Effect of the addition of Na to Pt/Al2O3 catalysts for the reduction of NO by C3H8 and C3H6 under lean-burn conditions

Article abstract of DOI:10.1016/S0021-9517(03)00200-8

A series of Pt/γ-Al₂O₃ catalysts promoted by various concentrations of sodium were prepared and tested for the SCR of NO by propylene and propane under lean-burn conditions. Na addition markedly modified the Pt catalytic behaviors. Under a C₃H₈/NO/O₂ atmosphere, the presence of sodium led to a drastic decrease of the C₃H₈ and NO conversions. In contrast, under C₃H₆/NO/O₂, Na promotion improved the Pt catalytic performance. The sodium effects on activity and selectivity of platinum were correlated to modifications of the strength of the chemical bonds between platinum and the different adsorbates. The opposite effects observed by using propane of propylene were attributed to their different mechanisms of adsorption on platinum. Similarities existed between the catalytic performances of an electrochemically promoted or film deposited on a NASICON electrolyte and a powdered Pt-supported alumina promoted by chemical additions of sodium. Whatever the mode of the Na addition (electrohemical or chemical), the interpretations of the Na effects operated in the same direction, i.e., a strengthening of the Pt-electron-acceptor (O and NO) bonds and a weakening on the Pt-electron-donor (C₈H₈ and C₃H₆) bonds.

Full text of DOI:10.1016/S0021-9517(03)00200-8

Journal of Catalysis 219 (2003) 247–257
www.elsevier.com/locate/jcat
Effect of the addition of Na to Pt/Al O catalysts for the reduction
2
3
of NO by C H and C H under lean-burn conditions
3
8
3 6
∗
P. Vernoux, A.-Y. Leinekugel-Le-Cocq, and F. Gaillard
Laboratoire d’Application de la Chimie à l’Environnement (LACE), UMR 5634 CNRS-Université Claude Bernard Lyon 1, 69622 Villeurbanne cedex, France
Received 3 February 2003; revised 14 April 2003; accepted 15 April 2003
Abstract
A series of Pt/γ -Al O catalysts promoted by various concentrations of Na were prepared and tested for the SCR of NO by propene and
2
3
propane under lean-burn conditions. The addition of Na markedly modifies the Pt catalytic behaviors. Under a C H /NO/O atmosphere,
3
8
2
the presence of Na leads to a drastic decrease of the C H and NO conversions. In contrast, under C H /NO/O , Na promotion improves the
3
8
3
6
2
Pt catalytic performance. For high Na loadings (5 wt%), we observed an enhancement of the selectivity to N and a widening of the operating
2
temperature window. The promotional phenomena observed have been explained in terms of modification of the adsorption strengths of the
various reactants on the metal surface. The Na effects should be to strengthen the Pt–electron-acceptor (O and NO) bonds and to weaken the
Pt–electron-donor (C H and C H ) bonds. The opposite effects observed by using propane or propene have been attributed to their different
3
8
3 6
mechanisms of adsorption on Pt.
2003 Elsevier Inc. All rights reserved.

Keywords: Selective catalytic reduction; SCR; NOx; Propene; Propane; Platinum; Sodium; Lean-burn conditions
1
. Introduction
creasing in Europe. Selective catalytic reduction (SCR) of
NOx by hydrocarbon under lean-burn conditions has been
the subject of intensive research effort since it was demon-
strated that hydrocarbons could be used as reducing agents.
Among many investigated catalyst formulations, supported
platinum-based catalysts have shown the highest activities
for NOx removal in real Diesel exhaust gas. Furthermore,
The pollution coming from automotive traffic is an in-
creasing problem, especially in urban areas. Intensive re-
search on the posttreatment of the exhaust gas led to the
development of the three-way catalytic converters (TWC)
which are simultaneously able to reduce NOx (NO and NO2)
into nitrogen and to oxidize carbon monoxide and unburned
hydrocarbons. These catalytic systems are efficient in the
case of engines that are operated with a fuel/air ratio close to
the stoichiometry, as the conventional gasoline ones. How-
ever, under lean-burn conditions such as encountered with
diesel and lean-burn gasoline engines, TWC are ineffective
to reduce NOx into nitrogen. Future legislations, such as
EURO 4 standards in Europe, which will be active from
◦
this metal is active at low temperatures (between 200 C and
◦
350 C) corresponding to that of a Diesel engine working
in an urban cycle. Supported platinum catalysts are resistant
to poisoning by steam or SO2 present in the exhaust stream
and exhibit a good thermal stability. However, these catalysts
have two limitations, high selectivity to N2O and a narrow
temperature window of activity. The development of the se-
lective NOx reduction by hydrocarbons is hindered by the
production of N2O because it is a powerful greenhouse gas,
200 times the worth of CO2. Much work tried to improve
the selectivity to N2 of platinum by varying the nature of the
support [1], the platinum dispersion [1], and the preparation
method [2] or by using various hydrocarbons [3]. Up to the
present time, there is no practical effective solution.
2
005, make necessary the posttreatment of the Diesel ex-
haust gas. Then, the development of an effective solution
for NOx removal under excess oxygen conditions is vital
to the future of Diesel cars, the use of which is ever in-
*
Corresponding author. LACE, UMR 5634 CNRS, Université Claude
Some interesting results have been obtained by adding
alkalis to platinum supported on γ -Al2O3. Most of these
studies have been carried with low concentrations of oxygen,
Bernard Lyon 1, Bat. Raulin/1, 43 Blvd du 11 Novembre 1918, 69622
Villeurbanne cedex, France
E-mail address: philippe.vernoux@univ-lyon1.fr (P. Vernoux).
0021-9517/$ – see front matter  2003 Elsevier Inc. All rights reserved.
doi:10.1016/S0021-9517(03)00200-8

