therefore the subject for
further investigation. The temperature at which it is active

2198 J. Phys. Chem. B, Vol. 103, No. 12, 1999
Acke and Panas
Figure 6. N2 formation during exposure to NO at 800 °C: effect of
intermediate exposure to O2 at 30, 170, and 380 °C.
Figure 9. Desorptions of NO and O2 during a heating ramp in a mixture
of NO and O2 mixture for a prereduced CaO surface.
Figure 7. Formation of N2 during heating ramps in NO for a
prereduced CaO: effect of intermediate exposure to O2 at 30, 170, and
380 °C.
Figure 10. Formation of N2O at 30 °C (a), as well as N2 during the
preceding heating ramp in Ar, (b) for an oxidized CaO surface exposed
to NO at room temperature.
Figure 8. Formation of N2 during exposure of a prereduced CaO
surface to a mixture of NO and O2 at 600 and 800 °C, respectively.
toward the reduction of NO can be determined by performing
a heating ramp in an NO-containing atmosphere (1200 ppm)
instead of Ar. The total gas flow was only 12 mL/min during
this experiment. A transient N2 production was observed in a
temperature interval between 500 and 800 °C as is shown in
Figure 7. In this figure, the N2 formation is plotted versus time,
not versus temperature, since not all the sites were reoxidized
upon reaching the final ramp temperature. The formation of N2
was accompanied by a corresponding decrease in the NO signal
(not shown), confirming the actual reduction of NO. The amount
of N2 formed is highest for an O2 exposure temperature of 30
°C and decreases with increasing O2 exposure temperatures, as
expected. A maximum N2 formation is found at about 760 °C
for all three O2 exposure temperatures.
Exposure of the bed material simultaneously to NO (2700
ppm) and O2 (9700 ppm) at 600 and 800 °C, respectively, results
in a transient N2 formation (Figure 8). Formation of neither N2
nor N2O was observed for exposures at 30, 200, or 400 °C.
However, the formation of small amounts of N2 is observed
during the succeeding heating ramp in a gas mixture containing
both NO and O2. In Figure 9, the corresponding NO and O2
signals are shown versus the ramp temperature. Two types of
NO desorption peaks can be seen: a low-temperature peak
which is due to preadsorbed NO and a high-temperature peak
which is likely to be due to a -NO2 intermediate, as can be
concluded by correlating the NO desorption and the O2
formation. The gas flow for these experiments was 15 mL/min.
3.3. Reduction of NO over an Oxidized CaO Surface. In
this experiment the activity of an oxidized CaO surface toward
the reduction of NO was investigated. The bed material was
exposed to a gas flow of NO (2400 ppm) at 800 °C for 15 min
and cooled in Ar to room temperature. The bed material was
then reexposed to NO (2400 ppm) for 15 min and the formation
of possible reaction products (as N2 and N2O) was monitored.
A transient production of N2O is observed (Figure 10a). Finally
the bed material was heated to 800 °C under a flow of Ar. The
N2 and N2O formations are displayed versus ramp temperature
in Figure 10b, where a temperature dependence in the formation
of N2 and N2O consistent with that displayed in Figures 1 and
2b is found.

Effects of O2 on the Reduction of NO Prereduced CaO
J. Phys. Chem. B, Vol. 103, No. 12, 1999 2199
3.4. Kinetics. Equations relating Arrhenius kinetic parameters
A and Ea (preexponential factor and apparent activation energy,
respectively), surface coverage θ, and linear heating rate (â)
have been reviewed by Lord and Kittelberger.26 It was shown
that the slope of a plot of ln(Tm2 θmn-1/â) versus 1/Tm reproduces
the apparent activation energy for desorption by multiplying it
with the gas constant R. Tm represents the temperature of
maximum desorption, θm the peak coverage, and n the order of
the desorption reaction. In the present work, n will be taken as
one. This simplifies the plot to ln(Tm2/â) versus 1/Tm. Calcula-
tions have shown that the assumption of a reaction order of 1
does not introduce major errors in the activation energy.26 The
activation energy for reduction of NO in the high-temperature
channel is determined by exposing a reduced CaO surface to
NO (2500 ppm) during a heating ramp between room temper-
ature and 900 °C, while measuring the formation of N2.
Repeating this for different heating rates (â) results in an
activation energy for the reduction of NO to N2 of about 25
kcal/mol.
