Pd/NaY-zeolite and Pd-W/NaY-zeolite catalysts: Preparation, characterization and NO decomposition activity

Article abstract of DOI:10.1016/j.molcata.2003.08.005

Pd/NaY-zeolite and Pd-W/NaY-zeolite catalysts were prepared from the compounds Pd(NO₃)₂ and [W(CO)₆]. Tungsten was added by photochemical activation of [W(CO)₆] to obtain a Pd-W interaction. The prepared catalysts were active for NO decomposition at 573 K leading mainly to N₂ and N₂O. NO conversion and the selectivity to N₂O were studied as a function of time and the reduction temperature. The Pd/NaY-zeolite and the Pd-W/NaY-zeolite samples showed an initial period of high and constant activity followed by deactivation. Tungsten prevented the sintering of small Pd particles. Bimetallic sample deactivated faster, indicating that the Pd-W interaction decreased the fraction of exposed Pd atoms. The onset of N₂O formation was in accord with the initiation of the deactivation.

Full text of DOI:10.1016/j.molcata.2003.08.005

Journal of Molecular Catalysis A: Chemical 209 (2004) 107–115
Pd/NaY-zeolite and Pd-W/NaY-zeolite catalysts: preparation,
characterization and NO decomposition activity
Sibele B.C. Perghera, Rogério M. Dallagoa,d, Renato Cataluña Vesesb,
Carlos E. Gigolac, Ione M. Baibichb,∗
a
Departamento de Qu´ımica, URI, Campus Erechim, Av. Sete de Setembro, 1621, CEP 99700-000, Erechim, RS, Brazil
b
Instituto de Qu´ımica, UFRGS, Av. Bento Gonçalves, 9500, CEP 91501-970, Porto Alegre, RS, Brazil
c
PLAPIQUI, Km 7 Camino “La Carrindanga” C.C. 717-8000, Bahia Blanca, Argentina
d
Intituto de Qu´ımica, UCS, Francisco Getúlio Vargas, 1130, Caxias do Sul, RS, CEP 95070-560, Brazil
Received 28 April 2003; received in revised form 7 August 2003; accepted 11 August 2003
Abstract
In this work, Pd/NaY-zeolite and Pd-W/NaY-zeolite catalysts were prepared from the compounds Pd(NO3)2 and [W(CO)6]. Palladium
catalysts supported on NaY-zeolite were prepared by ion exchange and impregnation methods. Subsequently, tungsten was added by pho-
tochemical activation of [W(CO)6] in order to obtain a Pd-W interaction. The catalysis materials were characterized by diffuse reflectance
(DRIFTS)–FTIR, chemical analysis (AAS), temperature programmed reduction (TPR), X-ray diffraction (XRD), H2 chemisorption and sorp-
tion measurements. The prepared catalysts were found to be active for NO decomposition at 573 K leading mainly to N2 and N2O. The
conversion of NO and the selectivity to N2O were studied as a function of time and the reduction temperature. The Pd/NaY-zeolite and the
Pd-W/NaY-zeolite samples exhibited an initial period of high and constant activity followed by deactivation. For samples reduced at 573 K the
deactivation process becomes more pronounced in the presence of tungsten. On the other hand when the reduction pretreatment was carried
out at 973 K the presence of tungsten allowed the catalysts to have a higher conversion for a longer period of time. Tungsten prevents the
sintering of small palladium particles. Another beneficial effect of tungsten was a significant suppression of N2O formation. The selectivity
for N2O at similar NO conversion level was lower on Pd-W/NaY-zeolite.
© 2003 Elsevier B.V. All rights reserved.
Keywords: NO decomposition; NaY-zeolite; Palladium; Tungstencarbonyl; Catalysts
1. Introduction
Recently, we have demonstrated that an alternative pro-
cedure for depositing tungsten or molybdenum on Al2O3
or Pd/Al2O3 is based on the photochemical reaction of
[M(CO)6] compounds (M = W, Mo) [10–14]. In this way
we have prepared and characterized catalysts where tung-
sten or molybdenum were found in close association with
palladium. Moreover, these catalysts showed high activity
for NO decomposition [12,15].
