8 ENVIRONMENTAL

ELSEVIER Applied Catalysis B: Environmental 13 (1997) 1-25

Review

The current state of research on automotive

lean NO, catalysis
A. Fritz, V. Pitchon*
LERCSZ, Laboratoire d’Etudes de la Rt!activitt Catalytique, Surfaces et Intelfaces, URA 1498 du
CNRS - Institut le Bel, Universite Louis Pasteur; 4, rue Blaise Pascal, 67070 Strasbourg, France

Received 15 May 1996; received in revised form 26 October 1996; accepted 30 October 1996

Abstract

A predominant challenge to
exhaust conditions. Despite a
credible and consistent diesel
researchers in their assessment
HC-SCR are discussed.
both industrial and academic research institutions is the selective reduction of NO, during lean
proliferation of publications, papers and reviews, a large number of catalysts fall short of a
application. Herein are reviewed systems designed specifically with the goal of assisting
of catalytic performance. By way of conclusion, three principal types of catalysts active for

Keywords: Lean NO,; Decomposition; Reduction; Zeolites; Noble metals; Oxides

1. Introduction these measures have proved less than sufficient and

so, to prevent pollution, an exhaust after treatment has
Nitrogen oxides (NO,) are emitted primarily become necessary. With this in mind, several technol-
from transportation and other industrial sources ogies have been developed in order to decrease the
and contribute largely to a variety of environmental harmful NO, emissions. They can be classified
problems: the formation of acid rain and the re- according to their use or non-use of a catalyst. The
sultant acidification of aquatic systems, ground-level aim of this paper is to review all possible catalytic
ozone (smog), and general atmospheric visibility candidates for the further development of this very
degradation. For these reasons, NO, emissions are necessary technology.
controlled.
In respect of the severe regulations imposed, the
motor car manufacturers have had to decrease con- 2. Origins and effects of nitrogen oxides
siderably the NO, emissions from their vehicles.
Initially, this was achieved either through action upon 2.1. Origins of nitrogen oxides
the nitrogen contents of the oil or by a modification of
the engine’s combustion chamber. More recently, The major source of nitrogen oxides is the combus-
tion of fossil fuels such as coke in electrical power
*Corresponding author. E-mail: pitchon@chimie.u-strasbg.fr. plants or petroleum in the engines of vehicles and

0926-860X/97/$17.00 0 1997 Elsevier Science B.V. All rights reserved.

PZZ SO926-3373(96)00102-6
2 A. Fritz, V PitchonIApplied Catalysis B: Environmental 13 (1997) l-25
aeroplanes. In both cases, the processes are so fast that
the thermodynamic equilibrium (C02+H20) is not
reached. For this reason, the exhaust gases contain
products of an incomplete combustion such as soot or
carbon monoxide. Nitrogen oxides are also formed.
Among them, nitrogen monoxide (NO) accounts for
95% of all nitrogen oxide emissions. From a thermo-
dynamic point of view, this compound is unstable
[l] (NO + ;Nz + iO2; A($ = -86kJ/mol). The
majority of NO is formed by the reaction between
nitrogen and oxygen following the mechanism estab-
lished by Zeldovich [2]:
N*+O* dNO+N*
N* + O2 + NO + 0*
N2+02+2N0
By assuming a constant concentration of N* radicals
and a large oxygen content when compared to NO
concentration, the following equation is obtained:
d[NO]/dt = 2 k exp( -EA/RZ’) [Nz] [O*]
Eq. (3) shows that the formation of NO is essen-
tially controlled by reaction (1). It also reveals the
importance of both temperature and atomic oxygen
concentration. The oxidation of organic nitrogen com-
pounds from oil is not significant from the point of
view of NO production [3].
2.2. Effects on health
In recent years, studies from within a variety of the
biological sciences have converged to reveal the
importance of NO to the living organism. One of
these reports [4] described NO as an essential mes-
senger, which transmits the necessary information to
the white corpuscles within the bloodstream to destroy
tumour cells and assorted bacteria and, in a second
role, to the neurotransmitters to dilate the blood
vessels. However, while the biologically active NO
is a poisonous product of the in vivo enzyme-catalysed
transformation of the amino acid, arginine, and is
maintained only at beneficent concentrations, that
emitted freely from a car engine presents a real hazard
of toxicity.
In the presence of air NO is oxidised to NO*, which
is a rapid conversion at high concentrations, further
accelerated in the presence of sunlight and hydrocar-
bon from unburned gasoline. This oxide is poisonous
for the respiratory system, provoking both lung infec-
tion and respiratory allergies since it is able to diffuse
through the alveolar cells and capillary vessels of the
lungs and damage their structure through their pro-
pensity towards oxidation [5]. Epidemiological stu-
dies have revealed that concentrations of nitrogen
oxides having hazardous effects for people in good
health are above 0.05 ppm for an exposure of over
24 h [6]. Nevertheless, this value is often exceeded in
towns with dense traffic during peak hours or during
the summer.
(1)
2.3. Impact on the environment
(2)
Nowadays, the deforestation of the northern hemi-
sphere by contamination is considered one of the most
important ecological problems. Although responsibil-
ity might be directed towards a variety of factors, that
of acid rain remains a major contributor among them
[7]. In this process, nitrogen oxides play an influential
(3) role in the photochemistry of both strotosphere and
stratosphere [8]. The chemical transformations of
nitrogen oxides are summarised in Fig. 1. Nitric oxide
is rapidly oxidised by ozone while radicals such as
OH* and HO2 are transformed into N02, HN02 and
H02N02 191. Acid rains usually form high up in the
cloud, a locality where nitrogen oxides react with
water, oxygen and other oxidants. These compounds
are transformed into HN03 which then acidifies rain,
Fig. 1. Chemical transformations of atmospheric NO,: (===>)
photochemical processes, (- >) thermal gas-phase processes,
(“*.a) dry deposition, and (- ->) heterogeneous reactions [5].
 A. Fritz, VI Pitchon/Applied Catalysis B: Environmental 13 (1997) 1-25

snow or fog because of its solubility in water. These

acid precipitations perturbate the aquatic ecosystems
greatly and can cause the biological death of lakes and
rivers.
Nitrogen oxides also contribute to photochemical
pollution by promoting the formation of ozone via the
reactions [ 10,111:
NO2 + hv(X < 415 nm) + NO + 0*
oz+o* +os

Other compounds such as peroxyacetylene nitrates

(PAN) formed by photochemical reaction have been
found [ 121. These are produced by a chain-reaction
process, described as follows:
NOz+hv+NO+O*
H20+0* --) 2OH*
co
CHs -CHO + OH* + CH3 -CO* + HZ0
Fig. 2. European norms for IDI diesel emissions (source: S.A.
CH3-CO* + O2 -+ CHs-CO; Renault).

CH3 -CO; + NO2 + CH3-C03-NO2

They contribute significantly to global photo-oxi-
dising pollution and have been shown to be more
harmful than ozone towards both human and plants.
Their principal source is the automobile.
2.4. Legislative aspect
As discussed above, NO is the starting material for
the production of both acid rain and photochemical
smog. As a result of these unfortunate ecological
consequences and continuing breaches of the legally
established threshold of NO, in air, a directive towards
renewed efforts from the motor car manufacturing
industry has been welcomed such that the NO emis-
sions are projected to comply with the EC-proposed
guidelines by the year 2000.
Since 1982, EC regulations concerning spark igni-
tion vehicles, more powerful purveyors of hydrocar-
bons (HC), NO and CO pollution than diesel, were
established to conform their engines to the extent of
unburned HC, CO and NO, legislated for. This was
achieved by the further development and implementa-
tion of exhaust devices and three-way catalysts. Reg-
ulation of diesel engines was introduced in 1988,
though mainly for particle emission, since this type
of engine was already in compliance with the estab-
lished gas-exhaust precedent. However, with the intro-
duction of a new series of EC objectives in 1993,
whose restrictions will become even more severe from
1996, it appears that the production of diesel vehicles
not in possession of an exhaust treatment will not be
possible in the very near future.
The norms represented in Fig. 2 for diesel vehicles
with direct injection show that the amount of
HC+NO, must be decreased, within seven years, by
a factor of 2.15 from 0.97 to 0.45 g/km.
3. A catalytic approach to NO, de-pollution
@eNOx)
3.1. The catalytic decomposition of NO
Nitrogen monoxide is thermodynamically unstable
(NO -+ $N2 + ;O z, AGF = -86 kJ/mol). Theoreti-
cally, it should decompose as soon as it is formed.
However, the decomposition reaction is inhibited by a
high activation energy of 364 kJ/mol [ 11. Therefore, a
catalyst is necessary to lower this activation energy in
order to facilitate this decomposition. Research in this
domain is extensive and this section will review the
various catalysts which can be divided into three
groups, i.e. precious metals, oxides and zeolites.
4 A. Fritz, V Pitchon/Applied Catalysis B: Environmental 13 (1997) 1-25

Table 1 of oxygen in the gas stream, the platinum is partially

Study of NO decomposition over noble metals (after Iwamoto and
oxidised and suffers a reduction in its activity. In the
Hamada [22])
absence of oxygen, the rate of decomposition is con-
Reference Catalyst stant since the molecular oxygen formed by decom-
1131 R wire position of NO is immediately desorbed.
[I41 Pt and Rh wire Wu et al. [21] have studied the effect of gold or
Cl51 Pt and Rh wire silver addition to a palladium-based catalyst. When
U61 Pt-Ni/AlaOa palladium is dispersed alone onto alumina, the results
u71 PtjAlaOa
are similar to those obtained with platinum: no activity
[181 PtlAlaOa
[I91 Pt/AlaOa, Pt foil occurs for temperatures lower than 773 K. The con-
r201 Ptwire version becomes substantial above 1023 K, at which
WI Pd-Au, AglAlaOa palladium oxide is decomposed producing some oxy-
gen. The addition of gold or silver leads to the for-
mation of Pd-Au or Pd-Ag alloys. The oxides of these
3.1.1. Precious metals alloys are decomposed beginning from 973 K and,
The majority of the decomposition studies have therefore, favour the NO decomposition at lower
been carried out either upon platinum metal alone temperatures, more than that with palladium alone.
or have engaged a bimetallic or alloy, both incorpor- Unfortunately, the catalyst inhibition due to oxygen
ating a second transition metal. Table 1 summarises has not been studied in this work.
their findings.
Studies upon the kinetics of the reaction have 3.1.2. Oxides
included experimental observation of parameters such The most active catalysts for NO decomposition are
as the partial pressures of oxygen and NO. The con- metallic oxide compounds. The principal studies are
clusions drawn from these enquiries are in support of summarised in Table 2.
the generally accepted mechanism which invokes a At a given temperature, the parameters and rate of
first step of NO adsorption and its subsequent in situ NO decomposition are found largely to depend upon
dissociation into nitrogen and oxygen atoms. Then, those of the oxygen desorption step. Also, a good
desorption of the newly constituted molecular nitro- correlation between catalytic activity and isotopic
gen and oxygen provides the sites for a new adsorp- exchange reaction between 02 and the surface has
tion: been observed [32], suggesting that the activity of
oxides as catalysts is strongly related to the strength of
NO (ads) + N (ads) + 0 (ads) the metal&oxygen bond in the lattice. The reaction
2N (ads) + NZ (g) shows strong similarities to NzO decomposition,
implying a first step of NO adsorption onto sites
2 0 (ads) + 02 (g)
formed by adjacent anion vacancies [23]. The overall
The desorption of oxygen from the surface is
strongly temperature dependant. Below 773 K, the Table 2
desorption is difficult. With a progressive covering NO decomposition studies on metallic oxides
of the surface by oxygen atoms, it becomes inactive
Reference Catalyst
for the adsorption of NO, until when the coverage is
complete so that the catalyst becomes totally inert. For ~231 40 metallic oxides
~241 Oxides of Cr, Fe, Co or Ni
the temperatures in the 973-1473 K range, Amimazmi
~251 BaO, SrO, CaO
et al. [ 191 have established the following rate equation: SrFeOa
1261
r = d[NO]/dt = k[NO]/(l + K[02]). ~271 BaaYCuaO,
The decomposition reaction has an order equal to 1281 La, Sr-Co, Mn, Fe, Cu-0,
unity towards NO and a negative order towards oxy- 12% Ag/Co304
I301 Rare earth oxides
gen; the reaction is therefore inhibited by oxygen. This
[311 Cuprates
might be explained by the fact that, with the presence
 A. Fritz, V Pitchon/Applied Catalysis B: Environmental I3 (1997) 1-25 5