248
P. Vernoux et al. / Journal of Catalysis 219 (2003) 247–257
near to the stoichiometry [4], or in the absence of oxygen
5,6]. Konsolakis et al. [4] have shown that the catalytic ac-
catalysts [5]. Indeed, pumping sodium from the electrolyte
to the Pt surface causes a drastic increase of the NO con-
version and the selectivity to N2. At high loadings of Na,
the system exhibits poisoning. We have recently shown [14]
that electrochemical promotioncan strongly enhance the cat-
alytic activity and the selectivity to N2 of Pt interfaced with
NASICON for the NO reduction by propene under lean-burn
conditions. NASICON, i.e., Na3Zr2Si2PO12, presents a high
[
tivity and selectivity of alumina-supported catalysts for the
reduction of NO by propene are strongly promoted by alkali
metals. At the stoichiometry, the addition of Na significantly
increases the selectivity to N2. The influence of sodium on
the catalytic properties of Pd and Rh for the NO reduction by
propene near the stoichiometry has also been demonstrated.
The addition of Na to Pd/γ -Al2O3 catalysts operated under
simulated TWC conditions has a beneficial effect on both
NO conversion and selectivity to N2 [7]. When using Rh
as a catalyst, doping by sodium promotes the NO reduc-
tion by C3H6 in the absence of oxygen [8] and inhibits it
in the presence of oxygen, under TWC conditions [9]. The
Na-induced effect is explained in terms of variation of the
electronic density of the metal and consequently of the ad-
sorption strengths of the various reactants onto its surface.
Under lean-burn conditions, Tanaka et al. [10] have shown
that the addition of small amounts of Na (0.1 wt%) does not
change the catalytic properties of Pt-supported SiO2 cata-
lysts. Burch et al. [11] have shown that the addition of K, Cs,
and Ba to Pt/Al2O3 catalysts does not improve the catalytic
performance as the selectivity to N2, even it reduces the NOx
conversion. Recently, Burch and Coleman [12] have inves-
tigated the reduction of NO by H2 under lean-burn condi-
tions. They found that the addition of low loadings of sodium
+
conductivity by Na ions even at low temperatures.
The aim of the present paper is to elucidate the influence
of a chemical addition of sodium to Pt/γ -Al2O3 catalysts
for the SCR of NO by hydrocarbons under lean-burn con-
ditions. Our objective is to correlate the promoting effect
on the selectivity to N2 and on NO reduction induced via
electrochemical promotion [14] to a chemical Na doping of
a supported Pt/alumina catalyst. Alumina is a good candi-
date as a support because of its thermal stability and high
surface area. We have selected two hydrocarbons, propane
and propene, which are major constituents of the hydrocar-
bon components of automotive exhaust. Furthermore, these
two reductants, despite similar steric factors, exhibit quite
different behaviors according to their strength of adsorp-
tion on platinum. Indeed, several studies [4,15–17] have
shown that alkenes adsorb strongly and compete with oxy-
gen, whereas alkanes adsorb weakly and therefore without
any significant adsorption competition with oxygen. Propene
adsorbs strongly on platinum because of the interaction of
its π-orbitals with Pt d-bands [18]. Several studies have
shown that π-bonded olefins are the central species in the
mechanism of catalytic hydrogenation reactions [19,20]. For
propane oxidation, the limiting step would be the cleavage of
a primary C–H bond [21]. The greater reactivity of alkanes
on the more open (1 × 2) reconstructions of Pt (110) com-
pared to the closed-packed (111) surfaces of these metals is
believed to result from the presence of a higher concentration
of sites on (110) surfaces which promote alkane dissocia-
tion. Other reactions involving alkane reactants, including
hydrogenolysis [22,23], are also reported to be structure sen-
sitive.
(
0.27–1 wt%) to Pt/Al2O3 catalysts increases the NO con-
version whereas large quantities of alkalis (5–10 wt%) de-
◦
creases it. At low temperatures (90–150 C), the selectivity
to N2 does not vary with the addition of sodium while at
higher temperatures, it decreases. These studies seem to in-
dicate that the dissociation of NO is not enhanced under
lean-burn conditions to the same extent as was found by
Konsolakis et al. [4] at the stoichiometry point.
The effect of the interaction between alkalis and metals
on the catalytic properties has also been investigated by us-
ing the electrochemical promotion (EP). In this concept, the
metallic catalyst is deposited, as a thin porous layer, onto a
dense substrate of solid electrolyte which is a sodium con-
ꢀꢀ
ductor, as β -alumina or NASICON (NA Super Ionic CON-
ductor). By polarizing the metal/solid electrolyte interface
This study reports the catalytic behaviors of sodium-
doped Pt/γ -Al2O3 catalysts for the NO reduction under
lean-burn conditions. Catalytic performances of various
catalysts were compared according to the Na loadings
and the nature of the hydrocarbon, propane, or propene.
Temperature-programmed desorption (TPD) measurements
were carried out in order to determine the influence of the
Na content on the oxygen reactivity with Pt.
+
with respect to a reference electrode, Na ions are pumped
from the electrolyte to the surface of the catalyst. EP pro-
vides an in situ control of the promoter concentration at the
surface of the working metal catalyst by varying the applied
potential. Thus, one may control reversibly metal coverage
by alkali ions. Furthermore, the total amount of sodium,
which is pumped out from the electrolyte structure, inter-
acts with the catalyst. On the contrary, in the case of highly
dispersed metal catalysts on an oxide support such as alu-
mina or silica, a large part of chemically added sodium is in
contact with the support only and not to the catalytically ac-
tive phase. Yentekakis et al. [13] have studied the EP by Na
of the Pt-catalyzed reduction of NO by propene in the ab-
sence of oxygen. Their results are in good agreement with
those obtained with a chemical Na doping of Pt/γ -Al2O3
2. Experimental
2.1. Preparation of the catalysts
Catalysts were prepared from a commercial γ -Al2O3
(Rhodia-SPH569) containing traces of α-Al2O3. Four cat-
alysts, containing different loadings of sodium, have been