The purpose of this work was to extend our previous
studies performed on ␥-Al2O3 to zeolites, as it is known
that metalcarbonyls interact more strongly with zeolites that
with alumina [16–19]. Another motivation for this work
are the reports that show that very active and selective cat-
alysts for the reduction of nitric oxide with methane are
obtained by supporting palladium on zeolites [20,21], the
role of the support being the stabilization of Pd2+ species.
Hence, Pd/NaY-zeolite catalysts were prepared by dry im-
pregnation and ionic exchange methods and the related
In an attempt to increase the catalytic activity of palla-
dium catalysts for the removal of hydrocarbons and NOx
several promoters such as WO3, Co3O4, La2O3 and MO3
have been used [1–9]. The interaction between the noble
metal and the oxide seems to be an essential condition for
changing the catalytic behavior. However, using inorganic
precursors a high metal oxide loading is in fact needed to
satisfy this condition. One way of preventing this problem
is the use of organometallic compounds; in this way prepa-
rations with low loadings and more controlled interactions
could be obtained.
∗
Corresponding author. Tel.: +55-51-333-30414;
fax: +55-51-331-66304.
E-mail address: ione@iq.ufrgs.br (I.M. Baibich).
1381-1169/$ – see front matter © 2003 Elsevier B.V. All rights reserved.
doi:10.1016/j.molcata.2003.08.005

108
S.B.C. Pergher et al. / Journal of Molecular Catalysis A: Chemical 209 (2004) 107–115
Pd-W/NaY-zeolite catalysts were obtained by photochem-
ical activation of [W(CO)6]. A high amount of Pd was
incorporated on the support by the ionic exchange method,
in the hope of increasing the interaction with tungsten,
since it is known that metalcarbonyls react preferentially
with large palladium particles [13]. The monometallic and
bimetallic catalysts were characterized by hydrogen sorp-
tion, chemisorption and temperature programmed reduction
(TPR) measurements. In addition, the effect of tungsten on
the activity and selectivity of Pd/NaY-zeolite for NO de-
composition at 573 K was investigated for samples reduced
at 573 and 973 K.
[W(CO)6] (Aldrich) and hexane as solvent in the presence
of the Pd/NaY-zeolite catalyst. The quantities used were
calculated to give 2% by weight of W on the zeolite (1 g of
Pd-zeolite; 0.038 g of [W(CO)6]; 40 cm3 of hexane).
The photochemical reactions were performed at room
temperature under an Ar atmosphere using a Philips HPL-N
(125 W) UV lamp fitted into a Pyrex cold finger. In or-
der to observe the reaction progress, small liquid samples
were withdrawn for analysis by FTIR. After an irradiation
time of 6 h, the solid was filtered and washed with hexane.
Small amounts of solution were removed for solution FTIR
analysis.
Finally, the generated subcarbonyl species were decom-
posed by thermal treatment at 573 K under vacuum.
2. Experimental
2.3. Characterization
The catalysts were prepared using a commercial NaY-
zeolite as a support. The NaY-zeolite was obtained from
ITQ, Valencia, Spain (surface area 655 m2/g; SiO2/Al2O3 =
5.2 and Na2O/Al2O3 = 1.02 molar ratio).
The prepared materials were characterized by DRIFTS–
FTIR, chemical analysis (AAS), temperature programmed
reduction (TPR), X-ray diffraction (XRD) and H2 chemisorp-
tion and sorption experiments. Diffuse reflectance (DRIFTS)
spectra were recorded at ca. 293 K on a Bomem MB-102
Spectrometer, coadding 32 scans at a resolution of
4 cm−1. The study was restricted to the ␯(CO) region
(2200–1700 cm−1). The solid phases were analyzed as
powder in a DRIFTS accessory, equipped with sampling
cups. The latter were filled with the support powder and
transferred to the equipment as fast as possible, to prevent
decomposition. These spectra were collected as reflectance
units and transformed to Kubelka–Munk (KM) units. The
solutions were analyzed with a pair of matched NaCl solu-
tion cells (0.10 mm pathlengths).