kinetics of the decomposition reaction upon oxides
could be described by the same rate equation as that
for noble metals. An order equal to unity towards NO
is observed and the reaction is also inhibited in the
presence of oxygen. The rate-determining step
remains the desorption of oxygen.
Oxygen inhibition can be reduced by a judicious
choice of promoter. For example, Hamada et al. have
shown that, by adding silver to cobalt oxide, both
activity and resistance to oxygen poisoning are
increased [29], as depicted by Fig. 3. The mixed oxide
with an Ag/Co ratio of 0.05 is the most active in the
absence of oxygen. However, with only 5% of oxygen
and at 773 K, the activity of this oxide is reduced to
about its half, while the pure cobalt oxide becomes
totally inactive. The promoting effect of silver remains
the same for catalysts prepared by co-precipitation or
successive impregnations of a silver salt on cobalt
oxide. The catalytically active species seems to be a
complex of Co and Ag oxides. The enhancement
observed with silver might be due to its low affinity
for oxygen.
Perovskite-type oxides have also been investigated
for this reaction [26-28,311. The driving force in the
development of these catalysts was the finding that
defective perovskite-type oxides permit easy desorp-
tion of a large amount of oxygen from oxide bulk,
thereby implying that inhibition by oxygen would not
be as extensive. Another advantage is the stability of
their structure. Unfortunately, no significant activity
for the decomposition of NO under reaction condi-
tions approximate to automotive catalytic converters
have been reported thus far.
3.1.3. Zeolites
Since the pioneering work of Iwamoto et al. in 1972
[33] demonstrating, by use of TPD, that Cu-Y zeolites
could adsorb large amounts of oxygen which are then
desorbed at temperatures as low as 846 K, zeolite
studies have been extensively reported in the litera-
ture. From among a host of possible structures, Cu-
ZSM-5 is by far the most studied. This catalyst
exhibits a low but constant activity for NO decom-
position [34]. Valyon and Hall have compared the
desorption of oxygen on zeolites Y and ZSM-5 [35].
The Cu-ZSM-5 zeolite presents four peaks of deso-
rption at 393,493,683 and 803 K, while Cu-Y desorbs
oxygen only at 363 and 823 K (see Fig. 4) leading to
desorption ratios of 0.2 oxygen per Cu atom for Cu-
ZSM-5 and only 0.1 for Cu-I:
The zeolites prepared by ion exchange are more
active than those prepared by impregnation. Gener-
ally, the Naf cations of the Na-ZSM-5 zeolite are
exchanged for Cu2+ cations coming from a solution of
a cupric salt. The amount of sodium exchanged
depends upon the counter anion; with copper nitrate
or sulphate, the level of exchange could reach 100%.
The use of copper acetate allows an over-exchange of

0.05 0.10 0.15 273 473 673 10 20 30

Ag /Co (atomic ratio1 Temperature,K/Time,min
Fig. 3. Effect of silver on Co104 catalyst in NO decomposition Fig. 4. Temperature-programmed desorption of O2 from Cu-zeolite
WI. catalyst [35].
6 A. Fritz, V PitchonIApplied Catalysis B: Envimnmental 13 (1997) 1-25

loo
200% [36]. Such a high level of exchange was
explained by the formation of multivalent ion such
as CUDS+, Cu(OH)+, CU*(OH)~+ or Cu3(OH)F,
accompanied by the formation of a proton which acts
as a Bronsted acid site according to the following
equation:

xCt?+(s) +yH20 * CuX(OH);‘-‘(s) +yH+(s) 20

The stoichiometry of exchange between the sodium 0

and the copper ions was determined by absorption
spectroscopy. These types of experiment have proved
interesting since a direct correlation between the
exchange rate and activity has been established by
Iwamoto et al. [37], as shown in Fig. 5.
Two proposals have attempted to explain this phe-
nomenon, The first one implies that there are two
different types of cation site in the ZSM-5 framework
and that the one most easily occupied provided less
activity for the decomposition reaction. In the second
one, the reaction should proceed only if the two active
sites are adjacent, the probability of these sites’ exis-
tence being higher for high exchange rates. Also,
experiments in the presence of oxygen have shown
that over-exchanged catalysts are less inhibited by
oxygen than catalysts having an exchange rate lower
than 100% [36].
ErCnen et al. [38] have studied the effect of the
incorporation of copper within the zeolitic framework
during synthesis. By exchanging such a zeolite with
copper acetate, two types of cuprous cations are
0 20 40 60 80
Exchange level of Cu*+ / 92
Fig. 5. Influence of the exchanged level of Cu(II) ions on the
catalytic activity of zeolite [27].
173 373 473 573 673 773 873
Reaction temperature / K
Fig. 6. Temperature dependence of NO decomposition over Cu-
ZSM-5 [22].
obtained: those incorporated into the framework and
others introduced by exchange. They reported an
activity towards decomposition which is higher than
that on a zeolite where copper is introduced only by
ion exchange.
Fig. 6 displays the temperature dependence of Cu-
ZSM-5 for NO decomposition [22]. Optimum con-
version of NO is achieved at ca. 773 K. The loss of
activity above this temperature is not due to deactiva-
tion of the catalyst, but rather to the fact that the
desorption of NO from the surface becomes favour-
able.
The rates of conversion of NO and formation of
oxygen and nitrogen are not equal. This discrepancy
could be attributed to the formation of NO2 in the gas
stream before the catalytic bed by oxidation of NO in
the homogeneous phase. When NO is totally con-
verted, NO2 formation does not occur. Hence, the
catalytic reaction is really the stoichiometric decom-
position NO + iN2 + i 02. The kinetics of the
decomposition has been studied by Li and Hall
[39], who have described the rate equation as:
r = d[NO]/dt = k[NO]/(l + K[02]i’*),
The order towards NO is equal to unity while that
towards oxygen is negative. The exponent of the
oxygen concentration is 0.5 instead of 1 as in the case
of noble metals. The authors suggested that this
variation is the consequence of a difference in the
recombination of the adsorbed oxygen atoms.
100
In the same study, Li and Hall observed, by use of in
situ electron spin resonance, the reduction of a number
of cu*+ cations to Cu+ during the desorption of
oxygen. Liu and Robota confirmed this by in situ
 A. Fritz, V Pitchon/Applied Catalysis B: Environmental 13 (1997) I-25
XANES experiments which measured the transition,
ls-+4p, of Cut during NO decomposition [40]. The
activity towards decomposition is strongly linked to
the cuprous ion concentration, suggesting that they
play an important role in this reaction.
Valyon and Hall have compared the effect of reduc-
tion by carbon monoxide over the two zeolites, Cu-
ZSM-5 and Cu-Y [41]. With the former, an oxidising
pre-treatment favoured the presence of Cu2+ while a
reducing pre-treatment, those of Cut. Whatever the
pre-treatment, both reported the presence of both Cu+
and Cu2+. The Cu-Y zeolite did not behave similarly
as that following an oxidising pre-treatment: copper
existed only in the form of Cu2+; a subsequent treat-
ment under CO then reduced all the copper to Cu+.
The comparatively high activity of Cu-ZSM-5 could
be due to its encouragement of a co-existence of the
two copper cations which suggests that a redox
mechanism may be involved.
The reaction intermediates have been studied by
infra-red spectroscopy. Following the adsorption of
NO upon Cu-ZSM-5, there was formation of an NO&+
species (ca. 1900 cm-‘), in liaison with the Cu2+
cation and each of both N06- (ca. 1810 cm-‘) and
(NO),‘- species (ca. 1824 and 1730 cm-‘), with the
Cut population [42,43]. Iwamoto et al. have studied
the presence of these species over time by analysing
the gaseous effluent [44]. At room temperature, the
intensity of the bands attributed to NO&+ increased
with the passage of time, while those of NO&- and
(NO),&- species decreased. There was formation of
nitrogen and nitrous oxide so long as the negative-type
species continued to be detected. However, N2 and
N20 were no longer formed in the exclusive presence
of NO’+ upon the surface.
The principal results obtained from the Cu-ZSM-5
experiments may be summarised as follows:
?? the decomposition is stoichiometric into nitrogen
and oxygen;
?? the reaction is inhibited by oxygen;
?? the presence of over-exchanged copper increases
the activity;
?? the zeolite can desorb oxygen, during which
sequence some copper Cu2’ are reduced to Cu+;
?? the activity is correlated with the concentration of
cu+;
?? the concentration of NO&+-Cu2+ increases while
there is a parallel decrease in the concentration of
I
NO&--Cu+ and (N0)26--Cut during the forma-
tion of molecular nitrogen at room temperature;
and
?? the nitrogen formation stops when the concentra-
tion of NO&--Cu+ and (N0)26--Cu+ species
becomes constant.
From these findings, Iwamoto [44] proposed the
mechanism presented in Fig. 7. In this reaction
scheme, the Cu2+ cations exist before the reaction,
while the N06+-Cu2+ species formed following the
introduction of NO are not represented. The NO
adsorption on Cu+ cations leads to the formation of
species I, II, III and IV. Species I and II are too far
away to interact with each other and therefore remain
inactive, while III and IV can react at room tempera-
ture with neighbouring species to form nitrogen and
nitrous oxide. The oxygen atoms remain on the surface
to oxidise the Cuf ions into Cu2+(V, VI and VII),
which adsorb NO in their VII form. At this stage, the
catalyst is inactive for the decomposition reaction. By
an increase in the temperature, regeneration of the
active sites occurs, proceeding by the desorption of
oxygen which leads in turn to the reduction of Cu2+
ions to Cu+.
More recently, Lei et al. [45] have reported a
confirmation of this mechanism by their proposal of
the existence of a [Cu-0-Cu12+ complex which is
able to adsorb NO. The reduction of the chelated
species would induce the formation of Cut cations.
Shelef admitted some doubts about the adsorption
of NO upon Cuf cations of the d” configuration;
preferring the view that NO is adsorbed preferentially
upon CU*+ cations which are of the d 9 configuration.
Rather than the redox mechanism of Iwamoto, Shelef
proposed that Cu2+ ions are the active sites [46]. The
adsorbed species are under the form of gemdinitrosyls
which undergo rearrangement in order to give nitrogen
and oxygen. The efficient activity of Cu-ZSM-5 is
explained by the zeolitic structure containing tetra-co-
ordinated Cu2+ cations especially favourable for the
adsorption of the gemdinitrosyl species.
3.1.4. Conclusion on NO decomposition
The fact that NO decomposition was possible at low
temperatures has given rise to great enthusiasm
amongst scientists working in the field of diesel de-
pollution. As a result, various branches of catalytic
science (solid state chemistry, kinetics, mechanisms,
 A. Fritz, V Pitchon/Applied
a+
a -N,O t-N,)
a+
Fig. 7. Surface mechanism
. . .) have been investigated and numerous publications
during the last ten years have reported important
advances in general knowledge.
Unfortunately, the most promising systems for NO
decomposition, such as copper-based zeolites, do not
suppress the inhibiting effect of oxygen which remains
the single most important obstacle to overcome in the
development of a successful DeNO, catalyst. A further
difficulty to be considered is that posed by contact
with water; the resultant deactivation is often irrever-
sible as it affects the very framework of the zeolite
itself. With this in mind, several studies have been
instigated upon the subject of structural stabilisation
by the use of promoters (e.g. Ga, Ag, Co, . .). How-
ever, it has become clear that this reaction of NO
decomposition cannot be contemplated for a real
technological application in the presence of a complex
exhaust mixture.
An alternative is the selective catalytic reduction
(SCR) of NO,, to accomplish which a suitable agent
Catalysis B: Environmental I3 (1997) 1-25
NO
NO
of NO decomposition over Cu-ZSM-5 [44].
would need to be involved. The research undertaken in
this field provides the subject for discussion in the next
section. It is a vast field of study and a recent paper by
Tabata et al. [47] proposed an extensive review of the
patent literature associated with a discussion on the
problems to be solved for practical applications of the
SCR process.
3.2. The selective catalytic reduction (SCR)
3.2.1. Selective reduction by ammonia (NH3-SCR)
In this process, ammonia is used as the reducing
agent for nitrogen oxides to produce nitrogen and
water:
6NO2 + 8NHs + 7N2 + 12H20
Remarkably, these reactions are favoured in the
presence of oxygen [5]. Indeed, Fig. 8 shows that
 A. Fritz, B Pitchon/Applied Catalysis B: Environmental 13 (1997) 1-25 9

