Catalytic Reduction of Nitric Oxide by Methane in the Presence of Oxygen on Palladium-Exchanged Mordenite Zeolites

Article abstract of DOI:10.1006/jcat.1998.2112

Pd was introduced by exchange in non-dealuminated and dealuminated mordenites. After activation under oxygen at 500°C, NO adsorption revealed the presence of isolated Pd²⁺ ions and Pd/PdO particles. Two types of Pd ions were shown: the first one would be located in the small cavities and leads to the formation of dinitrosyl complexes, whereas the second one, probably sitting in the main channels, is responsible for the obtainment of mononitrosyl complexes. Pd/PdO particles are active in the direct combustion of CH₄ by oxygen and poorly active in the selective reduction (SCR) of NO. Pd ions are active for the NO SCR; the ions in the main channels are thought to be more active than the ions in the small cavities for accessibility reasons. In the case of the non-dealuminated mordenite catalyst, an activation under reactants is observed. Moreover, the activity is increased by the addition of 10 vol% water which is the opposite of the classical behavior of metal-loaded zeolites. Both phenomena are attributed to a migration of Pd ions from hidden to accessible sites. For the dealuminated sample, the observed deactivation seems to be due to a migration of isolated Pd ions into PdO particles located outside the zeolite matrix.

Full text of DOI:10.1006/jcat.1998.2112

JOURNAL OF CATALYSIS 177, 352–362 (1998)
ARTICLE NO. CA982112
Catalytic Reduction of Nitric Oxide by Methane in the Presence
of Oxygen on Palladium-Exchanged Mordenite Zeolites
,1
ꢀ
ꢀ
ꢀ
Claude Descorme, Patrick Ge´lin, Christine Le´cuyer,† and Michel Primet
ꢀ
Laboratoire d’Application de la Chimie a` l’Environnement, UMR CNRS 5634, Universite´ Claude Bernard Lyon I, 43, Boulevard du 11
Novembre 1918, 69622 Villeurbanne Cedex, France; and †GAZ DE FRANCE, Direction de la Recherche, CERSTA, 361,
avenue du Pre´sident Wilson, P.O. Box 33, 93211 La Plaine St-Denis Cedex, France
Received January 26, 1998; revised April 15, 1998; accepted April 18, 1998
to remove the remaining CO, NOx, and unburned methane
from the exhaust gases. The catalytic reduction of NOx by
hydrocarbons has been studied extensively in the past few
years, appearing as a promising process for the removal of
dilute amounts of NOx from exhaust gases in the presence
of quite large amounts of oxygen (1–3). However, due to its
particularly high stability, methane is very difficult to acti-
vate and most catalysts, when active, tend to burn methane
by O2 without removing NOx. Therefore, the emission con-
trol of CNG-fueled vehicles requires the development of
specific catalysts. Since the pioneering work by Li and
Armor on Co2+-exchanged ZSM-5catalysts(4), newclasses
of transition metal-exchanged zeolite catalysts have been
shown to selectively reduce NOx by CH4 in excess oxygen
(5–17). Nontransition metal (Ga, In) loaded zeolites were
also found to be very active for the reaction between CH4
and NOx (18–23). However, the activity of the latter cata-
lysts is strongly decreased by the presence of water. On
the contrary, Pd exchanged H-ZSM-5 and MOR catalysts
(24–30), as well as, to a lesser extent, Co-H-ZSM-5 catalysts
(7), appear much less sensitive to water, which makes them
more attractive for being used in catalytic converters. For
automotive applications, another problem is raised by the
large range of temperatures (300–750ꢁC) which the catalyst
is submitted to, depending on the traffic conditions. The re-
sistance of the catalyst toward severe steaming conditions
is essential in view of its commercial application. It is well
known that zeolites are sensitive to steaming, which induces
the dealumination of the framework and, depending on the
structure and the Si/Al ratio, the possible collapse of the
structure. The catalytic properties of Pd-H-ZSM-5 catalysts
of varying Si/Al ratios have been previously studied (31).
In this paper, we were interested in the catalytic properties
of Pd-exchanged mordenite catalysts in the selective reduc-
tion of nitric oxide by methane in the presence of oxygen.
The influence of the Si/Al ratio on the catalytic activity and
selectivity in this reaction was examined. The adsorption of
NO on the Pd-H-MOR catalysts with varying Si/Al ratios
wascarefullyinvestigated byFT-IR spectroscopyin order to
Pd was introduced by exchange in non-dealuminated and dealu-
minated mordenites. After activation under oxygen at 500ꢁC, NO
adsorption revealed the presence of isolated Pd2+ ions and Pd/PdO
particles. Two types of Pd ions were shown: the first one would be
located in the small cavities and leads to the formation of dinitrosyl
complexes, whereas the second one, probably sitting in the main
channels, is responsible for the obtainment of mononitrosyl com-
plexes. Pd/PdO particles are active in the direct combustion of CH4
by oxygen and poorly active in the selective reduction (SCR) of NO.
Pd ions are active for the NO SCR; the ions in the main channels
are thought to be more active than the ions in the small cavities for
accessibility reasons. In the case of the non-dealuminated morden-
ite catalyst, an activation under reactants is observed. Moreover,
the activity is increased by the addition of 10 vol% water which is
the opposite of the classical behavior of metal-loaded zeolites. Both
phenomena are attributed to a migration of Pd ions from hidden to
accessible sites. For the dealuminated sample, the observed deacti-
vation seems to be due to a migration of isolated Pd ions into PdO
c
particles located outside the zeolite matrix.
ꢂ 1998 Academic Press
Key Words: SCR; palladium; mordenite; dealumination; nitric
oxide; methane.
INTRODUCTION
In view of environmental/pollution prevention, com-
pressed natural gas (CNG) appears to be an attractive al-
ternative fuel for vehicles. Besides for economical reasons,
CNG shows decided environmental advantages over gaso-
line or diesel: (i) CO2 emissions are reduced due to the high
H/C ratio ofCH4, which isthe main component (85–95% ) of
natural gas; (ii) despite the absence of CNG-dedicated en-
gines, the pollutants like CO, NOx, and unburned hydrocar-
bons (UHC) emitted by CNG-fueled vehicles are strongly
reduced, compared to noncatalyzed gasoline vehicles. How-
ever, in order to meet the more and more stringent regula-
tions for the control of vehicle emissions, the CNG-fueled
vehicles need to be equipped with catalytic converters able
1 Corresponding author. E-mail: Patrick.Gelin@univ-lyon1.fr.
352
0021-9517/98 $25.00
c
Copyright ꢂ 1998 by Academic Press
All rights of reproduction in any form reserved.