The NaY-zeolite has a structure that shows two chan-
nel systems interconnected. One system was formed by the
union of ␣ supercages (i.d. = 12.4 Å) that can be reached
by channels 12 membered rings (12 MR) with a diameter of
7.8 Å. Another channel system is formed by the alternated
connection of sodalite cages (i.d. of 6.6 Å) and the ␣ su-
percage, that can be reached by channels 6 MR (diameter
of 2.2 Å).
The NaY-zeolite support was first treated in synthetic air
for 3 h at 723 K and under vacuum for 1 h at 723 K. The
material was cooled at room temperature in vacuum and then
stored under argon.
The TPR experiments were carried out in a conventional
apparatus using an H2/Ar mixture (5% H2). Catalysts sam-
ples of 20 or 100 mg, depending on the metal content, were
packed in a glass tube and pretreated in a flow of 4.9% O2/Ar
at 773 K for 1 h. After oxygen treatment, the samples were
purged under a flow of argon and then cooled to 223 K. The
TPR profiles were obtained in the 223–773 K temperature
range with a heating rate of 10 K/min at a constant flow rate
of 20 cm3/min.
Hydrogen chemisorption and sorption isotherms were
measured in a volumetric glass adsorption apparatus
equipped with a MKS Baratron type PDR-A pressure gauge.
An ultimate vacuum near 9.3 × 10−4 Pa could be obtained,
as measured by a Granville-Philips Model 274002 ionization
gauge. The amount of hydrogen absorbed and chemisorbed
on the exposed palladium atoms was measured after sample
reduction in flowing H2 at 573 K followed by overnight evac-
uation at the same temperature and cooling to room temper-
ature. Subsequently, the amount of absorbed hydrogen form-
ing ␣ or ␤ palladium-hydride was measured after room tem-
perature evacuation for 30 min. In this way, only the amount
of absorbed hydrogen is removed while that corresponding
to strongly chemisorbed hydrogen, (H/Pd)s, remains on the
surface [23]. Unfortunately, reliable measurements of this
2.1. Pd incorporation
In the pretreated support, palladium was incorporated by
two different methods: ionic exchange (IE) and dry impreg-
nation (DI), following the procedure described by Cañizares
et al. [22].
In the ionic exchange method, a Pd(NO3)2 solution (20 ml
of 0.05 mol/l for 1 g of support) was added in a pretreated
support at 298 K and under agitation for 2 h. Then, the sus-
pension was filtered and washed with deionizated water.
In the dry impregnation method, a specific quantity of
Pd(NO3)2 solution was added in a pretreated support. The
quantity used, equal to the pore volume of the support ma-
terial was calculated to place approximately 1% of palla-
dium by weight on the support. The solvent was removed
by evaporation.
The samples prepared by both methods were dried at
393 K for 14 h, and treated at 673 K in synthetic air for 3 h
and then reduced with H2 at 573 or 973 K for 1 h.
2.2. W incorporation
The W incorporation was carried out through a pho-
tochemical activation of the organometallic compound

S.B.C. Pergher et al. / Journal of Molecular Catalysis A: Chemical 209 (2004) 107–115
109
amount of hydrogen, which provides information on the
fraction of exposed palladium atoms, was only possible for
the Pd-W/NaY (DI) sample due to the unavailability of the
appropriate amount of the Pd/NaY (DI) catalyst and the
very small metal dispersion (large particles) on the catalysts
prepared by ionic exchange. Consequently, on samples with
high metal loading the palladium particle size was mea-
sured by X-ray diffraction line broadening. Experiments
were performed on a Siemens D 5000 instrument operated
at 30 kV and 30 mA with radiation from a Cu target. The
particle size was calculated by the Scherrer equation with
the peak width corresponding to the Pd (1 1 1) reflection.