Despite a similarity in temperature range when com-

pared with an exhaust gas mixture, plus the positive
effect of oxygen upon the reduction, the NH,-SCR has
never been given a real application by the motor car
manufacturing industries. The toxicity and manipula-
tion problems of ammonia remain the major obstacles
preventing the use of this technology by the general
public, while those of NH3 corrosion and SO2 poison-
ing would be difficult to overcome from an engineer-
ing point of view.
Nevertheless, the NHs-SCR is widely used as a de-
pollution technique in conventional electric power
O2 concentration (96)
plants. A projection estimates that a third of the
Fig. 8. Effect of the oxygen concentration on NO conversion: (A) ammonia being produced in the United States by
V205/Ti02, (B) zeolite, (C) CuO/A1203, (D) V205/TiO&02, (E) the year 2000 will be for use as the reducing agent
Fe203/Cr203-A1203, and (F) V205 [5]. in this process [54]. In Japan, a pioneer country in the
development of this technology, more than 50% of the
power stations are equipped with such systems [55].
the addition of 0.1-l% of oxygen enhances the con- From a practical point of view, V205 is dispersed in
version. Above l%, the conversion remains stable. monoliths made of cordierite. A reduction of ca. 60-
A popular and efficient catalyst for this process is 85% is reached with an NHs/NO ratio of the range
vanadium pentoxide (V205) supported on oxides such 0.7-0.9, with the unreacted ammonia being partially
as Ti02, A1203 and SiOz [48]. The temperature of released. The NHs/NO ratio is maintained below the
activity for this catalyst varies in the 473-823 K range optimum since, despite the better conversion, a
according to conditions, i.e. temperatures comparable greater quantity of NH3 would be emitted to the
to those of a diesel exhaust gas. The active site seems atmosphere.
to be the (V=O)2+ complex on the catalytic surface
[49]. Two mechanisms have been proposed for this 3.2.2. Three-way catalysis
reaction. The first norms of de-pollution were established in
?? The Eley-Redeal mechanism which involves an the United States in the 1970s with limits upon
ammonia-adsorbed species reacting with NO in emissions of NO, CO and the unburned hydrocarbons
the gas phase [50,51], while oxygen regenerates being introduced. Initially, converters were made of an
the active sites as described: oxidation catalyst in order to transform CO and hydro-
carbons into CO2 and H20. By 1981, because of the
NH3 + s ---f [s-NH31
increase in the severity of the restrictions, a new
[s-NH31 + NO, + [s-H] + N2 + Hz0 catalytic system was developed. The three-way cata-
2[s-H] + +02 + 2s + Hz0 lyst (TWC) was able to eliminate simultaneously the
three pollutants: CO, NO and hydrocarbons [56].
?? The formation of a reaction intermediate of the Another system, the so-called dual-bed, comprising
nitrosamine type is postulated for the second a double catalytic surface, the first one for NO reduc-
mechanism [52,53]: here again, the role of oxygen tion and the second one for oxidising both CO and
is to regenerate the active site. hydrocarbons, had also been developed but was only
to present prominent weaknesses [57]. This system
NH3 + s + [H-s-NH21
required that the engine worked with a rich mode in
NO + [H-s-NH~] + [H-s-NH2 -NO] order to have a reducing exhaust gas composition on
[H-s-NH2-NO] -+ [s-H] + N2 + H20 the first bed and an additional air entrance on the
second in order to favour the oxidation processes.
2[s-H] +i02 --+ 2s+H20 Among its major faults was to encourage an increase
10 A. Fritz, V Pitchon/Applied
in the fuel consumption, while the efficiency of NO,
elimination is limited because of the production of
ammonia on the first bed, which was further trans-
formed into NO on the second bed.
At present, the choice is for a three-way catalyst
consisting mainly in a combination of noble metals
(Pt, Rh or Pd) deposited on an alumina carrier stabi-
lised by barium and lanthanum oxides. A consistent
amount of ceria is also present to play the role of an
oxygen reservoir in the rich transition and to help
obtain a well-dispersed metallic phase. The maximum
conversion is obtained for the three pollutants when
there is a stoichiometric ratio of oxidative and redu-
cing species in the exhaust. This is obtained by means
of a closed loop air/fuel control system. The reactions
involved are:
2N0 +2CO + N2 + 2C02
co +$02 --+ co2
C,H, + (X + $)02 -+ xC02 + $H20
The reaction of the three pollutants occurs as a func-
tion of the air/fuel mixture [58]. A deviation toward
enrichment will cause a decrease in the conversion of
CO and hydrocarbon because of the unavailability of
oxygen. Equally, with certain catalysts, the same
conditions promote a decrease in the NO+Nz
reduction because of the enhanced formation of
NH3 at high CO concentration, while deviation toward
the lean side causes a sharp fall off in the NO in this
reaction.
The fact that NO conversion decreases in oxidising
atmosphere -under conditions close to stoichio-
metric, as observed in three-way catalysis- has dis-
couraged the research on noble metals in a large
excess of oxygen - as with diesel conditions. These
catalysts have an improved activity when additional
hydrocarbon is injected even in a very oxidising atmo-
sphere (see following sections).
3.2.3. Selective reduction by hydrocarbons
(HC-SCR)
As discussed previously, while Cu-ZSMS is very
active for NO decomposition, it loses this efficiency in
the presence of oxygen. In 1990, Held et al. [59], and
Iwamoto [60], independently discovered that either
alkanes or alkenes are able to reduce NO on Cu-
Catalysis B: Environmental 13 (1997) 1-25
ZSMS, not only in the presence of oxygen, but also
that the reaction was favoured in an O2 excess. Fol-
lowing this discovery, much research upon the selec-
tive reduction of NO, by hydrocarbons was
undertaken and subsequently reported in the literature.
The active catalysts can be divided into three groups:
zeolites, metallic oxides and noble metals.
3.2.3.1. Zeolites (Cu-ZSMS)
3.2.3.1.1. Effect of the various parameters. As for the
case of catalytic NO decomposition, the exchange of
the Na+ or Hf for Cu2+ cations was the preparation
method which gave the best results [61]. The influence
of the main parameters on the catalytic activity, which
has been studied by Iwamoto et al. [62]. Fig. 9 shows
the effect of temperature upon the selective conversion
of NO into N2 when using the Cu-ZSM-5 catalyst.
When the gas stream contains 1000 ppm of NO, the
catalyst is only 25% efficient at 673 K. The reaction
can be completely inhibited by 1% oxygen in the
reactant feed. However, if 166 ppm of propene is
added to the stream, 100% of reduction is obtained
at 873 K. With 1% oxygen, the conversion presents a
volcano curve shape and the reduction is of ca. 30% at
573 K. If the concentration of propene is increased up
to 1000 ppm, the conversion of NO reaches 85% at
573 K. Clearly, the presence of oxygen promotes the
reaction for temperatures lower than 773 K. Above
e
. IO0
u IJ
/
z
Temperature / K
Fig. 9. Conversion into nitrogen over Cu-ZSM-5 as a funtion of
temperature: (0) 1000 ppm NO; (A) 1000 ppm NO + 1% 02; (0)
1000 ppm NO + 166 ppm C3Hs; (0) 1000 ppm NO + 1% O2 +
166 ppm C3H6; and (A) 1000 ppm NO + 1% O2 + 1000 ppm
C3Hs [Ql.
 A. Fritz, V! Pitchon/Applied Catalysis B: Environmental 13 (1997) 1-25 11

0’
1.

Oxygen concentration I I
Fig. 10. Conversion into N2 over Cu-ZSM-5 as a function
oxygen concentration [62].
0 100 200
of the Exchange level of copper ion I 46
Fig. 11. Dependence of the catalytic activity of copper ion-
exchanged ZSM-5 zeolite on the exchange level of copper ion [63].