CATALYTIC REDUCTION OF NITRIC OXIDE
353
characterize the Pd sites. A stoichiometric interconversion formed upon dealumination of mordenite by steaming–
between mono- and di-nitrosyl Pd complexes was shown. acid leaching cycles may be removed after a final washing
The localization of these complexes in the mordenite pores by a (NH4)2SiF6 solution (32). A suspension of 3 g of the
and their reactivity in the reduction of NO is discussed in H-MOR-D sample in 30 cm3 of a buffered ammonium
the light of catalytic activity results.
acetate solution (1.28 M) was stirred at 80ꢁC while adding
10.5 cm3 of a (NH4)2SiF6 solution (0.367 M) at an injection
rate of 0.745 cm3/min. Stirring of the suspension at 80ꢁC
was maintained for one hour after the complete injection
of the (NH4)2SiF6 solution. After being thoroughly washed
by distilled water until complete elimination of Fꢃions, the
resulting sample was dried in air at 120ꢁC overnight and
designated as H-MOR-D-W.
Pd was introduced by conventional exchange of the
starting materials (NH4-MOR and H-MOR-D-W) using
aqueous solutions of tetrammine Pd(II) nitrate at 80ꢁC
for 24 h. The amount of Pd(II) salt was adjusted so as to
obtain a theoretical value of 1 wt% Pd. After exchange,
the preparation was thoroughly washed with deionized
water, filtered, and dried at 120ꢁC overnight. In order
to decompose ammonium ions (if present) into protons
and remove the ammine ligands from Pd complexes, the
Pd-exchanged zeolite samples were subsequently activated
under flowing oxygen from 25 to 500ꢁC with a linear ramp
rate of 0.5ꢁC/min. The resulting samples were designated
as Pd-H-MOR and Pd-H-MOR-D-W. The results of atomic
absorption analysis are reported in Table 1. The expected
level of Pd exchange of the H-MOR and H-MOR-D-W
samples (corresponding to 1 wt% Pd) was reached.
EXPERIMENTAL
1. Sample Preparation
The starting mordenite sample obtained in the Na+ form
from Zeocat (ZM-060, Si/Al = 5.5) was first converted to
the NH+4 form by exchanging with NH4Cl. The resulting
material contained 0.1 wt% Na (dry basis) and its chemical
formula determined from elemental analysis was Na0.13
(NH4)7.76 Al7.89 Si40.11 O48 (Si/Al = 5.1). It is designated
as NH4-MOR. The NH4-MOR sample was subsequently
dealuminated according to the following procedure. 20 g
of NH4-MOR were deposited onto the frittered disk of
a quartz reactor (40 mm ID, bed thickness = 8 mm) and
heated from room temperature up 500ꢁC under a 10 l/h
oxygen stream at a heating rate of 0.5ꢁC/min. After one
hour calcination at 500ꢁC, the reactor was purged for one
hour with nitrogen flow (flow rate = 10 l/h) at the same
temperature before increasing the temperature up to 650ꢁC
(heating rate of 5ꢁC/min); 5 vol% water vapor was then
added to the nitrogen stream for one hour at 650ꢁC. An ad-
ditional one hour purge under nitrogen at 650ꢁC wascarried
out before decreasing the temperature down to room tem-
perature. The sample was subsequently acid-leached by a
3 M HCl aqueous solution at 80ꢁC for one hour under
stirring, centrifuged, and thoroughly washed by distilled
water. The complete sequence of thermal treatments
followed by acid leaching was repeated three times and the
resulting dealuminated mordenite sample was finally dried
in air at 120ꢁC overnight and designated as H-MOR-D.
The H-MOR-D sample contained significant amounts of
extraframework Al (EFAL) species (see Table 1). It has
been shown that part of the extraframework Al species
2. Catalytic Activity Measurements
The experimental details for activity measurements were
described elsewhere (26, 31). Briefly, the conversions of
NOx and CH4 were measured using a U-shaped quartz reac-
tor (16-mm ID) operating in a steady-state plug flow mode.
Typically, 200 mg of catalyst were activated in-situ in oxy-
gen flow at 500ꢁC (linear ramp rate of 0.5ꢁC/min) purged
for one hour at 500ꢁC byflowinghelium and cooled to 250ꢁC
TABLE 1
Main Physicochemical Characteristics of Fresh and Aged Pd Exchanged Mordenite Samples
Elemental analysis
27Al NMR
N2 ads.
Vꢀ (cm3/g)
XRD
% crystall.d
Samples
H-MOR
H-MOR-D
H-MOR-D-W
Pd-H-MOR
Si/Al
Al/uc
Pd/uc
Pd (wt% )
FAL/ucb
EFAL/uc
% crystall.
5.1a
7.5
9.9
6.3
11.9
7.9a
5.6
4.5
6.6
3.7
6.1c
2.7
2.4
5.9
2.4
1.8
2.9
2.1
0.7
1.3
100
81
88
n.m.
n.m.
0.22
0.18
0.19
0.20
n.m.
100
82
86
91
n.m.
0.27
0.25
0.98
0.92
Pd-H-MOR-D-W
Note. n.m. = not measured.
a NH+4 form.
b By measuring the intensity of the 4-coordinate Al signal; reference = H-MOR (6.1 FAL/uc).
c Equal to the elemental composition corrected for the 6-coordinate Al contribution.
d Reference = H-MOR (% C = 100).

354
DESCORME ET AL.
before contact with the reactants. The reaction mixture was 3. Physicochemical Characterizations
adjusted so as to examine the catalytic activity under stoi-
X-ray diffraction measurements were performed us-
chiometric or lean conditions. It consisted of 2000 vpm NO,
1000 vpm CH4, 1000 vpm O2 (stoichiometry) or 6240 vpm
O2 (lean conditions) and helium as balance, at a total flow
rate of 167 cm3/min; vpm stands for ppm (v/v). (The ap-
ing CuKꢁ radiation. Nitrogen adsorption–desorption
isotherms were measured at ꢃ196ꢁC with a homemade ap-
paratus. The microporous volume was determined using the
t-plot method. TEM micrographs were obtained by direct
observation of the steamed samples by using a JEOL 100
CX microscope. The 27Al MAS NMR spectra were ob-
tained using a BRUKER MSL-300 spectrometer operating
at 78.2 MHz.
The IR studies of NO or NO2 adsorption were per-
formed using self-supported samples wafers (18 mm diam-
eter, weight of 30 mg) introduced into a homemade IR-
cell allowing in situ studies at varying temperatures under
controlled atmosphere (34). The samples were pretreated
in situ in flowing oxygen at 500ꢁC for 30 min (after a linear
ramp rate of 10ꢁC/min from 25 to 500ꢁC). The cell was sub-
sequently connected to a UHV system allowing a base pres-
sure as low as 10ꢃ8 Torr (1 Torr = 101,325/760 Pa and the
sample evacuated at 500ꢁC for 3h before beingcooled down
to 25ꢁC and contacted with NO in the 0.02–0.4 Torr pres-
sure range. The IR spectra were recorded at a resolution
of 4 cmꢃ1 on a Fourier-transform IR spectrometer Nicolet
Magna 550. All the reported spectra have been corrected
for the baseline (spectrum of the sample activated in situ
and evacuated under vacuum at 25ꢁC) in order to highlight
the spectral changes resulting from the NOx adsorption.
parent gas hourly space velocity (GHSV) was 30000 hꢃ1
,
based on the apparent bulk density of the zeolites, ca 0.5 g/
cm3.) Under lean conditions, the S ratio, defined as
[oxidants]/[reductants] = ([O2] + 0.5[NO])/2[CH4] equal-
ed 3.62. Two cycles were successively operated, consisting
of increasing the temperature from 250 up to 500ꢁC (linear
rate of 1ꢁC/min), maintaining 500ꢁC for 10 h, and decreas-
ing the temperature to 250ꢁC. For all samples, the catalytic
activity of the activated sample was first examined under
stoichiometric conditions before switching to the lean con-
ditions.