The effect of lattice parameter on the hole hopping mecha-
nism for charge transport is illustrated semiquantitatively by
employing a regularized Hartree-Fock method.24,25 The activa-
tion energies are taken to be reflected in the relative stabilities
of the symmetry broken odd-electron system O--Ca2+-O2-
embedded cluster16,19 and the symmetric system, which delo-
calizes the hole over the two oxygen ions. Activation energies
in the range 20-32 kcal/mol are obtained by simply varying
the lattice constant in the range of the alkaline earth oxide series
from MgO to BaO. The activation energy for holes hopping in
CaO is determined to be 26.2 kcal/mol. It is gratifying to note
the semiquantitative agreement with the apparent activation
energies for NO reduction on these substrates.12 This was
expected on the basis of experience with similar systems.14-16
The main sources of error are due to the rigid cluster ansatz,
the assumption that nearest-neighbor interactions in an external
electrostatic potential determines the charge carrier properties
of the alkaline earth oxides, use of an effective electron
correlation description, and neglect of dispersion interactions.
The potential for such a model in explaining experimental
observations relies on the cancellation of errors. In this respect
the objective to seek the possible origin for a trend in an
activation energy is ideal. This is why the somewhat fortuitous
agreements between experiment and theory, regarding both trend
and magnitudes, can be taken to support the existence of a hole
transport controlled reaction channel for doped alkaline earth
oxides.
A semiconductor model based on an apparent activation
energy of 25 kcal/mol for hole transport cannot explain the
reduction of NO at room temperature since for the reduction of
NO at these temperatures a much lower activation energy is
expected. Figure 1 shows the transient formation of N2 and N2O
during exposure of the CaO surface to NO at room temperature.
A significant difference is observed in the reduction of NO with
and without an intermediate O2 exposure, i.e., a reduced NO
reduction capacity and a disappearance of the N2 formation.
Increasing the O2 exposure time results in a further decrease of
the N2O formation. The activity at these low temperatures is
complicated and can best be explained by a combination of two
reaction channels of which only one involves extra electrons.
These extra electrons must be easy to access, which makes
electrons localized in F-centers possible candidates. The nature
of these has been extensively reviewed in the literature.21,27-29
An intermediate room-temperature exposure to O2 would
stabilize the electrons of the F-center by forming peroxides and/
or superoxides. These sites, involved in a reaction channel
accessible at low temperatures and referred to as the “low-
temperature sites”, are thus expected to be eliminated by an
exposure of the substrate to O2. Left is then the activity observed
after the intermediate exposure to O2. At this stage there are no
electrons localized in F centers available and the activation
energy for hole transport is too high to have a significant
participation. A third reaction channel which does not involve
transport of holes and/or F centers and of which the effectiveness
decreases with increasing O2 exposure time is therefore needed.
Such a reaction channel has been proposed by Platero et al.30
and Cerruti et al.31 for MgO surfaces, based on infrared
spectroscopy studies:
4. Discussion
The effect of O2 on the reduction of NO over prereduced
CaO surfaces is investigated and the resistance toward poisoning
by O2 is shown to depend on the set up of the experiments, i.e.,
sequential (Figures 1-7) or simultaneous (Figures 8 and 9)
exposure to O2 and NO. The role of a prereduction has been
discussed before11,12 and results in a surface oxygen abstraction,
depositing the electrons of the abstracted oxygen ion in the
substrate. The fate of these electrons is not clear, and two models
are under consideration. One is based on a semiconductor model
using impurity states in the band gap as electron acceptors,
whereas in the other model the excess electrons are localized
in F-centers. An activation energy for the oxygen abstraction
step of about 25 kcal/mol was found.11
A common feature in the results of the sequential room-
temperature exposure to O2 and NO is that some sites are
inaccessible until elevated temperatures (Figure 3). For heating
ramps in NO (Figure 7), a maximum reduction is observed at
about 760 °C (heating rate of 0.5 K/s). Thus, even in the
presence of an oxidant (NO), the reduced CaO is not fully
oxidized until elevated temperatures. The sites involved in the
high-temperature reduction of NO will be referred to as the
“high-temperature sites” and a reaction channel based on an
activated hole transport is proposed to explain their activities.
This understanding results from three independent observations,
i.e., (i) an apparent activation energy of 25 kcal/mol for the
surface oxygen abstraction,11 (ii) an apparent activation energy
of 25 kcal/mol for formation of N2 under heating ramps in NO
(results presented in this paper), and (iii) a maximum apparent
activation energy of 10 kcal/mol for the breaking of a NO bond
in the presence of excess electrons.11 These experimental results
indicate that the transport of holes might be the rate-determining
step, i.e., a transport of holes from the O- species to the surface
during the oxygen abstraction step and back during the
reoxidation.
This understanding would explain the correlation between
apparent activation energy for the reduction of NO by H2 and
the lattice parameter of the respective alkaline earth oxides. In
previous work,12 an increase in apparent activation energy with
lattice parameter, and thus with increasing ionic radii of the
alkaline earth cation, was found. This is in full agreement with
the mechanism presented here. An increased lattice constant
results in a decreased overlap between the oxygen orbitals and
consequently in hole localization and thus in increasing the
apparent activation energy for hole transport.