P. Vernoux et al. / Journal of Catalysis 219 (2003) 247–257
249
synthesized by the impregnation method, as follows. The
support, γ -Al2O3, was suspended in aqueous solutions con-
taining various appropriate concentrations of NaOH (Pro-
labo, 99% purity). Water was removed under reduced pres-
sure at 60 C. After drying overnight in air at 120 C and
crushing, the resulting powder was impregnated with a solu-
tion of PtCl4 (Alfa Aesar, 99.9% purity) in order to yield
was used as the vector gas. The gas composition was con-
trolled by mass-flow controllers (Brooks, with accuracies
better than 1%). In the following, the gas composition will
be expressed in parts per million or percentage. The reac-
tive mixture is a simulated diesel exhaust gas containing
HC/NO/O2: 2000 ppm/2000ppm/5%, with HC = C3H6 or
C3H8. The overall gas-flow rate was kept with He constant
◦
◦
−1
−1
0
.9 wt% Pt metal loading. Water was removed as in the
at 10 L h (± 0.1 L h ). The corresponding space veloc-
−1
first step. Then, the powdered catalyst was dried overnight
ity is 20,000 h . The reactants and products were analyzed
by a NOx (NO and NO2) analyser (Cosma, Topaze 3000)
and an on-line microgas chromatograph (Varian CP2003).
Before these analyses, the water produced by the catalytic
reactions was trapped by an electrogas cooler (M&C Prod-
ucts, ECP1000-G). The microchromatograph was composed
of two modules, each one equipped with a thermal conduc-
tivity microdetector. The first module enabled the separation
of CO2, N2O, and the hydrocarbon(C3H6 or C3H8) by using
a Poraplot U column (10 m length and 0.32 mm in inter-
◦
at 120 C, crushed, and calcinated under flowing oxygen
(
−1
◦
10 L h ) at 500 C for 2 h in order to perform precursor
decomposition. After cooling and purging under nitrogen,
−
1
the catalyst was reduced under flowing hydrogen (10 L h
)
◦
at 300 C for 2 h. The BET surface of the samples was mea-
2
−1
sured by N2 adsorption at 77 K and found to be 135 m g
which is the area of the support. Chemical analysis of Pt
and Na was performed by ICP. Chlorine was analyzed by
ion-capture chromatography. The percentage of accessible
metallic fraction (AMF) was determined by H2–O2 titration.
◦
nal diameter) heated at 60 C. The second one was equipped
◦
with a molecular sieve heated at 90 C (10 m length and
2
.2. Temperature-programmed desorption of oxygen
0.32 mm in internal diameter) preceded by a back-flush col-
umn. This module analyzed O2, N2, and CO. The overall
duration of one analysis was only 110 s.
TPD experiments after adsorption of oxygen were per-
formed with fresh catalysts in order to determine the influ-
ence of the Na addition on the Pt–O bond strength. About
Carbon and nitrogen mass balances were found to be
within 2%. Blank experiments have shown that the fritted
quartz was catalytically inert up to 600 C. We have also
◦
1
00 mg of each sample was placed on a quartz wool bed in a
U-shaped quartz reactor, positioned vertically in a program-
mable tube furnace. The catalyst was pretreated at 500 C
investigated the catalytic performances of the bare support
γ -Al2O3. Catalytic performances were studied as follows:
◦
for 1 h under He in order to clean its surface. Then, it was
◦
◦
reduced under H2 at 300 C for 1 h. After purging under
– in situ reduction by H2 at 300 C for 1 h,
◦
◦
He, the oxygen adsorption occurred at 300 C for 1 h un-
– cooling from 300 to 200 C under a flow of helium,
– introduction of the reactive mixture at 200 C,
– heating from 200 to 500 C at 1 C min
◦
der pure oxygen atmosphere. Then, the sample was cooled
◦
◦
◦
−1
−1
at 25 C under the same atmosphere. After purging under
,
◦
He for 15 min, the TPD was performed under He flow
– plateau at 500 C for 10 h,
−
1
◦
−1
◦
◦
(
8
1.75 L h ) with a temperature ramp of 20 C min up to
– cooling from 500 to 200 C at 1 C min
.
◦
50 C. TPD was monitored by a VG Gasslab 300 quadru-
pole mass spectrometer. Signals at m/e = 16, 18, 28, 32,
The catalytic activities presented in this study correspond to
those recorded during the decrease of temperature. Carbon
monoxide was not detected, according to our lower detec-
tion limit, about 10 ppm. Then, we did not take into account
the reactions of partial oxidation of the hydrocarbons into
CO, those of reforming of the hydrocarbons with steam pro-
duced by the combustion, the water–gas shift reaction, and
the oxidation of CO into CO2.
+ + +
5, 36, and 44 amu corresponding to O , H2O , N2 or
3
+
+
+
+
+
CO , O2 , Cl , HCl , and CO2 ions, respectively, were
recorded. Blank experiments with the support γ -Al2O3 by
itself and doped with Na (5 wt%) were performed and we
did not observe any oxygen or chlorine desorption. After cal-
ibration, we estimated the amount of oxygen desorbed per
gram of catalyst.
We calculated the production of N2, N2O, and NO2,
which was defined as the percentage of NO converted to
these products. The selectivity to N2 represented the amount
of the production of N2 compared to that of N2O. In order to
compare the catalytic behaviors of the various catalysts, we
used these following parameters:
2
.3. Catalytic activity measurements
Catalytic testing was performed at atmospheric atmos-
phere in a conventional U-shaped fixed-bed quartz reac-
tor. The catalyst sample (about 200 mg) was placed onto
a fritted quartz and a K-type thermocouple was located
near the catalytic bed. The reaction gases were mixtures
of NO (Air Liquide, 7970 ± 160 ppm NO in He), O2 (Air
Liquide, 99.99% purity), C3H6 (Messer, 7780 ± 230 ppm
C3H6 in He), or C3H8 (Air Products, 8000 ± 80 ppm C3H8
in He) and helium (Air liquide, 99.999% purity) which
–
–
T50, which corresponds to 50% conversion of the hydro-
carbon,
TN , TN O, and TNO which are representative of the
2
2
2
temperature at optimal conversion of N2, N2O, and
NO2, respectively.

250
P. Vernoux et al. / Journal of Catalysis 219 (2003) 247–257
3
. Results
Table 2
Tm and T values obtained with the Pt(Na)/γ -Al O catalysts
50
2 3
◦
◦
( C)
50
3
.1. Catalyst characterization
Table 1 summarizes the data of the different catalysts.
Catalysts
Na loadings
T
T
( C)
T
m
50
(wt%)
(%)
(C H /NO/O )
(C H /NO/O )
3 6 2
3
8
2
Pt
0
0.12
1
5
658
666
679
689
338
268
The Na loading varies from 0 wt% (sample Pt) to 0.12 wt%
PtNa0.12), 1 wt% (PtNa1), and 5 wt% (PtNa5). The chlo-
PtNa0.12
PtNa1
PtNa5
327
454
529
248
238
229
(
rine analyses showed that the residual chlorine was almost
unaffected by the initial pretreatment successively under O2
and H2 since the theoretical initial chlorine is 0.8 wt%. The
chemical addition of Na decreases the Pt dispersion (Table 1)
by decreasing the number of exchange sites during the pro-
cedure of Pt impregnation.
The variations of signals at m/e = 32 and 36, represen-
tative of O2 and HCl, respectively, versus temperature are
plotted in Fig. 1. The catalyst Pt, which does not contain
Na, presents a symmetrical peak of oxygen desorption cen-
Tm, temperature for the top of the oxygen desorption peak. T , temperature
for 50% hydrocarbon (propane or propene) conversion.
50
to those observed with Pt and PtNa0.12. The total amount
of oxygen desorbed calculated from these TPD measure-
ments corresponds to high Pt dispersion, higher than 80% for
PtNa1 and close to 50% for PtNa5. These values are higher
than those measured by H2–O2 titration (Table 1). This in-
dicates that the first oxygen evolution peak, recorded at low
temperature, is not representative of oxygen adsorbed on Pt.
Our assumption is that this oxygen could arise from the de-
composition of a sodium peroxide (Na2O2). Indeed, the tem-
perature of decomposition of this compoundinto oxygen and
◦
tered at 658 C (Fig. 1). The amount of oxygen desorbed
corresponds to the AMF determined by H2–O2 titration (Ta-
ble 1). With the sample PtNa0.12, the oxygen desorption
peak, which is less important due to a lower Pt disper-
◦
◦
sion, presents a maximum at 666 C. With catalysts PtNa1
Na O, i.e., 460 C, corresponds to that of the oxygen desorp-
2
and PtNa5, we observed two oxygen desorption peaks. The
first one, starting around 400 C, was very broad. The sec-
ond peak took place at higher temperatures corresponding
tion peak. By using the support γ -Al O doped with 5 wt%
2
3
◦
Na, we did not observe any oxygen desorption peak during
TPD measurements. This could indicate that the presence of
Pt is necessary to produce Na2O2. The role of Pt would be to
adsorb oxygen which can spillover onto the Na-doped alu-
mina support and then react with Na.
Table 1
Properties of the 0.9 wt% Pt/γ -Al O catalysts
2
3
Table 2 summarizes the temperature, Tm, corresponding
to the maximum of the oxygen desorption peak, related to
the Pt–O chemical bond. The value of Tm increases when
Na loading increases in the catalyst. One can suggest that
this modification of Tm could be attributed to the varia-
tion of the platinum particle size which also increases with
the Na amount in the catalyst (Table 1). However, several
studies [24,25] have observed that the O2 desorption tem-
perature was higher on the small Pt particles relative to the
large Pt particles. In addition, Briot et al. [26] have seen,
by microcalorimetric measurements, a decrease in the heat
of oxygen chemisorption when the platinum particle size in-
creased. Consequently, the increase of Tm with the Na load-
ing cannot be related to the increase of Pt particle size but to
the promoting effect of Na. This clearly demonstrates that Pt
particles are in interaction with Na and that this interaction
strengthens the Pt–O chemical bond.
Catalysts Na loading Cl loading
AMF (%)
fresh catalysts
AMF (%) after the
pretreatment
a
(wt%)
(wt%)
Pt
0
0.12
1
0.72
0.75
0.7
60
36
25
20
20
23
9
PtNa0.12
PtNa1
PtNa5
5
0.7
6.5
a
◦
After the pretreatment at 500 C for 10 h under the reactive mixture.
TPD measurements also show an emission of chlorine, as
◦
HCl, which is detected from 685 C for the sample Pt. This
confirms that chlorine which is coming from the platinic pre-
cursor has not been entirely eliminated during the pretreat-
ment of the sample. Furthermore, a very high temperature
is necessary to remove it under He. The poisoning of the
catalysts by chlorine will be discussed during the next sec-
tion on the catalytic activity measurements. An interesting
point is that, when the Na loading is high, the temperature
of HCl evolution significantly increases. For PtNa5, we did
+
+
Fig. 1. Profiles corresponding to O ion (at 32 amu) and HCl (at 36 amu)
2
◦
detections during TPD of the catalysts, Pt, PtNa0.12, PtNa1, and PtNa5.
not detect HCl below 850 C. This retention of the residual