the amount of palladium corresponds to an exchange capac-
ity of about 38%. Due to the high metal loading the palla-
dium particle size measured by XRD on the monometallic
and the bimetallic sample was quite large; 32–33 nm. On the
other hand, using the DI method a palladium sample with
a much lower metal uptake was obtained; 1.26%. We were
able to perform a hydrogen chemisorption measurement on
the bimetallic sample and the (H/Pd)s ratio that is equal to the
fraction of exposed palladium atoms was found to be 0.25.
Using this value the palladium particle size was estimated
from the equation: d (nm) = 1.12/(H/Pd)s = 4.5 as shown
in Table 1 [25]. Taking into account that the chemisorption
measurement was carried out on the bimetallic sample, the
palladium particle size may be smaller if W atoms are de-
posited on the metal surface precluding the adsorption of
hydrogen.
2.4. Catalytic activity
The catalytic activity for NO decomposition of the
prepared catalysts was measured in a glass flow reactor
(i.d. = 13 mm) coupled to a FTIR MB100-BOMEM spec-
trometer equipped with a multiple-reflexion gas cell (7.0 m
path length and 2.1 l volume). The reaction was studied as
a function of time at 573 K, using a feed mixture of NO
(500 ppm) in Ar. The glass reactor was filled with 120 mg of
catalyst and the flow rate was adjusted at 60 cm3/min. The
NO, NO2 and N2O stretching bands were monitored. The
bands intervals were 1955–1790 cm−1, 1658–1565 cm−1
and 2266–2159 cm−1, respectively. In order to calculate the
NO conversion from the IR data a method developed in
our laboratory by Dallago et al. [24] was used. It enables
us to determine with precision the NO concentration at the
entrance of the gas cell (reactor exit) from the absorbance
values measured.
3.1. Pd/ NaY-zeolite catalysts
The TPR profiles of the two Pd/NaY-zeolite catalysts are
shown in Fig. 1. It is important to notice that the mass of
catalyst used in each experiment was varied in order to com-
pensate for the large difference in metal content. The low
temperature peaks in these profiles are clearly due to the
reduction of palladium oxide crystals and ␤-hydride forma-
tion. The subsequent negative peaks are due to hydrogen
desorption (␤-hydride decomposition). The sample prepared
by ion exchange shows a much higher hydrogen uptake, the
reduction temperature is the lowest, 269 K and the peak due
to hydrogen desorption at 347 K is very sharp. This behav-
ior, which is typical of preparations with a low metal dis-
persion (large particles), is due to the high palladium con-
centration. Homeyer and Sachtler [26] have reported similar
TPR profiles for 2% Pd/NaY sample prepared by IE fol-
lowing oxidation at 773 K. The presence of large palladium
particles on the Pd/NaY (IE) sample, demonstrated by the
XRD measurements, was confirmed by the H2 sorption data.
As shown in Fig. 2 the ␣-hydride phase to ␤-hydride phase
transition is quite sharp.