this temperature, the NO reduction decreases because

all of the hydrocarbon is consumed by the combustion
reaction (HC+O*). while the latter react preferentially with oxygen. For
The influence of oxygen concentration at 603 K is example, with Cu-ZSM-5, ethene, propene and butane
clearly represented by Fig. 10. In the absence of are selective while hydrogen, carbon monoxide or
oxygen, the activity is almost nil; yet by adding 2% methane are non-selective agents.
of oxygen, the conversion into nitrogen is largely Fig. 12 shows the influence of propene concentra-
increased. This concentration represents the optimum, tion upon NO reduction at 573 K. The NO conversion
after which the conversion decreases. increases linearly with a corresponding increase in
Another factor of interest is the rate of replacement propene concentration in the O-500 ppm range, but
of the zeolite’s cations by Cu*+ since there is an more gradually for still higher concentrations. Cho has
increase in activity in the decomposition reaction compared the reducing activity of ethene and propene
for an exchange greater than 100%. Sato et al. [63], [64], showing that the optimum reducing agent should
have tested several Cu-ZSM-5 zeolites with various have a temperature of half conversion close to that of
exchange rates of copper. The results are expressed in NO.
Fig. 11, and the conclusions are somewhat different
from those of the decomposition reaction, in which no
formation of nitrogen was observed for an exchange
lower than 40%. On the other hand, the rate of
exchange appears to have a positive effect on the
SCR reaction, but an over-exchange becomes detri-
mental. Maximum activity is observed for a rate of
exchange in the 80-100% range. However, it is nota-
ble that the activity is never zero even with 0% of
copper, suggesting that the activity of a given catalyst
for HC-SCR cannot be compared with that which it
would have for NO decomposition. In this sense,
reduction and decomposition are two distinct phenom- 0 200 400 600 SO0 IO00
ena implying different mechanisms.
C,H, concentration I ppnt
The reducing agents for SCR can be classified into
two groups; selective and non-selective agents. The Fig. 12. Conversion into N2 over Cu-ZSM-5 as a function of the
former reduce NO even in the presence of oxygen, propene concentration [62].
12 A. Fritz, R Pitchon/Applied
Radtke and Baiker have studied the products
formed during the catalytic reaction [65] and have
discovered that, in addition to nitrogen, undesirable
by-products such as N20, hydrogen cyanide (HCN) or
HCNO are formed under certain experimental condi-
tions. It is worth noting that under such oxidative con-
ditions, CO is not formed but a fraction of NO is always
converted into NO2 as a result ofagas phase process
[66]. The amount of NO2 forme depends on several
parameters such as the geometry I of the apparatus, the
space velocity or the NO/O2 ratio and could, therefore,
be very different from one laboratory to another.
3.2.3.1.2. In situ characterisations. The oxidation
state of copper during the course of the reaction has
been the subject of several in situ studies. Liu and
Robota have attempted to characterise it using
XANES spectroscopy [67]. In the presence of a
selective reducing agent such as propene, a
significant proportion of the copper is in the form
of Cu+, even in a large excess of oxygen. During the
course of the reaction, the concentration of cuprous
ions follows the same trend with temperature as that of
NO conversion. Moreover, the use of a non-selective
reducing agent such as methane was found not to lead
to the formation of Cu+ ions, all the copper being of
cupric-ion character. The authors concluded that
cuprous cations are essential for the reaction.
Following an FTIR study [68], Pieplu et al. were in
agreement with the above findings, while a paper from
Kucherov et al. reported to the contrary [69]. By
monitoring the reaction using an ESR technique,
they reported that all of the copper is in the cupric
form and then suggested that cations with an
unsaturated quadratic planar coordination might be
the active sites [70]. A further FTIR study by Kutyrev
et al. demonstrated a close correlation between the
catalytic activity and the concentration of isolated
cu2+ species [71]. They proposed a complex of
[Cu2+-O-N=01 as a reaction intermediate.
With such diverse studies, it is impossible to draw a
definite conclusion. However, the differences might be
explained by the variety of the characterisation tech-
niques employed and by the disparity in the experi-
mental conditions.
3.2.3.1.3. Reaction mechanism. The mechanism has
not yet been fully elucidated. An extensive paper
Catalysis B: Environmental 13 (1997) 1-25
reviews the principal features of the reaction over
Cu-ZSM-5 and discusses the mechanistic con-
clusions that different authors have proposed [72].
Several different examples have been suggested in
the literature, the most plausible ones being given in
the following paragraphs.
(a) The first mechanism proposed took, as its start-
ing point, the same as that for NO decomposition with
the zeolite, Cu-ZSM-5, i.e. that NO is adsorbed onto
Cu+ cations and decomposed into in situ molecular
nitrogen and Cu2+ -coordinated oxygen. The hydro-
carbon is then able to regenerate the active site by a
fast reaction with the surface oxygen [70,73,74]. The
mechanism can be represented by the following
scheme:
2 NO-Cu+ + N2 + 2 0-Cu2+
90-Cu+C3H6+9Cu++3C02+3H20
The increase in catalytic activity in the presence of
oxygen, according to Burch and Millington, is due to
the ability of Cu-ZSM-5 to avoid complete reduction
to metallic copper, as Cue would be inactive for the
reaction [75]. Cho has studied the reaction under
transient conditions [76] and suggests that
the hydrocarbon is adsorbed onto the zeolite at low
temperature, thus preventing the adsorption and sub-
sequent reduction of NO. Therefore, the catalyst is
only operational when the three species - hydrocar-
bon, NO and oxygen -have access to the active sites.
When hydrocarbon combustion by oxygen is
complete, the active sites are no longer reduced and
the conversion of NO is stopped.
(b) The formation of a carbonaceous deposit on the
surface of the catalyst is the key step in another
mechanism [77-791. This deposit is able to reduce
NO as follows:
xC3H6 + C,H,-S
C,H,-S + 2 NO + N2 + CO2 + H20
The effect of oxygen here is to increase the con-
centration of adsorbed NO species upon the catalytic
surface leading to a more efficient conversion.
(c) Several authors affirm that the oxidation of NO
to NO2 would be the first, or preliminary, step of the
overall reduction [80-831, then the nitrogen dioxide,
more active than NO, reacts directly with the
 A. Fritz, V Pitchon/Applied
hydrocarbon:
1
NO+-02+N02
2
Here, the oxygen plays a triple role:
?? oxidising NO to NO*,
?? by burning the carbonaceous deposit, it maintains
a clean surface, and
?? preventing an excessive reduction of copper ions
into metallic copper.
(d) A reactive intermediate formed by a partial
oxidation of the hydrocarbon adsorbed onto the sur-
face is postulated by certain other authors [66,84,85].
This intermediate should react directly with NO as
follows:
C3H6-S +; O2 -+ C30H6-S
C30H6-S + 8 NO-S + 4N2 + 3 CO2 + 3 Hz0
The exact structure of the oxygenated intermediate
is not described in the literature. Nevertheless, com-
pounds such as aldehydes, alcohols, ketones or car-
boxylates have been proposed. To answer to this
question, Montreuil and Shelef have used a very
interesting approach [80]. The have used oxygenated
organic compounds (alcohols, aldehydes and ketones)
as reducing agents and found that they are effective for
the SCR, supporting the hypothesis that a partially
oxidised hydrocarbon is a reaction intermediate.
3.2.3.1.4. Bench test evaluation. The activity of Cu-
ZSM-5 deposited on a monolith has been evaluated on
an engine bench [86,87]. The exhausts from lean-burn
spark ignition engines and from a diesel engine were
used in these tests. Generally, in respect of the effects
of both temperature and oxygen or hydrocarbon
concentrations, the results obtained with monoliths
are similar to those obtained with powder on a
laboratory scale. The zeolite was found to convert
NO for a range of air/fuel ratios larger than that
was seen with three-way catalysts, which are only
active for an exhaust having a composition of about its
stoichiometry. The zeolite activity could be improved
by increasing the hydrocarbon concentration in the
exhaust mixture; however, in this case, the catalyst
deactivates rapidly.
Catalysis B: Environmental 13 (1997) 1-25 13
a
TemperaturerC
Fig. 13. Nitric oxide conversions of Cu-ZSM-5 in a simulated gas
exhaust: (a) unheated, (b) heated at 600°C (c) heated at 7Oo”C, and
(d) heated at 800°C for 5 h [88].
The Cu-ZSM-5 deactivation is represented by
Fig. 13 [88]. With a fresh catalyst, the conversion
of NO reaches almost 45% at 773 K, while pre-treat-
ment under the reacting mixture for 5 h at 873 K leads
to a decrease of the conversion to ca. 30%. For higher
pre-treatment temperatures, i.e. 973 or 1073 K, the
activity falls to less than 10%. That the temperature of
optimum conversion remains almost unchanged by the
various treatments is remarkable.
Within the literature, there are differences of opi-
nion with regard to the reasons for the ageing of the
catalyst. By use of 27A1 and 29Si NMR, Tanabe et al.
have observed a de-alumination of the framework of
the zeolite during the reaction process [88]. Also, in
this paper, an ESR study demonstrated that the sinter-
ing of copper does not occur, but it remains dispersed
even after the deactivation. However, a difference in
the behaviour towards NO and propene was reported.
Apparently, the fresh zeolite interacted with these
gases whereas, on the deactivated solid, they did
not adsorb. The authors’ conclusion was that the
copper ions migrate during the ageing procedure
towards sites which are not accessible to NO and
propene. This migration, a model of which is repre-
sented in Fig. 14, is thought to be induced by the de-
alumination process.
In the fresh zeolite, the copper is located in the main
channels and is accessible to the reactants. However,
the de-alumination due to ageing leads to the migra-
tion of copper towards smaller sites ringed by only five
14 A. Fritz, V Pitchon/Applied
inaccessible site
Main Channel
accessible sit
Fig. 14. Schematic representation of the (100) face of ZSM-5
showing a ten-membered ring, accessible to adsorbates and a five-
membered ring inaccessicle to adsorbates [SS].
atoms. A paper from Matsumoto et al. offered similar
conclusions [ 891.
Kharas et al. observed by XRD, the sintering of
copper into CuO associated with a loss of crystallinity
of the zeolitic structure [90]. According to the authors,
the growth of the copper oxide crystallites is made to
occur on the internal surfaces of the zeolite. When
they exceed a critical size, imposed by the structure of
the zeolite, the copper sintering induces a local
destruction of the structure which is considered
responsible for micropore loss.
By employing in situ ESR, Kucherov et al. [91]
have monitored the changes of the state of copper in
the framework of the zeolite induced by thermal
ageing. According to their report, the exposure of
CuH-ZSM-5 to a high-temperature reducing environ-
ment causes fast reduction of copper possibly with the
formation of Cue crystallites leading to deactivation.
Indeed, they have shown that square-planar isolated
cupric ions are the most active species and, therefore,
such conditions lead to a strong deactivation.
3.2.3.2. Zeolites other than Cu-ZSM-5
3.2.3.2.1. Cu-exchanged zeolites. Gopalakrishnan et
al. have compared the catalytic activity of several
zeolitic structures, each containing 1% of cu
introduced by exchange [92]. Under the conditions
of operation, Cu-ZSM-5 is the most active catalyst
with a 90% conversion at 673 K, followed by Cu-
mordenite with 63% at 703 K. The activity of Cu-X
Catalysis B: Environmental 13 (1997) I-25
and Cu-Y is less than 10%. A study by Heimrich and
Deviney on zeolites dispersed on a monolith has
confirmed this range of activity [93]. Coq et al.
have studied, by TPR, the dispersion of copper in a
mordenite as a function of the rate of Cu exchange
[94]. For rates lower than 53%, CL?+ ions are isolated
and the volume of the micropores does not change,
while for higher rates, some copper clusters are
formed and the volume of the micropores decreases.
The isolated Cu*+ ions seem to be more active than the
clusters for NO reduction by decane. Centi et al. have
prepared a boralite which has the same structure as the
ZSM-5, only replacing A13+ cations with B3+ [95].
The boralite exchanged by copper presents the same
conversion as the Cu-ZSM-5, but the temperature of
the maximum conversion is decreased by 30 K.
3.2.3.2.2. Pt-ZSM-5. Hirabayashi et al. have
compared the activity of Pt-ZSM-5, Cu-ZSM-5 and
Fe-MOR at temperatures lower than 573 K [96]. The
results are displayed in Fig. 15, where it is clear that
the zeolite containing platinum is more active than the
other catalysts for this range of temperature with a
conversion of 18% at 473 K.
An interesting property of the Pt-ZSM-5 is its sturdy
resistance to steam deactivation. Indeed, the addition
of water at 485 K to the gas stream has no influence on
its performance, whereas the activities of Cu-ZSM-5
and Fe-MOR decrease as indicated in Fig. 16. The
major problem with Pt-ZSM-5 is its preference to
conversion into N20: at 473 K, 36% of NO is trans-
formed into N20 while only 18% is converted into N2.
Iwamoto et al. have evaluated the activity of Pt-
ZSM-5 in the form of pellets introduced into a diesel
exhaust line [96,97]. The above catalyst led to a 33%
reduction of NO, at 523 K, while a Cu-ZSM-5 catalyst
under the same experimental conditions reduced only
27% of NO, at 673 K. This activity can be enhanced
by the addition of ethene to the exhaust-gas stream.
However, under the same conditions, the formation of
N20 is once again significant.
3.2.3.2.3. Protonated zeolites. The protonated
zeolites have undergone intensive study by the
group of Hamada. In the presence of propene and
with a large excess of oxygen, the mordenite is more
active than the H-ZSM-5 and H-Y zeolites [98].
Methanol and ethanol are both selective reducing
 A. Fritz, V Pitchon/Applied Catalysis B: Environmental I3 (1997) l-25 15

experimental conditions, NOz is more easily

reduced than NO. The deduction being that NO is
primarily oxidised into NOz which is then
subsequently transformed into N2 by reduction.