The effluent gases were analyzed using two gas chro-
matographs equipped with TCD and FID detectors and
NOx infrared analyzers. Carbon and nitrogen balances
were checked. The NOx conversion was determined accord-
ing to the equation
NOx conversion % = ([NO]0 + [NO2]0 ꢃ [NO] ꢃ [NO2])
ꢀ 100/([NO]0 + [NO2]0),
where [NO]0 and [NO2]0 are the inlet concentrations of NO
and NO2, respectively, and [NO] and [NO2] are the outlet
concentrations. The NO2 formation was low in the whole
range of temperature ([NO2] ꢄ 40 vpm), almost indepen-
dent of the temperature and ascribed to the NO oxidation
in the dead volume of the apparatus. The CH4 conversion
was determined from the formation of CO2; carbon mono-
xide was never observed.
RESULTS
1. Aluminum Composition and Crystallinity of Pd
Exchanged Mordenite Catalysts
The Al composition of the starting and dealuminated
mordenite samples from elemental analysis is given in
Table 1 and compared to the framework Al (FAL) com-
position calculated from 27Al MAS NMR spectra.
Over Co-ZSM-5 catalysts, two reactions, NO reduction
by CH4 [1] and CH4 combustion [2], have been suggested
to occur (4–7):
The 27Al MAS NMR spectrum of the NH4-MOR sample
exhibits only one intense line at ca 53 ppm from Al(H2O)63+
due to 4-coordinate (framework) Al (35), which suggests
the absence of extraframework Al species. It is deduced
that the framework Al composition of the NH4-MOR sam-
ple is equal to the elemental Al composition; i.e. 7.9 Al
per unit cell (Al/u.c.). Upon calcination, the intensity of the
4-coordinate Al line decreases while a line at ca 0 ppm from
Al(H2O)36+ ascribed to 6-coordinate (extraframework) Al
appears (35), indicating the removal of some structural Al
ions upon calcination. Accordingly, the FAL composition
of the H-MOR sample determined from the integrated in-
tensity of the 4-coordinate Al line was found equal to 6.1
Al/u.c. Upon exchange with Pd, the FAL composition re-
CH4 + 2 NO + O2 → CO2 + N2 + 2 H2O
CH4 + 2 O2 → CO2 + 2 H2O.
[1]
[2]
Reaction [1] was based upon the fact that the catalyst was
not active for NO reduction in the absence of O2 (4–8) be-
low 500ꢁC. The same was observed on Pd catalysts (24, 33).
Accordingly, the selectivity towards NO reduction, SSCR
,
defined as the fraction of methane involved in reaction [1],
can be written
SSCR = 0.5[NO]0CNO /[CH4]0CCH
,
4
where CNO and CCH are the conversions of NO and CH4, mained unchanged and most of EFAL species present in
4
[NO]0 is the inlet concentration of NO, and [CH4]0 is the the H-MOR sample were removed as shown by the fairly
inlet concentration of CH4. In our experimental conditions, good agreement between NMR and chemical analysis re-
0.5[NO]0 = [CH4]0, and SSCR = CNO/CCH
.
sults reported in Table 1.
4

CATALYTIC REDUCTION OF NITRIC OXIDE
355
Repeated cycles of steaming and HCl leaching induced oxide. Conversely, the SCR selectivity smoothly decreases
a decrease of the elemental Al composition down to 5.6 upon temperature increase, indicating that higher temper-
Al/u.c. The FAL composition from 27Al MAS NMR de- atures favor the combustion of methane by O2. The same
creased more severely upon these treatments, indicating trend was reported on Pd-H-ZSM-5 catalysts containing
that a large fraction of Al (2.9 Al/u.c.) was still deposited more than ca 0.5 wt% Pd (36). It is noteworthy that con-
in the zeolite pores. This result was corroborated by the versions of methane and nitric oxide increase significantly
appearance of a broad line at ca 0 ppm. The final washing with time on stream at 500ꢁC and the conversions measured
with (NH4)2SiF6, while not affecting the FAL composition, upon decreasing temperature are higher than those mea-
removed 55% of the EFAL species produced upon the dea- sured during the heating ramp. This conversion increase
luminating treatment. Further exchange with Pd did not in- is accompanied by an increase of the selectivity for SCR.
fluence the framework Al content. The Pd-H-MOR-D-W A small amount of nitrous oxide (representing a NO con-
sample still contained some EFAL species (1.3 Al/u.c.) as version less than 2% ) is detected in the 400–300ꢁC range.
revealed by 27Al MAS NMR.
These results suggest modifications of the catalyst under re-
The crystallinity of the mordenite samples was mea- actants favoring the formation of sites responsible for the
sured from XRD measurements and determination of mi- nitric oxide reduction.
croporous volumes using nitrogen adsorption–desorption
The influence of oxygen concentration on the nitric ox-
isotherms at 77 K. The results, reported in Table 1, have ide reduction is examined by switching to the lean feed
shown a good agreement between both techniques. Dealu- immediately after the test in stoichiometric conditions and
mination by steaming–HCl cycles induced 18% loss of crys- running the same temperature program. Upon temperature
tallinity. Final leaching with (NH4)2SiF6 was accompanied increase, conversions of CH4 and NO (into N2) are much
by a small increase of crystallinity, fairly consistent with the higher than those measured in stoichiometric conditions.
partial removal of amorphous EFAL species.
However, according to the selectivity plot, the combustion
of methane by O2 is slightly favored under lean conditions,
compared to stoichiometric ones, and some N2O forms (NO
conversion into N2O less than 3% ) between 300 and 450ꢁC.
As observed with the stoichiometric mixture, the combus-
tion of methane by O2 is favored at higher temperatures.
Maintaining the temperature for 10 h does not affect the
NO conversion while the methane conversion decreases.
As a result the selectivity for SCR slightly improves. Upon
decreasing the reaction temperature, the NO conversion
curve exactly superimposes the one obtained upon temper-
ature rise, while the methane conversion diminishes more
slowly with temperature. Again N2O forms between 450
and 300ꢁC (less than 5% NO conversion).
2. Catalytic Activity of Pd-MOR Catalysts
2.1. Pd-H-MOR sample. The conversions of methane
into CO2 and nitric oxide into nitrogen over Pd-H-MOR are
reported as a function of temperature in Figs. 1A and 1B.