On the other hand, the catalyst prepared by DI shows a
TPR profile with a higher reduction temperature, 295 K, the
peak is broader and the hydrogen desorption peak is located
at a higher temperature; 351 K. Due to the lower metal load-
ing, there is a stronger interaction with the zeolite leading
to small particles. In other words, the palladium dispersion
is much higher. In accordance with this situation the hy-
drogen sorption results in Fig. 2 demonstrate the absence
of the ␣–␤-hydride transition. It is important to realize that
the hydrogen uptakes in Fig. 2 correspond to the amount
absorbed, (H/Pd)abs, not to hydrogen adsorbed. Extrapola-
tion to zero pressure gives the amount of hydrogen due to
the ␣-hydride phase which is equal for both samples, as ex-
pected. On the other hand, extrapolation of the high pressure
3. Results and discussion
The main properties of the mono and bimetallic catalysts
are presented in Table 1. The very large palladium uptake
shown by the sample prepared by IE, 8.9%, is in accord with
the use of a concentrated solution of Pd(NO3)2. Taking into
account that the NaY-zeolite has a Si/Al ratio of around 2.4,
Table 1
Main catalyst properties
Catalysts
Pd
(wt.%)a
W
(wt.%)b
Pd particle
size (nm)
Pd/NaY (IE)
Pd/NaY (DI)
Pd-W/NaY (IE)
Pd-W/NaY (DI)
8.9
1.26
8.9
–
–
1.9
1.6
33c
–
32c
4.5d
1.26
IE: Pd incorporation by ionic exchange; DI: Pd incorporation by dry
impregnation.
a
Determined by AAS.
From the decrease of the ␯(CO) IR band of W(CO)6 in hexane
b
region for sample Pd/NaY(IE) gives a value of (H/Pd)abs
0.65 which is close to the palladium-hydride stoichiometry
for the ␤-hydride phase [27].
=
solution.
c
From XRD line broadening.
From H2 chemisorption; d (nm) = 1.12/(H/Pd)s.
d

110
S.B.C. Pergher et al. / Journal of Molecular Catalysis A: Chemical 209 (2004) 107–115
5000
4000
3000
2000
1000
269 K
295 K
(b)
351 K
0
(a)
347 K
-1000
100
200
300
400
500
600
700
800
Temperature (K)
Fig. 1. TPR profiles of Pd/NaY (IE), m = 100 mg (a) and Pd/NaY (DI), m = 20 mg (b) following oxidation in O2/Ar at 773 K.
3.2. Pd-W/NaY-zeolite catalysts
After the photochemical reaction, the DRIFTS spectra of
the solid did not show any ␯(CO) band, in opposition to the
situation where a ligand (PPh3) was used to stabilize the
tungstencarbonyl derivatives [10,11]. The beneficial effect
of this ligand in stabilizing surface species on NaY-zeolite
is shown in Fig. 3a. This spectrum is more complex than
the one obtained when the same reaction was performed
on alumina [12]. The bands might be attributed to either a
The amount of tungsten incorporated on the Pd/NaY-zeolite
samples was followed by the decrease of the ␯(CO) IR band
of [W(CO)6] (1983 cm−1) in the hexane solution, as it has
been explained elsewhere [10–14]. As shown in Table 1 the
tungsten loading does not depend on the amount of palla-
dium; it was slightly higher in the ion exchange preparation.
1.60
1.20
0.80
0.40
0.00
0.00
20.00
40.00
60.00
80.00
100.00
PRESSURE (Torr)
Fig. 2. Hydrogen sorption vs. pressure for Pd/NaY (IE) (᭜) and Pd/NaY (DI) (ᮀ) at 308 K.

S.B.C. Pergher et al. / Journal of Molecular Catalysis A: Chemical 209 (2004) 107–115
111
80
(a)
70
60
50
40
30
20
2200
2100
2000
1900
1800
1700
WaveNumber(cm-1)
80
(b)
70
60
50
40
30
20
2200
2100
2000
1900
1800
1700
Wave Number (cm-1)
100
90
80
70
60
(c)
2200
2100
2000
1900
1800
1700
Wave Number (cm-1)
Fig. 3. DRIFTS–FTIR spectra of NaY-zeolite and Pd/NaY (IE); photochemical activation of [W(CO)6] + PPh3 on NaY-zeolite (a); photochemical
activation of [W(CO)6] on NaY-zeolite (b); photochemical activation of [W(CO)6] on Pd/NaY (IE) (c).