P. Vernoux et al. / Journal of Catalysis 219 (2003) 247–257
251
chlorine by the catalyst has already been observed by Yen-
tekakis et al. [30] by doping Pt/γ -Al2O3 catalysts with Ba
15.2 wt%). The authors proposed the formation of a BaCl2
(
phase which presents a good chemical stability. In our case,
the presence of NaCl as a distinct phase at high Na load-
ings is probable. Yentekakis et al. [5] have already observed
the presence of this compound with a Pt/γ -Al2O3 catalyst
doped with 4.18 wt% Na.
3
.2. Catalytic activity measurements under C3H8/NO/O2
atmosphere
◦
Despite the oxygen pretreatment at 500 C and the in
situ reduction, the performances of the fresh catalysts were
not stable and improved with time on stream. We observed
a hysteresis loop in a subsequent temperature increase–
decrease cycle. This activation under the reactive mixture
has already been reported in the literature. It has been linked
to the elimination of chlorine coming from the platinic pre-
cursor [27–30]. Several studies [29,31] have demonstrated
that the residual chlorine can be slowly removed under a
reactive mixture. In addition, previous works have shown
that the presence of water vapor accelerates the loss of chlo-
rine from the catalyst surface [32,33]. Consequently after a
pretreatment under a reactive mixture, the catalytic activity
becomes similar to that of platinum-based catalysts prepared
from a chlorine-free precursor. However, the presence of
chlorine is not the only explanation for the activation of the
catalytic performance under a reactive mixture because this
phenomenon has already been observed with chlorine-free
Pt-based catalysts [34,35]. Another possible origin is the Pt
sintering under a reactive mixture. Denton et al. have shown
Fig. 2. Conversions of C H into CO as a function of temperature for
3
8
2
catalysts Pt (×), PtNa0.12 (!), PtNa1 (2), PtNa5 (F), and Al O (+).
2
3
Reactive mixture: C H /NO/O .
3
8
2
[
1] that the greater the decrease in dispersion, the greater the
activation. This has been observed when platinum is sup-
ported on silica, but with alumina the effect is less significant
[
36,37].
In order to eliminate the residual chlorine and to stabi-
lize the Pt dispersion, all the samples were pretreated under
◦
the reactive mixture at 500 C for 10 h. Then, the catalytic
performances were stable and we did not observe hysteresis
any more between the increase and the decrease of temper-
ature. We have observed a decrease of the Pt dispersion,
◦
after the pretreatment at 500 C for 10 h under the reactive
Fig. 3. Conversions of NO into NO as a function of temperature for cata-
2
lysts Pt (×), PtNa0.12 (!), PtNa1 (2), and PtNa5 (F).
mixture (Table 1). The values of the Pt dispersion were the
same ones as the catalyst pretreated under C3H8/NO/O2 or
C3H6/NO/O2.
For the sample Pt, significant C3H8 conversion starts at
about 250 C and becomes complete from 500 C (Fig. 2).
The oxidation of NO into NO2 occurs at low temperatures
scribed in the literature [38,39]. The addition of moderate
amounts of Na, as for catalyst PtNa0.12, slightly promotes
the catalytic activity (Table 2). The promotional effect is par-
ticularly marked for the NO oxidation into NO2 since the
◦
◦
and the associated conversion passes through a maximum at
◦
◦
2
89 C, with a NO conversion into NO2 of 47% (Fig. 3). Ni-
conversion maximum occurs at a lower temperature, 263 C
◦
trogen dioxide is the main product since the conversions of
NO into N2 and N2O remain low (Fig. 4). The maximum
NO conversion into N2O was only 5.3% and occurred at
instead of 289 C with Pt, and is slightly higher. However,
when the Na loading increases, the catalytic activity dras-
tically decreases. The T50 value for propane conversion in-
◦
◦
◦
◦
◦
3
60 C while that of NO into N2 was 2.6% at 395 C. These
creases from 338 C for Pt to 454 C for PtNa1 and 529 C
for PtNa5. The catalytic performance of the sample PtNa5
catalytic behaviors are in good agreement with those de-