3.2.3.2.4. Zeolites containing gallium. The catalytic

activity for the SCR of NO by the gallium-exchanged
zeolites has been reported for the first time by the team
of Kikuchi [loll. With experimental conditions
(1000 ppm NO, 1000 ppm C3H6, 10% oxygen), Ga-
ZSM-5 exhibiting 91% of conversion at 673 K
revealed itself to be more active than Cu-ZSM-5
which gave only 60% at 573 K in the same
experimental conditions. The reduction proceeds in
the presence of ethene as reducing agent [ 1021.
Reaction temperature /K
In contrast to Cu-ZSM-5, Ga-ZSM-5 zeolite cata-
Fig. 15. Temperature dependences of the catalytic activity of (0) lyses the SCR of NO by methane [103,104]. This
Pt-MFI, (A) Fe-MOR and (0) Cu-MFI in the absence of water reduction, taking into account that neither Ga-H-Y nor
vapour [96]. Ga-Na-ZSM-5 are able to reduce NO with methane,
seems to be encouraged by a synergetic effect between
r cl 1 the Ga3+ and H+ cations and the zeolitic framework.
a Recently, Feeley et al. have proposed a sulphated
Ga-Zr-ZSM-5 zeolite which could be active with
reducing agents such as propane, propene and
methane [ 1051.

.-
g 00 3.2.3.2.5. Zeolites containing cobalt. In contrast to
the zeolite containing copper, those exchanged with
6 cobalt are able to reduce NO selectively with methane
“i 10
$ in the presence of oxygen. In this context, structures of
r’ ZSM-5 type [106,107], mordenite [108,109], and
ferrierites [llO,lll], have been studied. Armor et
e 0
al. have been able to draw many conclusions from
their work including the discovery that Co-ZSM-5
selectively catalyses the simultaneous reduction of
NO and N20 in the presence of oxygen [112]. Also,
a detailed kinetic study of the effect of water vapour
‘0 4 8 12 16 20 upon SCR indicated inhibition of the reaction at low
Time I h temperature, but an increase in the selectivity towards
Fig. 16. Effect of water vapour (8.6 ~01%) on the catalytic NZ. This effect proved reversible upon the exclusion of
activities of (a) Pt-MFI, (b) Fe-MOR and (c) Cu-MFI, for the SCR water from the system, whilst the catalytic activity
of NO at 485 K [96]. remained stable under wet conditions.

3.2.3.2.6. Zeolites containing cerium. Zeolites

agents with H-ZSM-5, while with Cu-ZSM-5 only the exchanged with cerium have been studied by the
latter is able to selectively reduce NO [99]. The group of Misono. Under the operating conditions
mechanism has been considered by Hamada et al. reported, Ce-ZSM-5 proved more active than Cu-
in one of their publications [ 1001. Under their ZSM-5 with 85% conversion and 55%, respectively,
16 A. Fritz, V PitchonIApplied Catalysis B: Envimnmental I3 (1997) 1-25

at 573 K [ 1131. The activity can be further increased authors have monitored the catalytic behaviour of
by the addition of alkaline earth metals such as MnO promoted by lithium [122].
strontium [ 1141. A mechanical mixture of Ce-ZSM-
5 and MnaOs was found to be more active than Ce-
3.2.3.3.2. Supported transition metal oxides. Hamada
ZSM-5 alone [ 1151. By employing MnaOs as an
et al. have compared the activity of several metal
oxidising agent for NO, Misono was able to
oxides (Cu, Co, Ni, Mn, Fe), supported on alumina
propose a mechanism whereby MnaOs oxidises NO
or silica [123,124]. In general, the alumina-based
to NOa and the catalyst Ce-ZSM-5 subsequently
catalysts prove better than those supported on silica.
reduces NOa into nitrogen. A similar mechanism
Their activity is strongly dependent upon both
was presented for Ce-ZSM-5 alone [116]; the first
preparation technique and thermal treatment.
step remained the oxidation of NO into NOz. The
Catalysts containing aluminates are more active.
latter then reacts with the hydrocarbon to form
Hamada has suggested that the oxidation of NO
compounds of nitro type which decompose into
into NOa is the first step of the mechanism on this
nitrogen or N20 via several intermediates.
type of catalyst.
The use of zirconia as a support for copper, nickel or
3.2.3.3. Oxides
cobalt is reported by Bethke et al. [125]. These
catalysts are less active than Cu-ZSM-5. Their study
3.2.3.3.1. Simple oxides. The oxides such as AlaOs,
by in situ FTIR proved the existence of cyan0 and
SiOa-AlaOs, TiOz or ZrOa are active for the selective
isocyanate species [ 1261.
reduction of NO [ 1171. Alumina is the most active
Miyadera has studied the activity of several transi-
with a reduction of 32% at 573 K (1000 ppm NO,
tion metals (Cu, Co, Ag, V, Cr), supported on alumina
330 ppm C3H6, 10% 0,). Pure silica is not at all active.
[ 1271. The catalyst containing silver is the most active
Ammonia-TPD experiments revealed that the amount
under reaction conditions giving a conversion of 80%
of desorbed ammonia can be correlated with the
at 673 K. With these catalysts, compounds such as
catalytic activity, leading the authors to conclude
ethanol or acetone are more efficient reducing agents
that acidity was a determining parameter. To
than propene. The maximal activity of Cu/AlaOs can
improve the acidity, A120s, TiOz, ZrOa or FeaOs
be achieved at lower temperatures by adding caesium
were treated with sulphuric acid [118]. This
[128] which, it is supposed, increases the electronic
treatment had differing effects upon the oxides:
density of copper, thereby, enhancing the reactivity of
while alumina became less active, the performances
adsorbed isocyanate species on Cu with NO [129].
of the other three solids improved significantly. The
Ueda et al. [ 1301 showed that the reduction of NO with
alumina deposited on a monolith was tested on a diesel
propene in the presence of an excess of oxygen and
exhaust line [119]. In the absence of additional
moisture takes place over gold supported on y-AlaOs.
injected hydrocarbon, the activity was almost nil.
The addition of MnsOa to Au/AlaOs appreciably
However, the conversion of NO reaches 70% at
increased the maximum of conversion of NO to Na
673 K if an addition of 3000 ppm methanol is
as well as the conversion, at lower temperature.
added to the exhaust gas. The catalytic activity of
alumina can further be improved by the addition of
copper. In this case, the temperature of the maximum 3.2.3.3.3. Perovskites. Furukawa et al. have reported
conversion decreased while there was a simultaneous the activities of oxides of perovskite type for the SCR
increase in activity [ 1201. The optimum concentration of NO by propene [ 13 11. As shown in Table 3, the
of copper is 0.3%. A similar enhancement was perovskites containing cobalt, copper or iron are
confirmed in the case of Si02-AlaOs. Zhang et al. active for the decomposition of NO, but do not
have studied the reduction of NO by methane on rare catalyse the SCR by propene. In these cases, the
earth oxides such as LaaOs, CeOz, NdaOs, Smz,03, only reaction to occur is the combustion of the
TmzOs and LuaOs, with strontium as promoter [ 1211. hydrocarbon. However, when the perovskites have
The Sr/La20s compound gave the highest activity in only a little combustion activity, such as LaAlOs,
both the presence and absence of oxygen. The same they are active for the SCR.
 A. Fritz, V Pitchon/Applied Catalysis B: Environmental 13 (1997) 1-25 17

Table 3 Finally, many researchers proposed the existence of

Activity of perovskites: T = 873 K, NO = 4400 ppm, and 02 =
oxygen-containing reaction intermediates [ 126,134].
4.4% [131]
This idea is supported by the fact that compounds such
Catalyst Conversion into N2 (%) as alcohols, ethers or ketones are active as reducing
LaA103 14 agents. For example, K. Matsuda et al. have demon-
La0.&0.2AlO3 21 strated the activity of methanol with an Ag/AlzOa
LaAhMgo.lO3 14 catalyst and T. Miyadera et al. have evidenced ethanol,
L~10.900.0503 3 acetone or propanol on a wide range of supported
l,a&.PPC% 0103 3
transition metals.
SrSnOs 8
Ca-%.8Co 0.203 3
CaTi03 11 3.2.3.4. Noble metals
CaTi 8Fe0.203 0
3.2.3.4.1. Effect of various parameters. Obuchi et al.
proposed a study of the SCR of NO over noble metals
3.2.3.3.4. Reaction mechanism. The performances of supported on alumina [ 1351. The dependence of NO
single metal oxides, sulphate-promoted metal oxides conversion upon temperature is shown in Fig. 17.
and binary metal oxides suggest that solid acidity or Platinum, rhodium and palladium gave a conversion
basicity is an important factor affecting catalytic at 623,573 and 573 K, respectively. The conversion of
activity. In order to explain the formation of NZ, NO is equal to 25% at 623 K and only 10% at 573 K,
three proposals exist in the literature. Obuchi et al. with iridium. The selectivities towards nitrogen are
[ 1321, using alumina, observed an oscillating given in Table 4.
behaviour which they have attributed to the With alumina-supported platinum there is a selec-
temporal accumulation and removal of a highly tivity towards nitrogen of only about 30%, while the
reactive carbonaceous deposit. They explained that other noble metals are more selective with over 75%.
a fast reaction takes place between NO and propene, The effect of the metal loading has been studied by
resulting in NO conversion and the deposit Burch et al. with supported alumina catalysts contain-
accumulation. Then, NO further reacts with the ing O.l-2% of platinum [137]. The increase of the
deposits, first slowly in radical deficient metal loading has a beneficial effect on both the
circumstances. As the deposits increase, NO
consumption increases and radical reactions begin
to occur, causing acceleration of deposit oxidation. I I 4 I 1

The overall reaction is then repeated and, therefore,

explain the occurrence of oscillations.
The group of Hamada [133] have investigated the
effect of oxygen on HC-SCR over Cu-ZSM-5, H-
ZSM-5 and alumina. A comparison of the reaction
with hydrocarbon oxidation reaction has led them to
conclude that the first step of the reaction was the
oxidation of NO into NOz over H-ZSM-5 and alumina
and partial oxidation of hydrocarbon over Cu-ZSM-5.
They mentioned three important parameters for the
reaction to occur: (i) activation of hydrocarbon, (ii)
activation of oxygen and (iii) selective reaction of
intermediate with NO, in preference to oxygen. The 150 200 250 300 350 400 450
first two factors are related to acidity and to the Temperature / “C
presence of transition metal oxides, while the third
Fig. 17. Catalytic activity of noble metals: (0) Pt, (0) Pd, (A)
one determines the NO reduction selectivity in an Ru, (W) Rh and (0) Ir. NO, 1000 ppm; C3H6, 870 ppm; and 02,
oxidising atmosphere. 5% [135].
18 A. Fritz, V Pitchon/Applied Catalysis B: Environmental I3 (1997) 1-25

Table 4
Selectivities of alumina supported noble metals [136]

Catalyst Temperature (K) NO conversion (%) Nz (%) Selectivity in Ns (%)

1% Pt/AlaOs 623 53 16 32
1% Rh/Al*Os 573 50 43 86
1% Pd/AlaO, 623 25 20 80
1% Ir/AlaOs 573 10 8 80

Table 5
Influence of the nature of the support upon catalytic activity [136]

Catalyst NO conversion Nz (%) Selectivity in

(%) NZ (%)

1% lWTiOa 13 7 55
1% Rh/ZnO 21 10 53
1% Rh/zrO, 33 21 68
1% Rh/Al,O, 50 43 86

have both the effect of making the particle size of the

active material smaller along with the ability to sup-
press sintering during thermal ageing. Giving an NO,
reduction of 51%, it has a practical application in the
Japanese market.