The selectivity in the nitric oxide reduction (SSCR) defined
as the fraction of methane involved in the SCR of NO by
CH4 (reaction 1) is plotted in Fig. 1C. From the experi-
ment performed under stoichiometric conditions, i.e. im-
mediately after activation in O2 at 500ꢁC, conversions of
NO and CH4 are observed to increase continuously with
the reaction temperature without the formation of nitrous
FIG. 1. Reaction of CH4, NO, and O2 over Pd-H-MOR catalyst as a function of the temperature: A, CH4 to CO2 conversion; B, NO to N2 conversion;
C, SCR selectivity for NO reduction as defined in the text. Full lines: lean mixture (2000 vpm NO, 1000 vpm CH4, 6240 vpm O2 in He). Broken lines:
stoichiometric mixture (2000 vpm NO, 1000 vpm CH4, 1000 vpm O2 in He). GHSV: 30,000 hꢃ1
.

356
DESCORME ET AL.
at 500ꢁC, corresponding to the total consumption of O2. The
conversion of NO into N2 shows a volcano curve, the max-
imum conversion of 25% being reached at 460ꢁC. Beyond
this temperature, traces of N2O are detected. The selectivity
for NO reduction (Fig. 4C) exhibits a continuous decrease
with reaction temperature, indicating that combustion is
more and more favored. At 500ꢁC, the conversion of NO
into N2O increases from 2 to 8% with time on stream, while
the formation of N2 is stable. Upon decreasing the temper-
ature, the formations of N2, N2O, and CO2 decrease.
On changing for lean conditions, the combustion is fa-
vored (Fig. 4C) and the SCR selectivity almost constant in
the whole interval of temperatures. N2O forms from 350ꢁC
up to 500ꢁC (7% NO conversion on average) and its con-
FIG. 2. Influence of water addition on CH4 and NO conversions over
Pd-H-MOR catalyst as a function of temperature for stoichiometric mix- centration isstronglyincreased, compared to stoichiometric
ture. Open symbols: dry reactants; full symbols: wet reactants (10 vol%
conditions. The conversion curves obtained upon cooling
show the same trend as upon temperature rise, with slightly
lower conversion values. This indicates that the catalyst
has not been significantly modified with time on stream at
500ꢁC.
water). 2000 vpm NO, 1000 vpm CH4, 1000 vpm O2 in He, GHSV:
30,000 hꢃ1
.
To evaluate the effect of water vapor on the reduction of
NO by methane over Pd-H-MOR, 10 vol% water vapor was
added to the stoichiometric feed and the conversions of NO
and CH4 measured between 250 and 500ꢁC. The results are
compared to those obtained in the absence ofwater in Fig. 2.
The selectivityplots are reported in Fig. 3. Surprisingly, both
conversions are enhanced when water is added to the feed.
Moreover, as shown in Fig. 3, the catalytic activity of Pd-
H-MOR in reduction of nitric oxide is significantly favored
by the presence of water. This behavior strongly contrasts
with that of a Pd-H-ZSM-5 catalyst (0.5 wt% Pd), for which
adding 10 vol% water vapor to the feed markedly inhibits
the reduction of NO (Fig. 3) by methane (25, 33).
3. Transmission Electron Microscopy
In both Pd-H-MOR and Pd-H-MOR-D-W samples, large
Pd (or PdO) particles can be observed on the outer surface
of mordenite crystallites. In Pd-H-MOR, their size is about
1–2 nm while bigger and fewer particles (5–30 nm) are ob-
served in Pd-H-MOR-D-W.
4. IR Study of NO Adsorption
4.1. Adsorption on Pd-H-MOR. Figure 5 shows the IR
spectra of 0.4 Torr NO adsorption at 30ꢁC on Pd-H-MOR
activated in O2 and evacuated at 500ꢁC. Three bands
form at 1906, 1875, and 1840 cmꢃ1, together with a broad
asymmetric band at 2236 cmꢃ1 of smaller intensity. The
same band at 2236 cmꢃ1 forms upon adsorption of NO2
onto H-Mordenite at room temperature. Its formation
parallels that of the 2136 cmꢃ1 band upon adsorption of
NO on Pd-H-ZSM-5 catalysts (27, 34, 36). It is therefore
suggested to relate the 2236 cmꢃ1 band to adsorbed NO2
in interaction with mordenite protons. The bands at 1906,
1875, and 1840 cmꢃ1 do not form when adsorbing NO on
the H-MOR support. Evacuating the Pd-H-MOR sample
under vacuum at room temperature does not affect the
spectrum. It is concluded that the bands at 1906, 1875, and
1840 cmꢃ1 are related to NO irreversibly bonded to Pd
sites. Curve fitting of the spectrum shown in Fig. 5 was per-
formed assuming gaussian peaks. Four distinct bands are
necessary to obtain a satisfactory curve fitting. The band
at 1878 cmꢃ1 (type A, HWHM = 19 cmꢃ1) was already
observed upon NO adsorption on Pd-H-ZSM-5 catalyst
(27, 34, 36). This band is attributed to the NO stretching
vibration of Pd mononitrosyl complexes anchored to the
zeolite framework. The bands at 1908 and 1839 cmꢃ1 (types
B1 and B2, HWHM = 7.5 and 8.5 cmꢃ1, respectively) are
2.2. Pd-H-MOR-D-Wsample. Figures4A and 4B show
the conversions of CH4 into CO2 and NO into N2 as a func-
tion of temperature over Pd-H-MOR-D-W under stoichio-
metric and lean conditions. In stoichiometric conditions,
the conversion of CH4 increases up to a maximum (63% )
FIG. 3. Influence of water addition on the SCR selectivity in NO
reduction over Pd-H-MOR (᭿, ❏) and Pd-H-ZSM-5 (᭹, ❍) catalysts;
open symbols: dry reactants; full symbols: wet reactants (10 vol% water).
2000 vpm NO, 1000 vpm CH4, 1000 vpm O2 in He, GHSV: 30,000 hꢃ1
.

CATALYTIC REDUCTION OF NITRIC OXIDE
357
FIG. 4. Reaction of CH4, NO, and O2 over Pd-H-MOR-D-W catalyst as a function of the temperature: A, CH4 to CO2 conversion; B, NO to N2
conversion; C, SCR selectivity for NO reduction as defined in the text. Full lines: lean mixture (2000 vpm NO, 1000 vpm CH4, 6240 vpm O2 in He).
Broken lines: stoichiometric mixture (2000 vpm NO, 1000 vpm CH4, 1000 vpm O2 in He). GHSV: 30,000 hꢃ1
.
specific of the Pd-H-MOR catalysts. Owing to their small were required for fitting all the spectra of adsorbed NO on
half-width at half-maximum, they must be related to well- our Pd-exchanged mordenite samples, whatever the treat-
defined Pd(NO)x complexes with x = 2 or 3. The band at ment preceding the adsorption was (vide infra).