252
P. Vernoux et al. / Journal of Catalysis 219 (2003) 247–257
Fig. 4. Conversions of NO into N O and N as a function of temperature for
2
2
catalysts Pt (P,Q), PtNa0.12 (!,"), and PtNa5 (E,F). Reactive mixture:
C H /NO/O .
3
8
2
Table 3
Fig. 5. Conversions of C H into CO as a function of temperature for cata-
3
6
2
TN , T
, and T
values and associated maximum NO conversion
NO2
2
N2O
lysts Pt (×), PtNa0.12 (!), PtNa1 (2), and PtNa5 (F). Reactive mixture:
into N , N O, and NO obtained with the Pt(Na)/γ -Al O catalysts un-
2
2
2
2
3
C H /NO/O .
3
6
2
der C H /NO/O
2
3
8
Catalysts
TN /conversion
T
/conversion
T
/conversion
2
N2O
NO2
(%)
(%)
(%)
Pt
395/2.6
360/5.3
289/47
PtNa0.12
PtNa1
PtNa5
391/3.5
364/5.5
263/51.5
362/29.5
409/25
–
–
–
–
TN , temperature for the maximum NO conversion into N . T , temper-
N2O
2
2
ature for the maximum NO conversion into N O. T
, temperature for the
NO
2
2
maximum NO conversion into NO .
2
for the C3H8 combustion is comparable to that of the support
itself (Fig. 2) whereas the Pt dispersion is 20% (Table 1). In
the same way, the NO oxidation is considerably inhibited by
the presence of Na (Table 3). The maximum NO conversion
into NO2 is divided by a factor 2 with an addition of 5 wt%
◦
of Na (PtNa5) and it is reached at a temperature 120 C
higher than that measured with Pt. Let us note that, whatever
the catalyst, the maximum NO conversion into NO2 always
takes place before that the propane conversion becomes sig-
nificant. With PtNa1 and PtNa5 samples, the conversion of
◦
NO into N2 and N2O is negligible below 500 C. Then, the
Fig. 6. Conversions of NO into N as a function of temperature for cata-
2
values of the selectivity to N2 are not significant.
lysts Pt (×), PtNa0.12 (!), PtNa1 (2), and PtNa5 (F). Reactive mixture:
C H /NO/O .
3
6
2
3
.3. Catalytic activity measurements under a
C3H6/NO/O2 atmosphere
ature at which the propene conversion becomes total. The
operating temperature window, defined as the full width at
half-maximum (FWHM) of the peak corresponding to NO
The catalytic performances of the four samples are pre-
sented on Figs. 5–9. The catalytic behavior of Pt is similar
to that reported in the literature for Pt-based catalysts [1,3].
The propene oxidation (Fig. 5) is strongly temperature de-
pendent and is complete from 285 C. The production of N2
passes through a maximum of 19.6% at 270 C (Fig. 6), tem-
◦
conversion into N2, can be estimated to 34 C in the range
◦
◦
260–294 C. The production of NO2 starts at 285 C when
the propene conversion is complete (Fig. 8). With propene
as a reducing agent, the main reaction is the oxidation of
propene. At temperatures sufficiently low to avoid the com-
plete oxidation of C3H6 into CO2, the NO reduction by
propene can take place. At low C3H6 conversions, the major
◦
◦
perature that corresponds to the 50% conversion of propene,
while that of N2O reaches 33% at 287 C (Fig. 7), temper-
◦

P. Vernoux et al. / Journal of Catalysis 219 (2003) 247–257
253
Fig. 9. Variation of the N selectivity as a function of temperature for cata-
2
Fig. 7. Conversions of NO into N O as a function of temperature for cata-
2
lysts Pt (×), PtNa0.12 (!), PtNa1 (2), and PtNa5 (F). Reactive mixture:
lysts Pt (×), PtNa0.12 (!), PtNa1 (2), and PtNa5 (F). Reactive mixture:
C H /NO/O .
3
6
2
C H /NO/O .
3
6
2
PtNa0.12, the maximum NO conversion into N2 does not
vary, whereas it slightly decreases for higher Na loadings.
We estimated the turnover frequency (TOF) of NO con-
sumption at the temperature corresponding to the maximal
overall NO conversion (Table 4). The TOF values were cal-
culated by using the dispersion of catalysts which have been
◦
pretreated under reactive mixtures at 500 C for 10 h. The
addition of Na increases TOF of NO which passes through
a maximum for 1 wt% Na (Table 4). For high Na load-
ings, the FWHM significantly increases. Let us note that
the maximum NO conversions into N2 observed with cata-
lysts PtNa1 and PtNa5 take place at temperatures that do not
correspond to the T50 values (Tables 2 and 4). With the sam-
ple PtNa0.12, the N2O and NO2 productions are promoted
(
Figs. 7 and 8). On the contrary, the presence of a higher Na
loading inhibits the conversion of NO into NO2. Optimum
performance is observed for catalyst PtNa5, characterized
◦
◦
by T50 = 229 C (39 C lower than for the unpromoted cata-
◦
◦
◦
lyst Pt), TN = 240 C (30 C lower than Pt), TN O = 244 C
2
2
◦
(
43 C lower than Pt), and a very low conversion to NO2 in
Fig. 8. Conversions of NO into NO as a function of temperature for cata-
2
◦
lysts Pt (×), PtNa0.12 (!), PtNa1 (2), and PtNa5 (F). Reactive mixture:
the temperature domain 200–400 C. Furthermore, the sam-
ple PtNa5 presents two interesting properties:
C H /NO/O .
3
6
2
product is N2 (Fig. 9). As soon as the temperature exceeds
T50, NO is mainly reduced into N2O.
– its operating temperature window is very wide; i.e., the
N2 production continues even after the complete oxida-
tion of propene (Fig. 6). The NO conversion into N2 is
The C3H6 conversion is improved by the addition of Na
◦
(
Fig. 5). The value of T50 linearly decreases with the Na
significant until 400 C while that of N2O is higher than
◦
loading (Table 2). Despite a lower Pt dispersion, the catalyst
10% until 325 C;
◦
PtNa5 exhibits a complete propene conversion from 249 C
– the selectivity to N2 is higher than that of the unpro-
moted catalyst (Fig. 9). In the temperature range 250–
◦
instead of 285 C with Pt. The NO conversion into N2 is
◦
shifted toward lower temperatures with the presence of Na
325 C, the C3H6 conversion is complete. The NOx
(
Fig. 6 and Table 4). The same trend is observed for the NO
conversion gradually decreases from 30 to 20% with a
selectivity to N2 that increases from 41 to 51%.
conversion into N2O (Fig. 7 and Table 4). With the sample