stronger interaction of M–CO with the Lewis acid sites of
the support and a consequent lowering of the symmetry or
to different sites for coordination [14]. Examining this spec-
trum, it can be seen that the bands at 2073 and 1943 cm−1
can be attributed to the a1rad and a1ax + e modes, respec-
tively, of the pentacarbonyl derivative [W(CO)5PPh3]. The
band at 2073 cm−1 does not change in comparison with
that of alumina, but the one around 1943 cm−1 is splitted in
three, showing that these modes are more sensible to the dif-
ferent interactions with the zeolite acid sites. The bands at
2021, 1916, 1888 and 1867 cm−1 are related to b2, b1, a1, e,
a1 modes, respectively, of the cis-disubstituted compound,
cis-[W(CO)4(PPh3)2]. The weak bands at 1992, 1976 and
1966 cm−1 can be related to the t1u mode of [W(CO)6]
interacting in different sites. The band at 1896 cm−1 (e1u
mode) indicates the presence of the trans-disubstituted

112
S.B.C. Pergher et al. / Journal of Molecular Catalysis A: Chemical 209 (2004) 107–115
3000
2500
2000
1500
1000
500
296 K
278 K
(a)
(b)
330 K
342 K
329 K
0
100
200
300
400
500
600
700
800
Temperature (K)
Fig. 4. TPR profiles for Pd-W/NaY (DI), m = 100 mg (a) and Pd-W/NaY (IE) m = 20 mg (b) following oxidation in O2/Ar at 773 K.
species. In Fig. 3b and c are the spectra obtained in this
work for the NaY-zeolite support and the Pd/NaY-zeolite
(IE) catalyst without the ligand PPh3. As it can be seen,
only traces of [W(CO)6−n] species are observed, because
the subcarbonyl species decomposed rapidly upon exposure
of the solid sample to air. However, a specific interaction
between hexacarbonyl species and the metal surface takes
place during the photochemical reaction. Xu and Zaera [28]
have shown that a complete and rapid decarbonylation of
[W(CO)6] takes place when this compound is adsorbed on
a nickel film. In a previous study we have found evidence
that partially decarbonylated species adsorb and decompose
rapidly on the surface of reduced palladium [13].
decomposition peaks found for the ionic exchange sample
suggests the presence of different palladium particles; large
ones with a close interaction with tungsten characterized
by a desorption peak at 329 K and small ones located in
the ␣-supercage or the sodalite cage with a desorption at
342 K. Consequently, the TPR measurements demonstrate
that both preparation methods lead to a palladium-tungsten
interaction. There would appear to be an advantage of using
zeolites over alumina, because in our previous studies we
have found that tungsten hexacarbonyls have a tendency to
interact mainly with large palladium particles supported on
alumina [13].
The effect of tungsten addition on the oxidation-reduction
behavior of the Pd-W/NaY-zeolite catalysts is shown in
Fig. 4. The TPR profile for the Pd-W/NaY (DI)-zeolite sam-
ple presents sharp and narrow peaks for hydrogen consump-
tion and hydrogen desorption. The hydrogen consumption
peak is positioned at the same temperature shown by the
monometallic sample, 296 K, and the calculated peak area
demonstrates that despite the change in peak shape the same
amount of palladium is involved in the oxidation-reduction
process. On the other hand, the hydrogen desorption peak
appears at a lower temperature; 330 K. This behavior reflects
an interaction between palladium and tungsten that facili-
tates the decomposition of the ␤-hydride. The profile for the
Pd-W/NaY(IE)-zeolite catalysts reveals a large reduction in
hydrogen consumption (about 50%) relative to that of the
monometallic catalyst. In addition, the TPR profile exhibits
two hydrogen desorption peaks, located at 329 and 342 K.