3.2.3.4.2. Proposed mechanisms. (a) Burch et al. have

studied SCR of NO by propene on 1% Pt/A1203 using
the temporary analysis product technique (TAP) [ 1391.
The mechanism proposed from the findings is shown
in Fig. 19. This two-step mechanism is initiated by
250 300 350 one molecule of hydrocarbon reducing a small region
of surface Pt-0 into metallic Pt. On these newly
Temperature / “C
prepared sites, the adsorption/dissociation of NO
Fig. 18. Influence of the platinum loading on the DeNO, activity can now proceed. The rate of NO dissociation is
of Pt/AlaOs: (+) 2%, (A) l%, (0) 0.5%, (W) 0.25% and (*) 0.1% temperature dependent. At lower temperatures, NO
[137]. is adsorbed molecularly, but only a fraction of it is
dissociated. Some NO will recombine with a N atom
temperature of the maximum conversion and the which explains the formation of NaO. However, as the
activity, as shown in Fig. 18. temperature increases the dissociation into N and 0
In Table 5, the important influence of the nature of atoms occurs and the formation of N20 is decreased
the support upon catalytic activity is presented. The because the most favoured reaction becomes the
alumina-based catalyst gave the highest NO conver- recombination of two adsorbed N atoms to give
sion and, therefore, resulting in nitrogen production. molecular nitrogen. The key step of the mechanism
Recently, a new type of catalyst for the lean-bum is the competition between nitrogen monoxide and
engine was developed by researchers at Mazda Motors oxygen for the reduced sites. In a further paper, the
[ 1381. This catalyst has a base support material of MFI rate of NO decomposition was investigated for this
zeolite loaded with active metal including platinum, catalyst and compared with that for the DeNO,
iridium and rhodium. This composition is thought to reaction with propene [140]. The authors concluded
 A. Fritz, V Pitchon/Applied Catalysis B: Environmental 13 (1997) I-25 19

000000
I I I I I I
Pt Pt Pl Pt P1 P,

1 HC

Fig. 20. Mechanism of HC-SCR on Pt/Si02 [141]

0 0
0 N N 0 experiments using the diffuse reflectance FTIR tech-
I I I I
Pt PI PI P, pt PI nique have investigated the surface intermediates.
Amongst the species discovered were adsorbed hydro-
carbons, isocyanates, cyanides and nitrosonium
cations. Of the supports examined, alumina gave
0 the best activity, selectivity towards nitrogen
0 0 N N 0 OONNOO
I I I I I I I I I I I (Table 5), and also the highest surface species
R R PI h PI PI Pl PI PL R Pt PI
concentration. The catalytic activity and the IR
1 -I$0 1 -Nz
intensities corresponding to the isocyanate and nitro-
0
I
0
I
0
I
0 0 0
I
0
I
sosnium species increase with temperature to reach
1 I
FTRPlPlPtR hRRRPtPt their maximum in the 573-623 K range, following
14 JO* which they decrease as temperatures are raised
000000 000000 (Fig. 21).
I I I 1 I I I I I I I I
PtRRRRR RRRRRPI This correlation between activity and IR intensity
led the authors to propose the following mechanism:
Fig. 19. Mechanism of NO reduction by hydrocarbons on P&based
catalysts [137].
the first step being the formation of an intermediate
species containing one oxygen atom [142], which
could react with adsorbed NO to form an intermediate
which then decomposes into fragments, each having
that the rate of NO decomposition over reduced Pt was
an isocyanate structure. Then, the -NC0 group reacts
suffkiently fast under lean DeNO, conditions to
with NO, NO2 and the adsorbed oxygen to give
confirm the above mechanism, while an estimate of
the fraction of Pt atoms reduced during the reaction
was deemed to be ca. 14%.
0.12
(b) Tanaka et al. have carried out mechanistic
experiments using FTIR on platinum deposited upon
0.10
silica [141], and reported the presence of organic
species of nitro, nitrite and carbonyl upon the surface. 0.08
These species can react rapidly with oxygen and NO2 g
b
to give N2, N20 and C02, but it is thought that they are 0.06 p
inactive towards NO and propene. 3
!G
The mechanism is explained in Fig. 20; initially, the 0.04
hydrocarbon reacts with the NO2 resulting from NO
oxidation to generate the nitro and nitrite species, a 0.02
part of which are transformed further into carbonyl-
containing compounds. The nitrogen is formed by two 0

pathways: the reaction between NO2 and the carbonyl 550 600 650 700

species and the reaction between oxygen and the Temperature/ K

species containing one nitrogen atom (R-ON0 (ads)

Fig. 21. Correlation between surface isocyanate concentration and
and R-NO* (ads)). catalytic activity with supported rhodium: (0) NO conversion, (0)
(c) A mechanism for the SCR on supported rhodium N2 selectivity, (V) NC0 intensity, (0) NOs+, and (- -) CN-
has been suggested by Bamwenda et al. [136]. Here, intensity [ 1361.
20 A. Fritz, V Pitchon/Applied Catalysis B: Environmental 13 (1997) l-25

Fig. 22. Proposed ‘dual-site’ reaction mechanism for NO, reduc-

tion in lean diesel exhaust gas on precious metal-containing 0-
100 200 300 400 500 600
catalyst [143].
Temperature I “C

Fig. 23. Activity of monoliths containing (0) 0.2% Pt, (W) 0.2%
nitrogen or N20. I+O.O5% Rh, (0) 0.2% Pta.04% Au, and (A) 0.7% Cu-ZSM-5
[135].
H,(ads) + 0 --+ C,H,O,
C,H,O, + NO -+ R-NC0
R-NC0 + NO, NOz --+ N2 + N20 that their activity for SCR has been brought to light so
late. One possible explanation is that in the literature
(d) Engler et al. have studied the SCR with hydro- devoted to three-way catalysis, the reducing agent
carbon over a catalyst of platinum base upon an receiving most attention is carbon monoxide which
alumina support, deposited on a monolith [143]. By is known to be inactive for SCR. Another potential
varying reaction parameters such as the nature of the reason is that these catalysts are usually studied with a
hydrocarbon, the HUN0 ratio coupled with diffuse composition mixture close to stoichiometry with a low
reflectance FTIR results, they were enabled to propose percentage of oxygen, thus disguising the promoting
a ‘dual-site’ mechanism schematised in Fig. 22, in effect of a high concentration of oxygen.
which the hydrocarbon and NO are adsorbed on
different sites reacting with oxygen to give adsorbed 3.2.4. NO, storage and reduction catalyst
NO* and C,HY02, before reacting together to produce The challenge to reduce NO, emissions by a direct
N2, CO2 and H20. catalytic process has proved difficult, the automobile
manufacturers have therefore developed a new tech-
3.2.3.4.3. Bench tests evaluation. In Fig. 23 are nology for passenger car engines. This is the ‘lean-
represented the activities of several catalysts which bum engine’ which operates by employing a lean air/
were deposited upon monoliths and placed into a fuel ratio at cruising speeds and stoichiometric ratios
diesel exhaust line enriched with 700 ppm of when more power is required. To meet these require-
propene [135]. Cu-ZSM-5 is the least efficient ments, the catalyst must play a double role. First, it
catalyst with less than 25% conversion of NO at must store NO, while the engine runs lean, then upon
673 K. The monolith containing 0.2% platinum switching to a short stoichiometric excursion, it must
converts 380% of NO at 483 K. The catalysts reduce NO, to nitrogen. The first catalyst of this type
loaded with 0.2% platinum and 0.05% rhodium is was developed by Toyota [145,146], its composition
the most active with a value of 50% at 473 K. being based upon that of a three-way catalyst. The
Therefore, the metals from group VIII and, parti- preparation was by impregnation upon alumina of
cularly, platinum and rhodium are efficient for HC- noble metals - essentially Pt - of several alkali and
SCR in the presence of an excess of oxygen 11441. alkali earth metals (Nat, K+, Ba2+) and of rare earth
These same metals have been used in three-way oxides (mainly La203). The NO conversion has been
converters for over twenty years. It is rather surprising measured under transient conditions by feeding the
 A. Fritz, V Pitchon/Applied Catalysis B: Environmental 13 (1997) 1-25 21