254
P. Vernoux et al. / Journal of Catalysis 219 (2003) 247–257
Table 4
The Na effects are spectacular and more marked for
TN , T
values and associated maximum NO conversion into N , and
2
2
N2O
propane than for propene. For instance, when adding 5 wt%
N O obtained with the Pt(Na)/γ -Al O catalysts under C H /NO/O
2
◦
2
3
3
6
2
Na, the value of T50 increases by 191 C with propane and
−2
−1
◦
Catalysts TN /conversion
T
/conversion
N2O
TOF NO (10
s
)
decreases by 39 C with propene. These opposite behaviors
2
(%)
(%)
at temperature
may not be attributed to a modification of Pt dispersion in-
duced by the presence of Na since it is the same for both hy-
drocarbons. Consequently, we may attribute the considerable
modification of the Pt catalytic properties to a chemical pro-
motion of Pt by Na. The Na effects on the surface chemistry
of Pt have already been observed, under TWC conditions, by
Konsolakis et al. [4] for the NO reduction by propene. Their
results are in good agreement with ours since they show
that the presence of Na promotes the activity (C3H6 and NO
conversions) and the selectivity to N2. However, under lean-
burn conditions, Burch and Watling [11] have shown that the
addition of K (2.42 wt%), Cs (7.25 wt%) and Ba (7.93 wt%)
to Pt/Al2O3 catalyst (1 wt% of Pt) does not improve the
catalytic performance of Pt. By adding 2.42 wt% K, they
observed an increase of the FWHM associated with a tem-
perature of the maximum NOx conversion slightly shifted to
of optimal NO conversion
Pt
270/19.6
250/19.6
243/15.2
240/14
287/33
262/40
250/29.5
244/19.5
14.7
15.7
34.7
30.8
PtNa0.12
PtNa1
PtNa5
TN , temperature for the maximum NO conversion into N . T , temper-
N2O
2
2
ature for the maximum NO conversion into N O.
2
The minimum value of the selectivity to N2 decreases for
the sample PtNa0.12 which contains the lowest amount of
Na. On the contrary, it increases with higher concentrations
of Na from 1 wt%. These results clearly demonstrate that
the SCR of NO by propene on Pt/γ -Al2O3 catalysts can be
promoted by the addition of Na, even under lean-burn con-
ditions. The operating temperature window of platinum can
be drastically expanded and the selectivity to N2 increases.
◦
lower temperatures (from 275 C for the unpromoted cata-
◦
lyst to 265 C), and a selectivity to N2 which was not mod-
ified. This is what we observed with 1 wt% Na. However,
the authors have seen that the maximum NOx conversion
decreased form 53 to 35%. This discrepancy may be due to
differences in the Pt dispersion of the samples or because the
4
. Discussion
4
.1. General
−1
space velocity used in this work (20,000 h ) was much less
−
1
The aim of this work was to demonstrate if Na doping
than one used in Ref. [11] (87,000 h ).
may influence the performance of Pt/Al2O3 in NOx SCR
under lean-burn conditions. Such an effect was detected and
it is markedly different according to the nature of the hydro-
carbon. By using propane as a reducing agent, the principal
effects of Na are
4.2. Origins of the chemical promotion of Na
The chemical promotion of Na to Pt suggests that Na is in
interaction with Pt particles. Postreaction spectroscopic data,
reported in the literature [13,40], indicate that Na is present
at the catalyst surface in the form of chemical compounds
such as Na CO or NaNO . One can think that NaCl might
–
–
a slight enhancement of Pt activity, essentially for the
oxidation of NO to NO2 for low Na amounts,
an important decrease of the catalytic performances, at
higher Na loadings (1–5 wt%).
2
3
3
also be formed with the residual chlorine originating from
the platinic precursor. This could explain why we observed,
by TPD, a retention of chlorine when the Na loading is high.
Our TPD measurements have demonstrated that Pt particles
are in interaction with Na since the temperature of the oxy-
gen desorption peak, which is related to the Pt–O chemical
bond, increases when Na loading increases. These observa-
tions prove that the presence of Na, in interaction with Pt,
strengthens the Pt–O chemical bond. Therefore, we may ra-
tionalize the promotional phenomena observed in terms of
modification of the adsorption strengths of the various reac-
tants on the metal surface. The ability of Na to enhance the
NO adsorption on Pt has been demonstrated by Harkness et
al. [41] by using NO TPD measurements. Furthermore, sev-
eral studies have shown that the addition of Na promotes the
NO dissociation [4,13]. According to Lang et al. [42], the
Na effect should be due to an electrostatic field induced by
the alkali cation. Taking into account that Na is present on
the catalyst in the form of a chemical compound, this as-
When using propene,
–
–
–
–
the presence of Na promotes the propene oxidation. The
value of T50 decreases with the Na loading,
the optimal NO conversions to N2 and N2O are shifted
toward lower temperatures according to the Na loading,
the TOF of NO increases with Na loading and passes
through a maximum for 1 wt% Na,
the N2O and NO2 productions are promoted for low
amounts of Na (0.12 wt%) and inhibited for higher load-
ings,
–
–
the selectivity to N2 slightly decreases for low Na con-
centration (0.12 wt%),
for the sample PtNa5, which contains the highest con-
centration of Na (5 wt%), we observed an enhancement
of the selectivity to N2 and also a widening of the oper-
ating temperature window.
−
sumption supposes that the counteranion, such as NO3 or