4NO + 2O2- w 2NO2- + N2O2
(R1)
(R2)
2-
N2O22- w N2O + O2-
It is gratifying to notice that even in the absence of a surface
oxygen abstraction step the reaction mechanism, as proposed

2200 J. Phys. Chem. B, Vol. 103, No. 12, 1999
Acke and Panas
O2 exposure time confirms such an electron consuming mech-
anism, although the latter effect can also partially be due to an
increased NO reduction during the heating ramp the lower the
intermediate O2 exposure time. The effect of this is expected
to be low since the activation energy of hole transport has been
shown to be high.
Increasing the O2 exposure temperature results in a decreased
formation of N2O at room temperature and of N2O and N2
during heating ramps. The formation of N2 during NO exposure
at 800 °C is also significantly decreased with increasing O2
exposure temperature, resulting in fewer high-temperature sites.
This is suggested to be related to an increase for the rate of
O-O bond dissociation and for surface diffusion of oxygen
species with increased O2 exposure temperature. The results
indicate that a temperature of 400 °C is sufficient for a
significant O2 dissociation and peroxide formation, besides the
ones due to O2 adsorption at F-centers. Evidence for this is found
in the high-temperature NO desorption peak in Figure 5c. This
desorption is ascribed to the desorption of an NO2 intermediate,
while the second oxygen of the peroxide remains at the surface
as an On- species consuming the electrons contained in the
impurity-induced sites, and decreasing the amount of electrons
available for NO bond breaking.
Figure 11. Activation energy determination for the reduction of NO
over the high-temperature sites.
Simultaneous exposure to O2 and NO of the reduced CaO
surface results in the disappearance of the formation of N2 and
N2O for exposure at room temperature, 200 and 400 °C, whereas
a transient N2 formation was observed during heating ramps.
Exposure at 600 and 800 °C results in the formation of N2
(Figure 8). The high-temperature sites are shown once more to
be quite stable. The results suggest, for the lower temperatures,
the formation of an intermediate from reaction between NO and
O2 inhibiting the formation of a N-N bond. This N-N bond
is necessary for the formation of N2 and/or N2O.
Figure 12. A schematical overview of the respective reaction channels.
by Platero et al.30 and Cerruti et al.,31 is based on the same
type of intermediate, N2O2n-. However, the probability of its
formation in the absence of excess electrons is expected to be
much lower. The activity of unreduced surfaces has also been
addressed by Yanagisawa32 and was again explained by the
formation of an (NO)2 intermediate, providing the necessary
N-N bond.
5. Conclusions
The presented results comprise basis for the development of
a mechanistic understanding of the NO reduction by H2 or CO
over CaO surfaces at temperatures above 500 °C. The apparent
activation energy for the overall reaction was determined to be
26-27 kcal/mol and independent of impurity content or
reducing agent.12 This value was ascribed to the surface oxygen
abstraction step (about 25 kcal/mol). The reduction of NO by
the high-temperature sites is also found to display an activation
energy of 25 kcal/mol. The similarity between the three
activation energies suggests that the rate-determining step in
the overall reaction is the transport of holes through the bulk.
Further support is found by ab initio quantum chemistry
calculations. The presence of low-temperature sites is further
substantiated and ascribed to F-centers. These do not affect the
overall reaction, since the formation of F-centers is believed to
have a higher activation energy than 25 kcal/mol. The third type
of sites does not involve a previous surface oxygen abstraction,
but the intermediate involved is the same, N2O2n-. The
probability of its formation is, however, much lower in the
absence of excess electrons and does not appear above 500 °C.
It can therefore be omitted in the final reaction mechanism.
A schematic overview of the respective reaction channels is
given in Figure 12.
Even after an exposure to O2 for 120 min, the amount of
N2O formed exceeds that observed for fully oxidized surfaces
(Figures 1 and 10.a). One possible explanation is that this
activity is not associated with the activity for fully (high
temperature) reoxidized surfaces and thus should include a
different type of site. This site could be related to the presence
of impurities close to the surface. Another possible explanation
is the poisoning of the N2O formation by adsorbed surface
oxygens. This effect has been observed in a previous work,22
i.e., the presence of adsorbed surface oxygens deteriorates the
N2O formation upon NO exposure at room temperature.
Extrapolation to the here presented results would suggest a
decrease in N2O formation with increasing O2 exposure time
to result from an alteration of the surface by adsorbed surface
oxygens.