Ob””
0 120 240 360 480 “0 2 4 6 8
Time /set 02 Concentration l%
Fig. 24. Activity of the NSR catalyst: (- - -) inlet concentration; Fig. 25. Influence of oxygen concentration on the NO, storage
(-) outlet concentration; period of one cycle = 240 s; tem- capacity [ 1461.
perature = 300°C. Rich (a) 70 ppm NO, 0.6% 02, 2000 ppm
C3H6, 0.5% CO, 10% H20, and 14.5% C02. Lean(b) 700 ppm NO,
NO 02
4% Oz. 800 ppm C3Hs. 0.1% CO, 10% H20, and 12.7% CO2 [146].
(a)
catalyst with a gas stream whose composition alter-
nated between lean and rich (Fig. 24).
During the lean excursion, the NO concentration
post-catalyst increases but does not reach the entrance
level. During the rich transition, the gas stream post-
catalyst does not contain NO any more in parallel with
an important formation of nitrogen. N2
It appears that the catalyst adsorbs NO under lean
mixture and reduces NO in a reducing gas @I f t
composition. Fig. 25 clearly illustrates that the NO,
storage only occurs in an oxidising media.
The basicity of the storage component determines
the amount of NO adsorbed. Nevertheless, the per-
formances of platinum, especially the hydrocarbon
oxidation, decrease with the basicity. The particle size
of platinum and BaO and the distance between them R: Reducing agents
are other factors which govern the NO storage. Small
particles in contact with each other adsorb more than Fig. 26. Mechanism of reduction on NO, storage catalyst [ 1461.
large particles. Takahashi et al. proposed that in lean
mixture, NO is oxidised into NOz and is stored under petrol containing 30 ppm of SO*, the reduction was
the form of a nitrate (Fig. 26). In rich mixture, the still of ca. 60%.
hydrocarbon adsorbed onto the platinum sites then Daimler Benz has also developed a lean-burn
reacts with these nitrates to form nitrogen. engine. The catalyst contains refractory oxides of
Toyota tested this catalyst on a vehicle equipped Al, Ce and Zr and noble metals Pt and Rh. In their
with a lean-bum engine of 1.8 1 cubic capacity. The publication [ 1471 the nature of the adsorbing phase is
fresh catalyst gave a 90% reduction of NO, emissions not given. The dual mechanism adsorption of NO,
according to the Japanese evaluation procedure. After during lean excursion and reduction of NO, during the
having been aged during 100.000 km with Japanese rich phase is similar to the one proposed above. With
22 A. Fritz, V Pitchon/Applied
Fig. 27. SNR catalyst behaviour in a 30/30 cycle (30.second
exposure to a stoichiometric phase and 30-second exposure to lean-
bum phase) and 30 (SV = 15 000 h-’ and T= 350°C) [147].
this formulation of absorber-catalyst technology hav-
ing a cycle of 30 s lean and 30 s rich, an average of
94% NO, conversion is obtained over the two modes
with a prototype lean-burn engine (Fig. 27).
4. Conclusion
The selective reduction of NO, during lean exhaust
mixture conditions has proved a challenging problem
which, particularly during the last five years, motor car
manufacturers have shared with many research labora-
tories. The resulting large number of publications
serve both academic and applied interests. Equally,
a terrific number of catalysts, in continuing association
with and developed from differing technologies, do
not appear to meet the requirements for real diesel
application. The aim of this paper has been to review a
variety of systems designed to help the researchers to
assess the performances of their own catalysts while
seeking to provide an alternative solution to meeting
the emissions standards.
In summary, there are three principal types of
catalysts active for the HC-SCR whose advantages
and problems, in respect of a technical application, are
discussed below.
4.1. Zeolites
This group can be subdivided into two classes, the
first one containing the redoubtable Cu-ZSM-5, while
the second comprises the ZSM-5 framework with the
Catalysis B: Envimnmental I3 (1997) I-25
copper substituted by a large variety of cations or other
zeolitic structures. The former, under certain experi-
mental conditions, exhibits a good activity and high
selectivity towards nitrogen. However, these catalysts
are subject to hydrothermal deactivation and are only
active at temperatures too high to be considered for
diesel applications. The substitution of copper by
other metals or the use of a zeolite structure other
than ZSM-5 has lead to the development of formula-
tions which exhibit acceptable conversions even at
high space velocity.
The present aim, with respect to these catalysts, is a
verification of the true mechanism from amongst the
several proposed in the literature. Also numerous
further questions remain such as that of the oxidation
state of the active site, which might be Cu’, Cu+ or
cl.?+, the identification of the spectator species to the
reaction intermediates, or to assess the role of the
acidity of the structure. Moreover, studies of the
stabilisation of the structure upon hydrothermal deac-
tivation and the development of zeolite-type catalysts
active at low temperatures are both current research
areas.
4.2. Oxide-type catalysts
A wide variety of oxides (bulk or supported), exhibit
a moderate activity. The temperature of operation is
high (> 673 K), but a large selection of reducing
agents can be used. However, the activity is low at
high space velocity and they are strongly subjected to
deactivation by sulphur dioxide and to thermal ageing
due to loss of specific area.
Future research on these materials should allow a
greater understanding of the mechanisms of deactiva-
tion for an improvement in their resistance to SOa to
be secured. Another aspect of the work could be the
comprehension of the correlation between activity and
surface area, linked to the mobility of oxygen, struc-
ture defects or anionic vacancies, by a development in
techniques of preparation in solid state chemistry in
order to have solids with a wide range of surface area
and electronic properties.
4.3. Supported noble metals
The case of the noble metal catalyst has proved very
challenging, not least because the realisation that this
 A. Fritz, V Pitchon/Applied
group was very active for the reaction in strongly
oxidising conditions and with high space velocities
was not immediate. A high activity is generally
observed at low temperatures, but with too narrow a
range. For most examples, the efficiency is only very
little affected by the presence of water, while it is also
known that noble metals show some resistance to
sulphur dioxide poisoning.
Future research into supported noble metals might
perhaps be focussed in such a way that the operating
temperature range is broadened. This goal might be
attained by intensive studies upon the effects of pro-
moter addition, combinations with another transition
metals or the effect of the support on the nature of the
active site. However, it remains important to under-
stand the mechanisms in order to give a judicious
choice for the second metal, or a combination of
several elements as it has already appeared in the
literature. A problem associated with supported noble
metals is that of nitrous oxide formation in the exhaust
emissions. Obviously, this should be avoided and the
development of future catalysts must take this into
consideration.
Finally, lean-bum engine development has brought
to light a new challenge for catalysts, the discovery of
new material possessing both a very good and selec-
tive NO, storage capacity.
Acknowledgements
The authors would like to thank Dr. S. Brooks for
proofreading the manuscript.
References
[II H. Glick, J.J. Klien and W. Squire, J. Chem. Phys., 27
(1957) 850.
PI Y.B. Zeldovich, Acta Physico-chern. URSS, 21 (1946) 577-
656.
[31 R. Prasad, Catal. Rev.-&i. Eng., 26(l) (1984) 1.
[41 S. Snyder and D. Bredt, Pour la Science, 177 (1992) 70.
[51 H. Bosch and F. Janssen, Catal. Today, 2 (1988) 369.
El M. Chiron, Stud. Surf. Sci. Catal., 30 (1987) 1.
[71 R. Impens, Stud. Surf. Sci. Catal., 30 (1987) 11.
[81 M.E. Kowalok, Environment, 35(6) (1993) 12.
[91 J.H. Seinfeld, Science, 243 (1989) 745.
[lOI J.N. Armor, Appl. Catal. B, 1 (1992) 221.
[Ill Y. Kondo, T. Kitada, M. Koike, S. Kawakami and Y.
Makino, J. Geophysical Res., 98 (1993) 20527.
Catalysis B: Environmental 13 (1997) 1-25 23
WI I. De Kermikri, La recherche, 279 (1995) 887.
Cl31 T.E. Green and C.N. Hinshelwood, J. Chem. Sot., 128
(1926) 1709.
[I41 P.W. Bachman and G.B. Taylor, J. Phys. Chem., 33 (1929)
447.
u51 J. Zawadski and G. Perlinsky, Compt. Rend., 198 (1934)
260.
1161 R.R. Sakaida, R.G. Rinker, Y.L. Wang and W.H. Corcoran,
AIChE J., 7 (1961) 658.
[I71 M. Shelef, K. Otto and H. Gethi, Atmos. Envir., 3 (1969)
107.
[181 A. Amirnazmi, J.E. Benson and M. Boudart, J. Catal., 30
(1973) 55.
[191 A. Amimazmi and M. Boudart, J. Catal., 39 (1975) 383.
PO1 S. Pancharatnum, K.J. Lim and D.M. Manson, Chem. Eng.
Sci., 30 (1975) 781.
WI R.J. Wu, T.Y. Chou and C.T. Yeh, Appl. Catal. B, 6 (1995)
105.
WI M. Iwamoto and H. Hamada, Catal. Today, 10 (1991) 57.
~231 E.R.S. Winter, J. Catal., 22 (1971) 158.
~241 B. Rebenstorf, L. Larsson and R. Larsson, Acta Chim.
Scetin. A, 31 (1977) 877.
~251 P. Meubus, J. Electrochem. Sot., 124 (1977) 49.
WI S. Shin, K. Ogawa and K. Shimomura, Mater. Res. Bull., 14
(1979) 133.
1271 K. Tabata, J. Mater. Sci. Lett., 7 (1988) 147.
m Y. Teraoka, H. Fukuda and S. Kagawa, Chem. Lett., (1990)
1.
P91 H. Hamada, Y. Kintaichi, M. Sasaki and T. Ito, Chem. Lett.,
(1990) 1069.
[301 X. Zhang, A.B. Walters and M.A. Vannice, J. Catal., 155
(1995) 290.
[311 I. Halasz, A. Brenner, M. Shelef and K.Y.S. Ng, Catal. Lett.,
11 (1991) 327.
~321 G.K. Boreskov, Disc. Faraday Sot., 41 (1966) 263.
[331 M. Iwamoto, K. Maruyama, N. Yamazoe and T. Seiyama, J.
Chem. Sot., Chem. Commun., (1972) 615.