P. Vernoux et al. / Journal of Catalysis 219 (2003) 247–257
255
CO32 , should exert a negligible opposite effect to cation.
Konsolakis et al. [4] suggest that the electrostatic field due to
the anion is low mainly due to its larger size compared to the
cation. Recently, Yentekakis et al. [30] have clearly shown
that the positive effect of Ba added to Pt/alumina catalysts,
on the NO reduction by propene, overwhelms the negative
effect of chlorine coming from the platinic precursor. We
also observed in this present study, under C3H6/NO/O2 at-
mosphere, a promotional effect induced by the addition of
Na despite the possible presence of NaCl on the catalyst.
Konsolakis et al. [4] have exposed a similar beneficial effect
of Na under TWC conditions by using H2PtCl6 as a platinic
precursor. Therefore one can conclude that, with the chemi-
−
alumina catalysts for the NO reduction by propene under
TWC conditions. Contrary to the case of Pt, propene is not
strongly adsorbed on Rh. Then, the authors suggest that an
excessive promotion of oxygen adsorption due to the pres-
ence of Na inhibits the propene adsorption and oxidation. We
have shown that the NO conversion to NO2 also decreases
with the Na loadings. This could indicate that the strength-
ening of the Pt–O bond is much more pronounced than that
of the Pt–NO bond because O is a much stronger electron
acceptor than NO. Then, the NO coverage decreases lead-
ing to the drop of the NO conversion. At low loadings of Na
(0.12 wt%), we observed a slight promotion of NO oxidation
to NO2. Let us note that we observed the same promotion by
using propene with 0.12 wt% Na. One can suppose that the
origin of this moderate activation could arise from a slightly
higher Pt dispersion (Table 1). Indeed, we observed that the
Pt dispersion of this sample decreases less than that of the
catalyst which does not contain Na, during the pretreatment.
Konsolakis et al. [4] have found similar results with their
samples containing low loadings of Na.
+
cal promotion by Na , the nature of the counteranions seems
not to be important.
Again referring to Lang et al. [42], the electrostatic field
+
∗
induced by Na should shift the π -orbital energy of a NO
molecule adsorbed in the vicinity of an alkali ion below the
Pt Fermi level. Then, valence electrons from the metal can
∗
populate the NO π -orbital energy, resulting in weakening
of the N–O bond and a strengthening of the Pt–N bond. In
a general way, the electrostatic field induced by the positive
charge of sodium ions promotes the adsorption of electron
acceptors such as NO and O2 and inhibits the adsorption of
electron donors such as propene. Therefore, we can consider
these modifications of the strength of the chemical bonds
between Pt and the different adsorbates in order to elucidate
the Na effect on the Pt catalytic performances. The oppo-
site behaviors between propane and propene, depending on
their strength of adsorption on Pt, induce different coverages
of adsorbates on a Pt surface. This could explain that the
addition of Na leads to opposite effects according to the hy-
drocarbon considered.
Under a C3H6/NO/O2 atmosphere, the Pt surface is pre-
dominantly covered by carbonaceous species while the oxy-
gen coverage is negligible. Therefore, the weakening of the
Pt–C3H6 bond induced by Na addition should lead to an
increase of O and NO coverages on Pt. This considerably
promotes the catalytic behavior of Pt. The beneficial effect
is much more pronounced for the C3H6 combustion. This
could be due to the fact that the oxygen coverage is much
more increased than that of NO, because as already noted
above, O is a much stronger electron acceptor than NO. For
the highest Na concentration (5 wt%), the NO dissociation is
favored by the presence of Na since we observed an increase
of the selectivity to N2. This confirms that the Na loading
must be sufficiently important in order to promote the Pt cat-
alytic behavior. This could explain why Tanaka et al. [10] ob-
served that small amounts of Na (0.1 wt%) do not change the
catalytic properties of Pt-supported silica catalysts. An in-
teresting point is the widening of the operating temperature
window observed with 5 wt% Na. Indeed, in the temperature
Under a C3H6/NO/O2 atmosphere, Burch and Watling
[
43] suggested that the Pt surface is predominately cov-
ered by carbonaceous species while the oxygen coverage
is negligible. In contrast, under a C3H8/NO/O2 reactive
mixture, the oxygen coverage on a Pt surface would be
high [39]. Burch and co-workers [39,44] and Inaba et al. [45]
have proposed that the mechanism of the NOx reduction on
Pt/Al2O3 catalysts is different according to the nature of the
hydrocarbon. By using propene, NOx reduction only occurs
on the Pt surface while, with propane, the reaction appears
to occur via the formation of a surface nitrate species on the
alumina support that activates the reductant.
When using propane as a reducing agent, our catalytic re-
sults confirm that the NO oxidation reaction is predominant
since the propane conversion is low before the maximum
conversion of NO into NO2. The addition of Na strengthens
the Pt–O bond, as confirmed by TPD measurements. One
can think that this increase of the oxygen coverage saturates
the Pt surface in oxygen and then inhibits the adsorption of
C3H8 and NO. This leads to a significant drop of the C3H8
and NO conversions, as experimentally observed. Macleod
et al. [9] proposed the same interpretation concerning the
detrimental effect induced by Na addition with Rh-supported
◦
range 250–325 C when the propene conversion is total, the
conversions of NO into N2O and essentially into N2 remain
significant whereas that into NO2 becomes negligible. This
could indicate that a new mechanism occurs when the quan-
tity of Na becomes important on the catalyst surface. This
is also confirmed by the fact that the maximum NO conver-
sions into N2 observed with catalysts PtNa1 and PtNa5 take
place at temperatures that do not correspond to the T50 val-
ues. One can suggest that NO2 is trapped by the excess of
Na present at the surface of alumina, in the form of sodium
nitrates, and then is reduced by propene into N2 and N2O.
4.3. Correlation with electrochemical promotion using
+
Na -conducting solid electrolytes
One may correlate the assumption on the role of the
+
Na electrostatic field with studies undertaken in electro-