1843 cmꢃ1 (type C) is also attributed to NO adsorbed on Pd
The influence of the NO adsorption temperature was
sites. Its broadness (HWHM = 44 cmꢃ1) suggests multiple evaluated by subsequently heating the sample in 0.4 Torr
Pdx+ sites, possibly coordinatively unsaturated Pdx+ sites NO at 100, 150, and 200ꢁC. The IR spectra were recorded
at the surface of PdO small aggregates or particles. It must at room temperature after each heating step and decon-
be kept in mind that all the studied mordenite samples voluted. A quantitative evaluation of the variation of inte-
contain large PdO particles. Hoost et al. have reported that grated intensity of the four different lines as a function of
the adsorption of NO on PdO supported on alumina gives heating temperature in NO is reported in Fig. 6. Clearly, the
rise to broad absorption in the 1900–1700 cmꢃ1 region (37). intensity of the A band significantly increased upon heat-
Two additional bands of very weak intensity, a shoulder ing up to 100ꢁC and remained constant upon further heat-
at ca 1780 cmꢃ1 and a very broad band between 1700 and ing. Simultaneously the intensity of the doublet (B1 and B2
1600cmꢃ1 can also be observed. The former isnot attributed
while the latter could be possibly related to nitrate/nitrite
groups resulting from the reaction of NOx with the zeolite
surface (27) and/or adsorbed water. These two bands have
been neglected in the curve fitting. It must be pointed out
that the same four lines with about the same parameters
FIG. 6. Adsorption of NO onto the Pd-H-MOR sample as a func-
tion of the temperature: integrated intensities of the A, B, and C type
bands (NO pressure = 0.02 Torr). Full symbols: spectra recorded at room
FIG. 5. FTIR spectrum of NO adsorbed at room temperature onto
the Pd-H-MOR sample (NO pressure = 0.02 Torr). Experimental (❍)
and fitted spectra (———).
temperature; open symbols: spectra recorded at the temperature of ad-
sorption; (᭿, ❏) Type A band (1878 cmꢃ1); (᭹, ❍) Type B (B1 + B2) band
(1908 + 1839 cmꢃ1); (᭡, ∇) Type C band (1843 cmꢃ1).

358
DESCORME ET AL.
bands) slightly decreased while the broad C band was not trum is dominated by the sharp intense doublet (type B
affected. It is deduced that the formation of Pd mononitro- bands) characteristic of the Pd(NO)x (x = 2 or 3) com-
syl complexes is favored by increasing the temperature of plexes while the type A band forms with a smaller intensity.
NO adsorption.
Increasing adsorption time causes a decrease of the doublet
The spectra were also recorded in situ at 100, 150, and and the simultaneous increase of the type A band. Two in-
ꢁ
–
200 C in order to evaluate the strength of the Pd NO bond variant points can be observed where all the curves cross
of the various nitrosyl species. The integrated intensity of perfectly. An isosbestic point is characteristic of a system
the various bands is compared to that obtained at room where only two species are present and involved in a stoi-
temperature after the same heating step in Fig. 6. Taking chiometric interconversion process. Considering that the
into account the uncertainty of the curve fitting, no signif- 1908 and 1840 cmꢃ1 bands characterize the same Pd(NO)x
icant variation of intensity of the types A and C bands is (x = 2 or 3) species, two isosbestic points are expected as
observed following heating and subsequent cooling down observed experimentally, characteristic of the interconver-
to room temperature. On the contrary, the intensity of the sion process between Pd(NO)x (x = 2 or 3) species and Pd
doublet (B1 + B2 bands) progressively decreases with in- mononitrosyl species.
creasing temperatures, compared to the intensity measured
The spectra shown in Fig. 7A were all resolved into four
at room temperature. This points to the lower thermal sta- distinct gaussian lines as indicated above. The integrated in-
bility of the species at the origin of the B1 and B2 bands, tensities of these bands are reported in Fig. 7B as a function
compared to Pd mononitrosyl complexes and to NO ad- of the adsorption time. It will be assumed that the extinction
sorbed onto Pd or PdO particles.
coefficient of NO adsorbed in the various complexes is the
After evacuation at 250ꢁC, a broad band is observed same. Clearly the intensity of type C ꢂNO band does not
and can be resolved into a band at 1880 cmꢃ1 (type A, vary with adsorption time. It can be also observed that the
HWHM = 15 cmꢃ1) and a broad line at 1846 cmꢃ1 (type mononitrosyl Pd species (type A ꢂNO band) forms exactly
C, HWHM = 46 cmꢃ1) with the same relative intensities at the expense of the Pd(NO)x ones (B1 + B2 intensities).
as before evacuation. Pd dinitrosyl complexes have been to- These two plots can be fitted with an excellent agreement
tally decomposed as expected from their low thermal stabil- by log functions of the type AA log(time) + KA and ꢃAB
ity and the remaining adsorbed NO species are removed in log(time) + KB. Since the interconversion process between
the same proportions. The integrated intensity of the ꢂNO the two nitrosyl Pd complexes can be written as
bands after evacuation at 250ꢁC represents 25% of the in-
tensity of the massif before evacuation, which indicates that
most of the adsorbed NO has been removed from the sam-
Pd(NO)x → Pd(NO) + (x ꢃ 1)NO
ple. The sample was then contacted again with 0.02 Torr NO with x = 2 or 3, by deriving the two previous functions
at 30ꢁC. The spectra were recorded at increasing times and the rate of appearance of the A band (AA/time) and dis-
an expanded view in the 2000–1700 cmꢃ1 region is shown appearance of the B bands (AB/time) can be obtained.
in Fig. 7A. Immediately after contact with NO, the spec- A ratio of AB/AA equal to 1.8 was found supporting the
FIG. 7. Adsorption of NO at room temperature onto the Pd-H-MOR sample: A, FTIR spectra. NO (pressure = 0.4 Torr) was first adsorbed at
200ꢁC, then evacuated at 250ꢁC. NO was subsequently adsorbed at room temperature (P = 0.02 Torr) and spectra were recorded as a function of time
after 50 s (a), 90 s (b), 640 s (c), and 2260 s (d). The arrows indicate the invariant or isosbestic points. B, Evolution with time of the integrated intensities
of the A, B, and C type bands: ❏, Type A band (1878 cmꢃ1); ᭿, Type B (B1 + B2) band (1908 + 1839 cmꢃ1); ❍, Type C band (1843 cmꢃ1).

CATALYTIC REDUCTION OF NITRIC OXIDE
359
still present with Pd-H-ZSM-5, has a much smaller inten-
sity than on our Pd-H-MOR catalysts. This could be due to
a higher dispersion of Pd in Pd-H-ZSM-5 catalysts because
of a lower Pd loading. It was recently shown that increas-
ing the Pd loading beyond 0.5 wt% Pd favors the forma-
tion of large PdO particles on the outer surface of ZSM-5
crystallites in addition to isolated Pd ions dispersed in the
zeolite channels (34). All our Pd-H-MOR samples exhibit
a significant amount of large Pd/PdO particles detected by
electron microscopy. This strongly supports the attribution
of the 1840 cmꢃ1 band to NO adsorbed on Pd/PdO particles.