Increasing the temperature by a heating ramp in Ar affects
the substrate. All oxygens that were adsorbed during the O2
exposure, altering the surface and responsible for the decreased
N2O formation at room temperature, will be integrated in the
crystal structure since no O2 desorption was observed at any
time. This integration involves the formation of O2- ions by
the adsorbed oxygen species and thus implies the involvement
of a hole transport. The higher the temperature, the faster this
effect. The decreased N2 formation at 800 °C with increasing
Acknowledgment. The authors are indebted to the Swedish
National Board for Industrial and Technical Development for
financial support (NUTEK).
References and Notes
(1) Burch, R.; Millington, P. J. Catal. Today 1985, 26, 185.

Effects of O2 on the Reduction of NO Prereduced CaO
J. Phys. Chem. B, Vol. 103, No. 12, 1999 2201
(2) Acke, F.; Westerberg, B.; Eriksson, L.; Johansson, S.; Skoglundh,
M.; Fridell, E.; Smedler, G. Stud. Surf. Sci. Catal., Catal. Automot. Pollut.
Control 1998, 116, 285.
(3) Obuchi, A. Appl. Catal. B: EnViron. 1993, 2, 71.
(4) Armor, J. N. Appl. Catal. B: EnViron. 1992, 1, 221.
(5) Iwamoto, M.; Furukawa, H.; Mine, Y.; Uemura, F.; Mikuriya, S.;
Kagawa, S. J. Chem. Soc., Chem. Commun. 1986, 1272.
(6) Iwamoto, M.; Yahiro, H.; Kursumo, T.; Bunyu, S.; Kgawa, S. Bull.
Chem. Soc. Jpn. 1989, 62, 583.
(18) Snis, A.; Miettinen, H. J. Phys. Chem. 1998, 102, 2555.
(19) Snis, A.; Panas, I. Surf. Sci. 1998, 412-413, 477.
(20) Lin, C.-H.; Wang, J.-X.; Lunsford, J. H. J. Catal. 1988, 111, 302.
(21) Wu, M.-C.; Truong, C. M.; Coulter, K.; Goodman, D. W. J. Catal.
1993, 140, 344.
(22) Acke, F.; Skoglundh, M. J. Phys. Chem. B 1999, 103, 972-978.
(23) Pierloot, K.; Dumez, B.; Widmark, P.-O.; Roos, B. O. Theor. Chim.
Acta 1995, 90, 87.
(24) Panas, I. Chem. Phys. Lett. 1995, 245, 171.
(7) Shin, S.; Hatakeyama, Y.; Ogawa, K.; Shimomura, K. Mater. Res.
Bull. 1979, 14, 133.
(25) Panas, I.; Snis, A.; Theor. Chem. Acc. 1997, 97, 232.
(26) Lord, F. M.; Kittelberger, J. S. Surf. Sci. 1974, 43, 173
(27) Marchese, L.; Colcuccia, S.; Martra, G.; Giamello, E.; Zecchina,
A. Mater. Chem. Phys. 1991, 29, 437.
(28) Giamello, E.; Ferrero, A.; Colcuccia, S.; Zecchina, A. J. Phys.
Chem. 1991, 95, 9385.
(29) Giamello, A.; Murphy, D.; Ravera, L.; Colcuccia, S.; Zecchina, A.
J. Chem. Soc., Faraday Trans. 1994, 90, 3167.
(30) Platero, E. E.; Spoto, G.; Zecchina, A. J. Chem. Soc., Faraday
Trans. 1 1985, 81, 1283.
(8) Tabata, K. J. Mater. Sci. Lett. 1988, 7, 147.
(9) Shimada, H.; Miyama, S.; Kuroda, H. Chem. Lett. 1988, 1797.
(10) Xie, S.; Mestl. G.; Rosynek, M. P.; Lunsford, J. H. J. Am. Chem.
Soc. 1997, 119, 10186.
(11) Acke, F.; Panas, I.; Sto¨mberg, D. J. Phys. Chem. B 1997, 101,
6484.
(12) Acke, F.; Panas, I. J. Phys. Chem. B 1998, 102, 5127.
(13) Acke, F.; Panas, I. J. Phys. Chem. B 1998, 102, 5158.
(14) Snis, A.; Stro¨mberg D.; Panas, I. Surf. Sci. 1993, 292, 317.
(15) Snis, A.; Panas, I.; Stro¨mberg, D. Surf. Sci. 1994, 310, L579.
(16) Snis, A.; Panas, I. J. Chem. Phys. 1995, 103, 7626.
(17) Snis, A.; Panas, I. Chem. Phys. 1997, 221, 1.
(31) Cerruti, L.; Modone, E.; Guglielminotti, E.; Borello; E. J. Chem.
Soc., Faraday Trans. 1 1974, 70, 729.
(32) Yanagisawa, Y. Appl. Surf. Sci. 1996, 100/101, 256.