256
P. Vernoux et al. / Journal of Catalysis 219 (2003) 247–257
+
chemical promotion with Na -conducting solid electrolytes
agent, the addition of Na is detrimental, leading to a drastic
decrease of the C H and NO conversions. In contrast, with
propene, Na promotion improves the activity of Pt. For high
concentration of Na (5 wt%), we observed an enhancement
ꢀꢀ
such as NASICON or β -Al2O3. We have recently observed
14] that the application of negative overpotentials through
a Pt/NASICON interface significantly improves the Pt cat-
alytic performances in the same oxygen-rich C3H6/NO/O2
atmosphere used in this present study. Applying negative po-
tentials to a Pt film on NASICON leads to a spillover of
Na ions from the electrolyte onto the Pt surface. A 295 C,
we have seen that the application of −100 mV leads to the
enhancement of NO and C3H6 conversions while the selec-
tivity to N2 increased from 41 to 61%. This promoting effect
is quite similar to that observed in this study with 5 wt%
Na. Therefore, the two ways of Na providing to Pt (elec-
trochemically or chemically) lead to similar improvements
of the Pt catalytic performances. Then, we may be certain
that, in the present case, the promotion is mainly due to the
3
8
[
of the selectivity to N and also a widening of the operating
2
temperature window. The Na effects on activity and selectiv-
ity of Pt have been correlated to modifications of the strength
of the chemical bonds between Pt and the different adsor-
bates. The opposite effects observed by using propane or
propene have been attributed to their different mechanisms
of adsorption on Pt.
Similarities exist between the catalytic performances of
an electrochemically promoted Pt film deposited on a NASI-
CON electrolyte [14] and a powdered Pt-supported alumina
promoted by chemical additions of Na. Whatever the mode
of the Na addition (electrochemical or chemical), the inter-
pretations of the Na effects operate in the same direction,
namely a strengthening of the Pt–electron-acceptor (O and
NO) bonds and a weakening of the Pt–electron-donor (C3H8
and C3H6) bonds.
+
◦
+
effect of Na on the platinum particles, rather than on the
support. The comparison between the electrochemical and
chemical Na addition suggests that the interactions between
Pt and electrochemically supplied Na are many more than
those with chemically added Na. Indeed, in this study the
molar ratio of Pt to Na is 1:50 despite 100:1 in Ref. [14]. We
tried to explain the observed electrochemically promotional
effects [14] on the basis of the variation of the work func-
tion (WF) of Pt according to the rules proposed by Vayenas
et al. [46,47]. These authors established that a change in the
overpotential ꢀVWR between the catalyst-electrode such as
Pt and a solid electrolyte modifies the work function eΦ of
the gas-exposed surface of the porous catalyst-electrode film
according to
References
[1] P. Denton, A. Giroir-Fendler, H. Praliaud, M. Primet, J. Catal. 189
(2000) 410.
[2] E. Seker, E. Gulari, J. Catal. 194 (2000) 4.
[
[
3] R. Burch, D. Ottery, Appl. Catal. B 13 (1997) 105.
4] M. Konsolakis, N. Macleod, J. Isaac, I.V. Yentekakis, R.M. Lambert,
J. Catal. 193 (2000) 330.
[5] I.V. Yentekakis, M. Konsolakis, R.M. Lambert, N. Macleod, L. Nal-
bantian, Appl. Catal. B 22 (1999) 123.
ꢀ
(eΦ) = eꢀVWR.
(1)
[
[
[
[
6] M. Konsolakis, I.V. Yentekakis, Appl. Catal. B 29 (2001) 103.
7] N. Macleod, J. Isaac, R.M. Lambert, J. Catal. 198 (2001) 128.
8] N. Macleod, J. Isaac, R.M. Lambert, J. Catal. 193 (2000) 115.
9] N. Macleod, J. Isaac, R.M. Lambert, Appl. Catal. B 33 (2001) 335.
10] T. Tanaka, K. Yokota, N. Isomura, H. Doi, M. Sugiura, Appl. Catal.
B 16 (1998) 199.
The variation of the Pt WF alters the strength of the chem-
ical bonds between the metal and the adsorbates. The conse-
quence of applying a negative overpotential, then supplying
Na to Pt, is to weaken the Pt chemical bonds with electron-
[
+
donor adsorbates, such as C3H6 and to strengthen those with
electron-acceptor ones, such as O2 and NO. According to
the classification of C. Vayenas et al. [48], propane is also
considered as an electron-donor, then the strength of the Pt–
C3H8 bond is supposed to decrease in the presence of Na.
Therefore, whatever the mode of Na addition (electrochemi-
cal or chemical), the tentative interpretation of the Na effects
operates in the same direction, namely a strengthening of
the Pt–O and Pt–NO bonds and a weakening of the Pt–HC
[
11] R. Burch, T.C. Watling, Appl. Catal. B 11 (1997) 207.
[12] R. Burch, M.D. Coleman, J. Catal. 208 (2002) 435.
[13] I.V. Yentekakis, A. Palermo, N.C. Filkin, M.S. Tikhov, R.M. Lambert,
J. Phys. Chem. B 101 (1997) 3759.
[14] P. Vernoux, F. Gaillard, C. Lopez, E. Siebert, J. Catal. 217 (2003) 203.
[15] R. Burch, D.J. Crittle, M.J. Hayes, Catal. Today 47 (1999) 229.
[16] R. Burch, T.C. Watling, Catal. Today 43 (1997) 19.
[17] P. Vernoux, F. Gaillard, L. Bultel, E. Siebert, M. Primet, J. Catal. 208
(2002) 412.
[18] A.B. Anderson, D.B. Lang, Y.J. Kim, J. Am. Chem. Soc. 106 (1984)
6597.
(
HC = C3H6 or C3H8) bonds.
[
[
[
19] H. Öfner, F. Zaera, J. Phys. Chem. 101 (1997) 396.
20] F. Zaera, G.A. Somorjai, J. Am. Chem. Soc. 106 (1984) 2288.
21] J.F. Weaver, M.A. Krzyzowski, R.J. Madix, J. Chem. Phys. 112 (1)
5
. Conclusions
(
2000) 396.
22] J.T. Carter, J.A. Cusumano, J.H. Sinfelt, J. Phys. Chem. 70 (1966)
257.
23] G.A. Somorjai, in: Chemistry in two Dimensions, Cornell Univ. Press,
Ithaca, NY, 1981, p. 479.
24] V. Labalme, N. Guilhaume, E. Garbowski, M. Primet, React. Kinet.
Catal. Lett. 64 (1998) 207.
25] E.S. Putna, J.M. Vohs, R.J. Gorte, Surf. Sci. 391 (1997) L1178.
[
[
[
[
2
The activity and nitrogen selectivity of platinum-sup-
ported alumina catalysts for the NO reduction by hydrocar-
bons under lean-burn conditions can be significantly mod-
ified by the addition of Na. The effect of the latter on the
catalytic properties is markedly different according to the
nature of the hydrocarbon. By using propane as a reducing
[26] P. Briot, A. Auroux, D. Jones, M. Primet, Appl. Catal. 59 (1990) 141.

P. Vernoux et al. / Journal of Catalysis 219 (2003) 247–257
257
[
[
27] F.J. Gracia, J.T. Miller, A.J. Kropf, E.E. Wolf, J. Catal. 209 (2002) 341.
28] P. Denton, A. Giroir-Fendler, H. Praliaud, M. Primet, Top. Catal. 16/17
[39] R. Burch, T.C. Watling, Catal. Lett. 43 (1997) 19.
[40] N.C. Filkin, M.S. Tikhov, A. Palermo, R.M. Lambert, J. Phys. Chem.
B 103 (1999) 2680.
(2001) 377.
[
[
29] D. Roth, P. Gélin, M. Primet, E. Tena, Appl. Catal. A 203 (2000) 37.
30] I.V. Yentekakis, R.M. Lambert, M. Konsolakis, N. Kallithrakas-Kon-
tos, Catal. Lett. 81 (3–4) (2002) 181.
[41] I.R. Harkness, R.M. Lambert, J. Chem. Soc., Faraday Trans. 93 (7)
(1997) 1425.
[42] N.D. Lang, S. Holloway, J.K. Norskov, Surf. Sci. 150 (1985) 24.
[43] R. Burch, T.C. Watling, in: N. Kruse, A. Frennet, J.-M. Bastin (Eds.),
Proceedings of Fourth International Congress on Catalysis and Auto-
motive Pollution Control, Vol. 116, Brussels, 1998, p. 199.
[44] R. Burch, J.A. Sullivan, T.C. Watling, Catal. Today 42 (1998) 13.
[45] M. Inaba, Y. Kintaichi, H. Hamada, Catal. Lett. 36 (1996) 223.
[46] C.G. Vayenas, S. Bebelis, I.V. Yentekakis, H.G. Lintz, Catal. Today 11
(1992) 303.
[
[
31] E. Marceau, H. Lauron-Permot, M. Che, J. Catal. 197 (2001) 394.
32] I. Mochida, A. Uchino, H. Fujitsu, K. Takeshita, J. Catal. 51 (1978)
72.
[33] A.A. Castro, O.A. Scelza, G.T. Baronetti, M.A. Fritzler, J.M. Parera,
Appl. Catal. 6 (1983) 347.
[34] F. Jayat, C. Lembacher, U. Schubert, J.A. Martens, Appl. Catal. B 21
(1999) 221.
[35] J.H. Lee, H.H. Kung, Catal. Lett. 51 (1998) 1.
[36] R. Burch, P.J. Millington, Catal. Today 26 (1995) 185.
[37] R. Burch, P.J. Millington, Catal. Today 29 (1996) 37.
[38] R. Burch, T.C. Watling, J. Catal. 169 (1997) 45.
[47] C.G. Vayenas, M.M. Jaksic, S. Bebelis, S.G. Neophitides, in: J.O.M.
Bockris, B.E. Conway, R.E. White (Eds.), in: Modern Aspects of Elec-
trochemistry, Vol. 29, Plenum, New York, 1996, p. 57.
[48] C.G. Vayenas, S. Brosda, C. Pliangos, J. Catal. 203 (2001) 329.