The formation ofdinitrosyl Pd complexesisspecific ofthe
mordenite structure. These complexes are characterized by
two ꢂNO sharp vibrations (type B) at 1908 and 1839 cmꢃ1
,
corresponding respectively to the symmetric and antisym-
metric vibrations associated to Pd complex in a C2v sym-
metry. As for the gem dicarbonyl complexes, the relative
intensity of the two bands is related to the angle (ꢄ) be-
tween the two N-O oscillators (38):
FIG. 8. FTIR spectra of NO adsorbed at room temperature onto the
Pd-H-MOR-sample (a) and Pd-H-MOR-D-W (b) sample under a NO
pressure of 0.02 Torr.
hypothesis of an interconversion process between mono-
and di-nitrosyl Pd complexes. The NOꢃ2+ species are simul-
taneouslyobserved to slightlyincrease with adsorption time
but this variation does not correlate with the formation of
the mononitrosyl Pd complexes.
I
1908/I1839 = [cotan(ꢄ/2)]2.
An invariant ratio of the doublet peaks is found in our
experiment of NO readsorption on Pd-H-MOR evacuated
at 250ꢁC after NO adsorption. Its value of 0.43 corresponds
ꢁ
4.2. Adsorption on Pd-H-MOR-D-W. The IR spec-
trum in the 2000–1600 cmꢃ1 region obtained upon 0.02 Torr
NO adsorption at 30ꢁC on Pd-H-MOR-D-W activated in O2
and evacuated at 500ꢁC is shown in Fig. 8b. It is dominated
by the band (type A) at ca 1875 cmꢃ1 while the doublet
(type B bands) at ca 1908–1840 cmꢃ1 has a much smaller
intensity than for Pd-H-MOR (Fig. 8a). It can be deduc-
ed that dealumination strongly inhibits the formation of
Pd dinitrosyl complexes while favoring the formation of
mononitrosyl species. As a result, the integrated inten-
sity of the A type band is 30% higher on Pd-H-MOR-
D-W (16.3 ꢅ 103 cmꢃ1 gꢃ1 Pd) than on Pd-H-MOR (12.7 ꢅ
103 cmꢃ1 gꢃ1 Pd). The appearance of bands due to adsor-
bed NO is accompanied by the band at 2236 cmꢃ1 relat-
ed to adsorbed NO2 as for Pd-H-MOR.
–
–
to a NO Pd NO angle of 113 . This value is comparable to
ones obtained with other dinitrosyl species reported in the
literature: 123ꢁ on Co-Y (39), 103ꢁ on Cu-ZSM-5 (40).
The question arises to explain the formation of stable
dinitrosyl Pd complexes in the mordenite structure and the
interconversion process between these complexes and the
Pd mononitrosyl complexes. A possible explanation might
originate in the pore structure of the mordenite, made of
main parallel elliptical channels (0.65 ꢅ 0.70 nm) and an
interconnecting array of smaller channels, so-called side
pockets, presenting a restricted aperture of about 0.28 nm
in diameter (41, 42).
Accordingly, Pd2+ ions might occupy the main channels
and side pockets. Upon NO adsorption, the former are ex-
pected to form Pd mononitrosyl complexes with a struc-
ture similar to that obtained in the straight channels of the
H-ZSM-5 while the latter sites would be responsible for the
formation of dinitrosyl complexes (not formed in H-ZSM-5
for which no such sites exist). The distribution of the Pd2+
DISCUSSION
1. NO Species Adsorbed on Pd-H-MOR
From the above IR results of NO adsorption on Pd- cations between side pockets and main channels depends,
exchanged mordenites, three distinct Pd nitrosyl species of course, on the repartition of aluminum between these
have been identified by their ꢂNO vibrations (type A at two types of sites. For a dealuminated MOR matrix, the
ca 1878 cmꢃ1, type B at ca 1908 and 1839 cmꢃ1, and type formation of Pd dinitrosyl complexes is strongly inhibited.
C at ca 1843 cmꢃ1). Two of them form also in the Pd-H- A first hypothesis could be that dealumination induces a
ZSM-5 catalysts, namely the species related to the 1880 preferential extraction of aluminum atoms responsible for
and 1840 cmꢃ1 bands. The 1880 cmꢃ1 band (type A) was exchange sites in side pockets. Another explanation would
related to a Pd mononitrosyl complex anchored to the zeo- be the blockage of these sites by EFAL species not removed
lite framework and sitting as isolated species in the channels upon leaching treatments. This would explain why the
of the zeolite by analogy with the results obtained on Pd- Pd-H-MOR-D-W sample exhibits almost no formation of
H-ZSM-5 catalysts (27, 34). The 1840 cmꢃ1 band, although Pd dinitrosyl complexes.

360
DESCORME ET AL.
The Pd dinitrosyl complexes, because of their localization eral factors such as the Pd loading, the acidic properties of
in the small cavities, are expected to be coordinatively satu- the zeolitic support or the activation conditions of the cata-
rated with the surrounding lattice oxygen atoms acting as lyst. The latter factor might explain the changes of catalytic
ligands. The conversion process of the dinitrosyl complexes behavior observed for the non-dealuminated Pd-H-MOR
into mononitrosyl complexes under NO at room tempera- catalyst during the first reaction cycle (stoichiometric mix-
ture requires the substitution of the lattice oxygen ligands ture). The activity both in NO and CH4 conversions in-
by NO, NO2, or H2O, allowing their migration towards large creases with time on stream at 500ꢁC (Figs. 1A and B). The
channels and their subsequent stabilization as mononitrosyl changes in SCR values show that the NO reduction is fa-
complexes.
vored, which indicatesan increase in the number ofcatalytic
sites responsible for SCR, i.e. isolated Pd ions. This would
suggest the partial redispersion of PdO particles under re-
actants, in accordance with recent works (43–44). X-ray ab-
sorption experiments on supported Pd catalysts have shown
that the metallic Pd particles initially present on acidic sup-
ports (H-ZSM-5, H-MOR, or sulfated zirconia) are rapidly
transformed into Pd2+ ions under a CH4 + NO + O2 mix-
ture (43). The reaction would be assisted by the protonic
acidity of the support. Of course, this mechanism could ap-
ply only to the PdO particles of small size and dispersed
inside the pores of our Pd-H-MOR catalyst, the large par-
ticles observed on the outer mordenite surface remaining
intact. It must be mentioned that no such changes are ob-
served during the second cycle (lean cycle), which indi-
cates that the catalyst has been stabilized during the stoi-
chiometric cycle. The observed decrease of the selectivity
for SCR upon changing from stoichiometric to lean condi-
tions is indicative of the Pd2+/PdO ratio, PdO particles be-
ing responsible for CH4 combustion activity in oxygen-rich
mixtures.