[341 M. Iwamoto, S. Yokoom, K. Sakai and S. Kagawa, J. Chem.
Sot., Faraday Trans. 1, 77 (1981) 1629.
[351 J. Valyon and W.K. Hall, J. Catal., 143 (1993) 52.
[361 M. Iwamoto, M. Nakamura, H. Nagano, S. Kagawa and T.
Seiyama, J. Phys. Chem., 86(2) (1982) 153.
[371 M. Iwamoto, H. Yahiro, Y. Mine and S. Kagawa, Chem.
Lett., (1989) 213.
[381 K. Eranen, N. Kumar and L.E. Lindfors, Appl. Catal. B, 4
(1994) 213.
1391 Y. Li and W.K. Hall, J. Catal., 129 (1991) 202.
[401 D.J. Liu and H.J. Robota, Catal. Lett., 21 (1993) 291.
[411 J. Valyon and W.K. Hall, J. Phys. Chem., 97 (1993)
1204.
~421 E. Giamello, D. Murphy and G. Magnacca, J. Catal., 136
(1992) 510.
[431 J. Valyon and W.K. Hall, Stud. Surf. Sci. Catal., 75 (1992)
1339.
[441 M. Iwamoto, H. Yahiro, N. Mizuno, W.-X. Zhang, Y. Mine,
H. Furukawa and S. Kagawa, J. Phys. Chem., 96 (1992)
9360.
24 A. Fritz, V Pitchon/Applied Catalysis B: Environmental
[45] G.D. Lei, B.J. Adelman, J. Sarkany and W.M. Sachtler,
Appl. Catal. B, 5 (1995) 245.
[46] M. Shelef, Catal. Lett., 15 (1992) 305.
[47] T. Tabata, M. Kokitsu and 0. Okada, Catal. Today, 22
(1994) 147.
[48] C.J. Pereira and K.W. Phumlee, Catal. Today, 13 (1992) 23.
[49] P. Ciambelli, G. Bagnasco, L. Lisi and M. Turco, Appl.
Catal. B, 1 (1992) 61.
[50] M. Inomata, A. Miyamoto and Y. Murakami, J. Catal., 62
(1980) 140.
[51] F. Jansen, F. van den Kerkhof, H. Bosch and T.J.R. Ross, J.
Phys. Chem., 91 (1987) 5921.
[52] G. Ramis, G. Busca, F. Bregani and P. Forzatti, Appl. Catal.,
64 (1990) 259.
[53] V. Tufano and M. Turco, Appl. Catal. B, 2 (1993) 9.
[54] W.S. Kyte, J. Bettelheim and J.R.P. Cooper, Inst. Chem.
Eng. Symp. Ser. No. 78, 1982.
[55] J. Eto, Eur. Patent application 600/7-85-040, 1985.
[56] K.C. Taylor, Catal. Rev.-Sci. Eng., 35(4) (1993) 457.
[57] J.C. Summers and D.R. Monroe, Ind. Eng. Chem. Prod. Res.
Dev., 20 (1981) 23.
[58] C. Howitt, V. Pitchon and G. Maire, J. Catal., 155 (1995) 47.
[59] W. Held, A. Konig, T. Richter and L. Puppe, SAE paper No.
900496, 1990.
[60] M. Iwamoto, Proc. meet. Catal. Technol. Removal of NO,
Tokyo, Jan. 1990, p. 17.
[61] Z. Chajar, M. Primet, H. Praliaud, M. Chevrier, C. Gauthier
and E Mathis, Appl. Catal. B, 4 (1994) 199.
[62] M. Iwamoto, N. Mizuno and H. Yahiro, Stud. Surf. Sci.
Catal., 75 (1992) 1285.
[63] S. Sato, H. Yuu, H. Yahiro, N. Mizuno and M. Iwamoto,
Appl. Catal., 70 (1991) Ll.
[64] B.K. Cho, J. Catal., 142 (1993) 418.
[65] F. Radtke, R.A. Koeppel and A. Baiker, Appl. Catal. A, 107
(1994) L125.
[66] C.N. Montreuil and M. Shelef, Appl. Catal. B, 1 (1992) Ll.
[67] D. Liu and H. Robota, Appl. Catal. B, 4 (1994) 155.
[68] T. Pieplu, F. Poignant, A. Vallet, J. Saussey and J.C.
Lavalley, Stud. Surf. Sci. Catal., 96 (1995) 619.
[69] A.V. Kucherov, J.L. Gerlock, H.W. Jen and M. Shelef, J.
Catal., 152 (1995) 63.
[70] A.V. Kucherov, A.A. Slinkin, S.S. Goryashenko and K.I.
Slovetskaya, J. Catal., 118 (1989) 459.
[71] M.Y. Kutyrev, A.A. Ukharsky, A.N. Ilichev and V.A.
Matyshak, Proc. 1st Int. Cong. on Environ. Catal., Pisa,
1995, p. 311.
[72] A.P. Walker, Catal. Today, 26 (1995) 107.
[73] R. Burch and S. Scire, Appl. Catal. B, 3 (1994) 295.
[74] Y. Li and W.K. Hall, J. Phys. Chem., 94 (1990) 6145.
[75] R. Burch and P.J. Millington, Appl. Catal. B, 2 (1993) 101.
[76] B.K. Cho, J. Catal., 155 (1995) 184.
[77] G.P. Ansell, A.F. Diwell, R.R. Rajaram and AX Walker,
Appl. Catal. B, 2 (1993) 81.
[78] C.J. Bennet, P.S. Bennett, S.E. Golunski, J.W. Hayes and
A.P. Walker, Appl. Catal. A, 86 (1992) Ll.
[79] E. Kikuchi, K. Yogo, S. Tanaka and M. Abe, Chem. Lett.,
(1991) 1063.
13 (1997) 1-25
[80] M. Shelef, C.N. Montreuil and H.W. Jen, Catal. Lett., 26
(1994) 277.
[81] J.O. Petunchi and W.K. Hall, Appl. Catal. B, 2 (1993) L17.
[82] H. Hamada, Y. Kintaichi, M. Sasaki and T. Ito, Appl. Catal.,
70 (1991) L15.
[83] Z. Chajar, M. Primet, H. Praliaud, M. Chevrier, C. Gauthier
and F. Mathis, Catal. Lett., 28 (1994) 33.
[84] J.L. D’itry and W.M.H. Sachtler, Catal. Lett., 15 (1992) 289.
[85] Y. Torikai, H. Yahiro, N. Mizuno and M. Iwamoto, Catal.
Lett., 9 (1991) 91.
[86] D.R. Monroe, C.L. Dimaggio, D.D. Beck and F.A.
Matekunas, SAE paper No. 930737, 1993.
[87] M. Konno, T. Chikahisa, T. Murayama and M. Iwamoto,
SAE paper No. 920091, 1992.
[88] T. Tanabe, T. Iijima, A. Koiwai, J. Mizuno, K. Yokota and
A. Isogai, Appl. Catal. B, 6 (1995) 145.
[89] S. Matsumoto, K. Yokota, H. Doi, M. Kimura, K. Sekizawa
and S. Kasahara, Catal. Today, 22 (1994) 127.
1901 K. Kharas, H.J. Robota and D.J. Liu, Appl. Catal. B, 2
(1993) 225.
[91] A.V. Kucherov, C.P. Hubbard and M. Shelef, J. Catal., 157
(1995) 603.
[92] R. Gopalaktishnan, P.R. Stafford, J.E. Davidson, W.C.
Hecker and C.H. Bartholomew, Appl. Catal. B, 2 (1993)
165.
[93] M.J. He&rich and M.L. Deviney, SAE paper No. 930736,
1993.
[94] B. Coq, D. Tachon, F. Figueras, G. Mabilon and M. Prigent,
Appl. Catal. B, 6 (1995) 271.
[95] G. Centi, S. Perathoner and L. Dall’olio, Appl. Catal. B, 4
(1994) L275.
[96] H. Hirabayashi, H. Yahiro, N. Mizuno and M. Iwamoto,
Chem. Lett., (1992) 2235.
[97] M. Iwamoto, H. Yahiro, H.K. Shin, M. Watanabe, .I. Guo,
M. Konno, T. Chikahisa and T. Murayama, Appl. Catal. B,
5 (1994) Ll.
[98] H. Hamada, Y. Kintaichi, M. Sasaki, T. Ito and M. Tabata,
Appl. Catal., 64 (1990) Ll.
[99] H. Hamada, Y. Kintaichi, M. Sasaki, T. Ito and T. Yoshinari,
Appl. Catal. A, 88 (1992) Ll.
[loo] H. Hamada, Y. Kintaichi, M. Sasaki, T. Ito and M. Tabata,
Appl. Catal., 70 (1991) L15.
[loll K. Yogo, S. Tanaka, M. Ibara, T. Hishiki and E. Kikuchi,
Chem. Lett., (1992) 1025.
[102] K. Yogo, M. Ihara, I. Terasaki and E. Kikuchi, Appl. Catal.
B, 2 (1993) Ll.
[103] Y. Li and J.N. Armor, J. Catal., 145 (1994) 1.
[104] T. Tabata, M. Kokitsu and 0. Okada, Appl. Catal. B, 6
(1995) 225.
[105] J.S. Feeley, M. Deeba, R.J. Farrauto, G. Beri and A. Haynes,
Appl. Catal. B, 6 (1995) 79.
[106] Y. Li, P.B. Battavio and J.N. Armor, J. Catal., 142 (1993)
561.
[107] Y. Li and J.N. Armor, Appl. Catal. B, 2 (1993) 239.
[108] Aribotta, M. Lezcano, J. Vassallo, E. Miro, E.A. Lombardo,
J.O. Petunchi, J.M. Dereppe and C. Moreaux, Proc. 1st Int.
Cong. on Environ. Catal., Pisa, 1995, p. 255.
 A. Fritz, V Pitchon/Applied Catalysis B: Environmental
[109] J. Vassallo, M. Lezcano, E. Miro and J. Petunchi, Stud. Surf.
Sci. Catal., 96 (1995) 697.
[llO] Y. Li and J.N. Armor, J. Catal., 150 (1994) 376.
[ill] Y. Li, T.L. Slager and J.N. Armor, J. Catal., 150 (1994) 388.
[112] Y. Li and J.N. Armor, Appl. Catal. B, 3 (1993) 55.
[113] C. Yokoyama and M. Misono, Chem. Lett., (1992) 1669.
[114] C. Yokoyama and M. Misono, Catal. Lett., 29 (1994) 1.
[115] C. Yokoyama and M. Misono, J. Catal., 150 (1994) 9.
[ 1161 Y. Kintaichi, H. Hamada, M. Tabata, M. Sasaki and T. Ito,
Catal. Lett., 6 (1990) 239.
11171 Y. Kintaichi, H. Hamada, M. Tabata, M. Sasaki and T. Ito,
Catal. Lett., 6 (1990) 239.
11181 H. Hamada, Y. Kintaichi, M. Tabata, M. Sasaki and T. Ito,
Chem. Lett., (1991) 2179.
11191 M. Tabata, H. Tsuchida, K. Miyamoto, T. Yoshinari, H.
Yamazaki, H. Hamada, Y. Kintaichi, M. Sasaki and T. Ito,
Appl. Catal. B, 6 (1995) 169.
11201 Y. Torikai, H. Yahiro, N. Mizuno and M. Iwamoto, Catal.
Lett., 9 (1991) 91.
11211 X. Zhang, A.B. Walters and M.A. Vannice, J. Catal., 155
(1994) 290.
[ 1221 X. Zhang, A.B. Walters and M.A. Vannice, J. Catal., 146
(1994) 568.
11231 H. Hamada, Y. Kintaichi, M. Sasaki, T. Ito and M. Tabata,
Appl. Catal., 75 (1991) Ll.
[ 1241 M. Inaba, Y. Kintaichi and H. Hamada, Proc. 1st bit. Cong.
on Environ. Catal., Pisa, 1995, p. 327.
[ 1251 K.A. Bethke, D. Alt and MC. Kung, Catal. L&t., 25 (1994)
37.
[ 1261 C. Li, K.A. Bethke, H.H. Kung and M.C. Kung, J. Chem.
Sot., Chem. Commun., (1995) 813.
11271 T. Miyadera, Appl. Catal. B, 2 (1993) 199.
I1281 Y. Ukisu, S. Sato, G. Muramatsu and K. Yoshida, Catal.
Lett., 16 (1992) 11.
[ 1291 Y. Ukisu, S. Sato, A. Abe and K. Yoshida, Appl. Catal. B, 2
(1993) 147.
[ 1301 A. Ueda, T. Oshima and M. Haruta, Proc. 1st Int. Cong. on
Environ. Catal., Pisa, 1995, p. 343.
13 (1997) 1-25 25
[131] H. Furukawa, T. Harada, Y. Teraoka and S. Kagawa, 68th
meeting of Catal. Sot. Jpn. (A), 1991, p. 4H215.
[132] A. Obuchi, M. Nakamura, A. Ogata, K. Mizuno, A. Ohi and
H. Ohuchi, J. Chem. Sot., Chem. Commun., 16 (1992)
1150.
[133] H. Hamada, Catal. Today, 22 (1994) 21.
[134] K. Masuda, T. Tsujimura, K. Shinoda and T. Kato, Appl.
Catal. B, 8 (1996) 33.
[135] A. Obuchi, A. Ohi, M. Nakamura, A. Ogata, K. Mizuno and
H. Ohuchi, Appl. Catal. B, 2 (1993) 71.
[ 1361 G.R. Bamwenda, A. Ogata, A. Obuchi, J. Oi, K. Mizuno and
I. Skrzypek, Appl. Catal. B, 6 (1995) 311.
[137] R. Burch, P.J. Millington and A.P. Walker, Appl. Catal. B, 4
(1994) 65.
[ 1381 A. Takami, T. Takemoto, H. Iwakuni, F. Saito and K.
Komatsu, SAE paper No. 950746, 1995.
[139] R. Burch and P.J. Millington, Catal. Today, 26 (1995)
185.
[140] R. Burch and T.C. Walting, Catal. Lett., 37 (1996) 51.
[141] T. Tanaka, T. Okuhara and M. Misono, Appl. Catal. B, 4
(1994) Ll.
[142] M. Sasaki, H. Hamada, Y. Kintaichi and T. Ito, Catal. Lett.,
15 (1992) 297.
[143] B.H. Engler, J. Leyrer, ES. Lox and K. Ostgathe, Stud. Surf.
Sci. Catal., 96 (1995) 529.
[144] A. Ogata, A. Obuchi, G.R. Bamwenda, J. Oi and K. Mizuno,
Proc. 1st Int. Cong. on Environ. Catal., Pisa, 1995, p. 347.
[145] K. Kato, H. Nohira, K. Nakanishi, S. Iguchi, T. Kihara and
H. Mu&i, Eur. Patent application 0573 672 Al (1993) to
Toyota.
[146] N. Takahasi, H. Shinjoh, T. Ijima, T. Suzuki, K. Yamakazi,
K. Yokota, H. Suzuki, N. Miyoshi, S. Matsumoto, T.
Tanizawa, T. Tanaka, S. Tateishi and K. Kasahara, Proc. 1st
Int. Cong. on Environ. Catal., Pisa, 1995, p. 45.
[147] W. Bogner, M. Kramer, B. Krutzsch, S. Piscinger,
D. Voigtllnder, G. Wenniger, F. Wirbeleit, MS. Brogan,
R.J. Brisley and D.E. Webster, Appl. Catal. B, 7 (1995)
153.