The hypothesis that PdO particles are redispersed into
Pd ions under reactant is, however, contradicting our re-
sults obtained over Pd-H-ZSM-5 catalysts (34). We have
shown that there exists a relationship between the Pd load-
ing, the Si/Al ratio of the H-ZSM-5, and the Pd dispersion:
the higher the Pd loading the lower the Pd dispersion. For
Pd loading exceeding 0.5 wt% Pd, large particles are ob-
served by electron microscopy and the selectivity of the
catalyst for SCR is strongly reduced. Moreover, the ac-
tivity for NO reduction decreases with time on stream at
temperatures above 500ꢁC, which suggests the transforma-
tion of isolated Pd ions (responsible for SCR) into PdO
particles favoring the methane combustion. It would be ex-
pected upon these results that, owing to its high Pd loading
(1 wt% ), the Pd-H-MOR would exhibit a deactivation with
time on stream similar to that of the Pd-H-ZSM-5 catalysts.
Surprisingly, an increase of the activity is observed. An al-
ternative explanation to the change in the catalytic behav-
ior of the Pd-H-MOR catalyst under reactants might be
given by the interconversion process between mono- and
di-nitrosyl species evidenced by our IR results. The Pd2+
ions dispersed in the mordenite structure might occupy two
different sites and form two different interconverting nitro-
syl complexes upon reaction with NO while only one type
of Pd nitrosyl complex (mononitrosyl species) forms in the
2. Relationship between Pd Sites and Catalytic
Behavior of Pd-H-MOR Catalysts
Pd-exchanged H-ZSM-5 catalysts have been shown to be
highly active to convert NOx into N2 in the presence of O2
(24–31, 34, 43, 44). At least two reactions are supposed to
compete according to the following scheme:
CH4 + 2 NO + O2 → CO2 + N2 + 2 H2O
CH4 + 2 O2 → CO2 + 2 H2O.
[1]
[2]
Reaction [1] corresponds to the reduction of NO by me-
thane, it involves O2 as reactant since NO2 is suggested as
a reaction intermediate and since oxygen has a promoting
effect on the NO reduction. Reaction [2] corresponds to the
direct combustion of methane by O2. Very different cata-
lytic properties have been observed, depending on the Pd
loading (28, 34). At low Pd loading, reaction [1] (NO re-
duction) is favored and the reaction temperature has little
influence on the selectivity for NO reduction. Conversely,
higher Pd loading induces the increase of the selectivity
for reaction [2] (methane combustion). Moreover, increas-
ing reaction temperatures favor the combustion reaction.
This behavior was attributed to the presence of PdO par-
ticles located on the outer surface of zeolite crystals. The
high selectivity of the catalyst towards NO reduction in the
presence of excess O2 was attributed to the catalytic prop-
erty of isolated Pd ions and/or complexes anchored to the
zeolite framework and located inside the zeolite channels.
The O2 concentration has no effect on the catalytic activity
of Pd ions.
Our Pd-exchanged mordenite samples do contain both
PdO particles and isolated Pd2+ ions; the presence of large
PdO particles is confirmed by electron microscopy. The in-
creasing proportion of Pd as PdO particles should reveal
the catalytic properties of bulk PdO: (i) lower selectivity
for NO SCR; (ii) dependence of the selectivity for NO
SCR with temperature, the combustion being favored at
increasing temperatures; (iii) the combustion being more
favored under lean conditions, compared to stoichiometric
mixture; (iv) the formation of N2O in the 300–450ꢁC range.
Accordingly the catalytic properties of our Pd-exchanged
mordenite samples reflect the proportion of Pd as isolated
cations and PdO particles. This might be dependent on sev-

CATALYTIC REDUCTION OF NITRIC OXIDE
CONCLUDING REMARKS
361
H-ZSM-5 structure. The latter species, related to the reac-
tion of NO with isolated Pd ions dispersed in the zeolite
(26–27), was thought to be responsible for the high activity
and selectivityofPd-H-ZSM-5catalystsfor SCR in the pres-
ence of O2. It might be assumed that the dinitrosyl species,
ascribed to species sitting in the side pockets of the morden-
ite structure, are inactive for SCR, the side pockets being
less accessible to reactants than main channels. This hypoth-
esis is confirmed by catalytic activity results obtained in the
presence of water. It is shown that the non-dealuminated
Pd-H-MOR catalyst exhibits an enhancement of its cata-
lytic activity for SCR under wet conditions, while the addi-
tion of water to the feed generally induces a severe decrease
of the catalytic activity for Pd catalysts supported on other
supports, even though this decrease is somewhat limited in
the case of Pd-H-ZSM-5, compared to most of the cata-
lysts (25, 33). The loss of activity could be explained by a
competitive adsorption of water on the active site. The sur-
prising catalytic behavior of Pd-H-MOR in the presence
of water can be rationalized by considering that the in-
terconversion process between di- and mono-nitrosyl Pd
species is dependent upon the amount of water in the feed.
Water would be acting as a substituting ligand of the dini-
trosyl complex, favoring the migration of this complex to-
ward main channels and therefore the formation of mono-
nitrosyl complexes, inducing the increase of the catalytic
activity.
The IR study of NO adsorption on Pd exchanged mor-
denite catalysts calcined in O2 at 500ꢁC has revealed the
existence of three distinct states of adsorbed NO, charac-
terized by distinct ꢂNO vibrations:
(i) Pd mononitrosyl complexes vibrating at ca
1880 cmꢃ1, also observed in MFI structures and thought
to be responsible for the catalytic activity in NO reduction
in the presence of excess O2; these complexes could be
anchored to the framework and sit in the main channels of
the mordenite;
(ii) Pd dinitrosyl complexes characterized by a sharp
doublet at 1908–1839 cmꢃ1 and suggested to be inactive in
the catalytic reaction;
(iii) NO adsorbed on Pd/PdO particles characterized by
a broad ꢂNO band centered at ca 1840–1850 cmꢃ1 and
thought to be active in NO reduction but less selective than
isolated Pd mononitrosyl species.
The specific formation of dinitrosyl species in the mor-
denite structure is related to the presence of small cavities
with restricted aperture, which could stabilize Pd cations
but would not allow the reactants to penetrate. A stoi-
chiometric interconversion process between mono- and di-
nitrosylspecieswasclearlyshown. Thisprocessisthought to
be responsible for the specific catalytic behavior of Pd mor-
denite catalyst in wet conditions; the activity is enhanced
by addition of water to the feed.
Pd in dealuminated mordenite. The Pd-H-MOR-D-W
(dealuminated) sample does contain PdO particles as evi-
denced by transmission electron microscopy. After activa-
tion under reactants at 500ꢁC under stoichiometric condi-
tions, the SCR selectivity measured during the decrease
in temperature from 500 to 350ꢁC was found between 0.3
and 0.4 for the dealuminated sample instead of 0.4–0.6 for
the non-dealuminated one in the same conditions (Figs. 1C
and 4C). Such smaller SCR values suggest that the dealu-
minated catalyst exhibits larger amounts of palladium ox-
ide in the form of big particles. It has been proposed for
Pd-H-ZSM-5 catalysts that protons allow the redispersion
of palladium oxide particles into isolated Pdx+ ions (43).
For dealuminated samples, the relative number of protons
compared to the number of palladium atoms is not high
Dealumination of the zeolite structure was shown to fa-
vor the formation of the mononitrosyl species at the ex-
pense of the dinitrosyl complexes and leads to an increase
of NO SCR selectivity under stoichiometric conditions.
REFERENCES
1. Shelef, M., Chem. Rev. 95, 209 (1995).
2. Armor, J. N., Catal. Today 26, 147 (1995).
3. Burch, R., and Millington, P. J., Catal. Today 26, 185 (1995).
4. Li, Y., and Armor, J. N., Appl. Catal. B 1, L31 (1992).
5. Li, Y., and Armor, J. N., Appl. Catal. B 2, 239 (1993).
6. Li, Y., and Armor, J. N., Appl. Catal. B 3, L1 (1993).
7. Li, Y., Battavio, P. J., and Armor, J. N., J. Catal. 142, 561 (1993).
8. Witzel, F., Sill, G. A., and Hall, W. K., J. Catal. 149, 229 (1994).
9. Li, Y., and Armor, J. N., J. Catal. 150, 376 (1994).
enough to limit the growth of the PdO particles outside of 10. Li, Y., Slager, T. L., and Armor, J. N., J. Catal. 150, 388 (1994).
11. Cowan, A. D., Du¨mpelmann, R., and Cant, N. W., J. Catal. 151, 356
the zeolite crystallites.
Nevertheless, the conversion of NO into N2 is higher
in the case of the dealuminated sample compared to the
(1995).
12. Adelman, B. J., Beutel, T., Lei, G.-D., and Sachtler, W. M. H., J. Catal.
158, 327 (1996).
non-dealuminated one. Such behavior is certainly due to
the larger number of A species (mononitrosyl complexes)
found in the Pd-H-MOR-D-W solid.
It seems there is a compensating effect between the total
number of Pd2+ ions and the distribution between the mono
and dinitrosyl species; the mononitrosyl species is favored
in the case of the dealuminated sample.
13. Lukyanov, D. B., d’Itri, J. L., Sill, G., and Hall, W. K., in “11th Interna-
tional Congress on Catalysis—40th Anniversary,” Studies in Surface
Science and Catalysis, Vol. 101 (J. W. Hightower, W. N. Delgass, and
A. T. Bell, Eds.), p. 651. Elsevier, Amsterdam, 1996.
14. Ogura, M., and Kikuchi, E., in “11th International Congress on
Catalysis—40th Anniversary,” Studies in Surface Science and Cata-
lysis, Vol. 101 (J. W. Hightower, W. N. Delgass, and A. T. Bell, Eds.),
p. 671. Elsevier, Amsterdam, 1996.

362
DESCORME ET AL.
15. Aylor, A. W., Lobree, L. J., Reimer, J. A., and Bell, A. T., in “11th
International Congress on Catalysis—40th Anniversary,” Studies 29. Uchida, H., Yamaseki, K., and Takahashi, I., 2nd Japan-EC joint work-
in Surface Science and Catalysis, Vol. 101 (J. W. Hightower, shop, JECAT’95, Catal. Today 29, 99 (1996).
W. N. Delgass, and A. T. Bell, Eds.), p. 661. Elsevier, Amsterdam, 30. Nishizaka, Y., and Misono, M., Chem. Lett. 2237 (1994).
28. Loughran, C. J., and Resasco, D. E., Appl. Catal. B 7, 113, 351 (1995).
1996.
31. Descorme, C., Ge´lin, P., Le´cuyer, C., and Primet, M., Appl. Catal. B
13, 185 (1997).
16. Armor, J. N., and Farris, T. S., Appl. Catal. B 4, L11 (1994).
17. Li, Y., and Armor, J. N., Appl. Catal. B 5, L257 (1995).
18. Li, Y., and Armor, J. N., J. Catal. 145, 1 (1994).
19. Yogo, K., Ihara, M., Terasaki, I., and Kikuchi, E., Chem. Lett. 229
(1993).
20. Kikuchi, E., and Yogo, K., Catal. Today 22, 73 (1994).
21. Tabata, T., Kokitsu, M., and Okada, O., Appl. Catal. B 6, 225 (1995).
22. Kikuchi, E., Ogura, M., Terasaki, I., and Goto, Y., J. Catal. 161, 465
(1996).
32. Almanza, L. O., Ph.D. thesis, Lyon 1 University, 1993.
33. Descorme, C., Ph.D. thesis, Lyon 1 University, 1996.
34. Ge´lin, P., Goguet, A., Descorme, C., Le´cuyer, C., and Primet, M., in
“IVth Intern. Cong. on Catalysis and Automotive Pollution Control,
April 1997, Brussels,” Vol. 2, p. 197.
35. Thomas, J. M., and Klinowski, J., Adv. Catal. 33, 199 (1985).
36. Aylor, A. W., Lobree, L. J., Reimer, J. A., and Bell, A. T., J. Catal. 172,
453 (1997).
23. Tabata, T., Kokitsu, M., and Okada, O., Catal. Lett. 25, 393 (1994).
24. Fakche, A., Pommier, B., Garbowski, E., Primet, M., and Le´cuyer, C.,
French Patent Appl. 93 08 006.
37. Hoost, T. E., and Laframboise, K. A., J. Catal. 155, 303 (1995).
38. Cotton, F. A., and Wilkinson, G., “Advanced Inorg. Chemistry,” IVth
ed., p. 697. Wiley, New York, 1980.
25. Descorme, C., Fakche, A., Garbowski, E., Primet, M., and Le´cuyer,
C., in “1995 International Gas Research Conference, Cannes (France),
6–9 Nov. 1995,” Vol. IV, p. 505.
26. Descorme, C., Ge´lin, P., Primet, M., Le´cuyer, C., and Saint Just, J.,
in Zeolites: A refined tool for designing catalyst, Studies in Surface
39. Windhorst, K. A., and Lunsford, J. H., J. Am. Chem. Soc. 97, 1407
(1975).
40. Iwamoto, M., Yahiro, H., Mizuno, N., Zhang, W. X., Mine, Y.,
Furukawa, N., and Kagawa, S., J. Phys. Chem. 96, 9360 (1992).
41. Breck, D., “Zeolite Molecular Sieves,” Wiley, New York, 1974.
Science and Catalysis, Vol. 97 (L. Bonneviot and S. Kaliaguine, Eds.), 42. Meier, W. M., and Olson, D. H., Zeolites 12, 1 (1992).
p. 287. Elsevier, Amsterdam, 1995.
27. Descorme, C., Ge´lin, P., Primet, M., and Le´cuyer, C., Catal. Lett. 41,
133 (1996).
43. Ali, A., Alvarez, W., Loughran, C. J., and Resasco, D. E., Appl. Catal.
B 14, 13 (1997).
44. Adelman, B. J., and Sachtler, W. M. H., Appl. Catal. B 14, 1 (1997).