One possible explanation is that the initial sample treatment
at 773 K in 5% O2 might be not sufficient to oxidize all Pd
atoms due to its interaction with tungsten. The two ␤-hydride
3.3. Catalytic activity and selectivity for NO decomposition
3.3.1. Pd/NaY (DI)-zeolite and Pd-W/NaY (DI)-zeolite
catalysts
Fig. 5 shows the activity for NO decomposition at 573 K
on samples reduced at the same temperature. Both the
monometallic and bimetallic catalysts exhibited a period of
very high activity followed by a steady deactivation due to
oxygen retention by Pd◦ atoms and perhaps WO3−x species
formed during the reduction pretreatment. It is generally
accepted that the catalytic decomposition of NO occurs
on reduced metals. However the oxygen produced by the
NO dissociation remains strongly adsorbed and depending
on the reaction temperature may even lead to oxidation of
the metal. It is shown in Fig. 5 that the bimetallic sample
deactivates faster, indicating that, under the chosen con-
ditions, the palladium-tungsten interaction decreased the
fraction of exposed palladium atoms. According to these
results the addition of tungsten to a Pd/NaY-zeolite catalyst
prepared by dry impregnation and containing a moderate

S.B.C. Pergher et al. / Journal of Molecular Catalysis A: Chemical 209 (2004) 107–115
113
120
100
80
60
40
20
0
(a)
(b)
0
50
100
150
200
250
300
350
400
Reaction Time (min)
Fig. 5. NO decomposition at 573 K. Conversion (%) vs. time on stream for samples reduced at 573 K. Pd/NaY (DI) (a) and Pd-W/NaY (DI) (b).
Fig. 6. NO decomposition at 573 K. Conversion (%) vs. time on stream for samples reduced at 973 K. Pd/NaY (DI) (a) and Pd-W/NaY (DI) (b).
palladium loading does not improve its activity for the NO
decomposition reaction. As observed in Fig. 6 similar sam-
ples reduced at 973 K behave quite differently. In this case,
the monometallic catalyst deactivates very rapidly because
large palladium particles, with a smaller fraction of exposed
metal atoms, are formed by the sintering process that takes
place in H2 at 973 K. In other words, the initial (H/Pd)s ratio
(≥0.25) decreases and the 4.5 nm particles growth to some
extent. Bulk oxidation of large particles is a slow process
that limits the rate of NO decomposition. On the other hand,
the presence of tungsten on Pd-W/NaY (DI)-zeolite leads to
a brief period of high activity followed by a slow and steady
deactivation. The improved behavior of this sample is in part
due to the presence of reduced tungsten species and also to
inhibition of the palladium sintering process in hydrogen.
In addition to changes in catalytic activity there is a pro-
motional effect of tungsten on the reaction selectivity to N2.
It is important to mention first that N2O and N2 were the
only reaction products; the formation of NO2 was negligi-
ble. Another characteristic of the NO decomposition reac-
tion on metals observed here was that the onset of N2O for-
mation is in accord with the initiation of the deactivation
process [12]. As shown in Table 2 the presence of tung-
sten decreases the selectivity to N2O (for a NO conversion
Table 2
Selectivity to N2O at 573 K for a NO conversion of 60%
Catalyst
Reduction
temperature
(K)
Reaction
time (min)
Selectivitya
to N2O (%)
Pd/NaY (DI)
Pd-W/NaY (DI)
Pd/NaY (DI)
Pd-W/NaY (DI)
Pd/NaY (IE)
Pd-W/NaY (IE)
Pd/NaY (IE)
573
573
973
973
573
573
973
973
320
238
25
300
>500
165
390
390
81
27
17
67
87b
60
75
52
Pd-W/NaY (IE)
a
Selectivity to N2O = 2 × [ppm N2O]/[ppm NO converted].
b
Extrapolated value.

114
S.B.C. Pergher et al. / Journal of Molecular Catalysis A: Chemical 209 (2004) 107–115
120
100
80
60
40
20
0
(a)
(b)
0
50
100
150
200
250
300
350
400
Reaction Time (min)
Fig. 7. NO decomposition at 573 K. Conversion (%) vs. time on stream for samples reduced at 573 K. Pd/NaY (IE) (a) and Pd-W/NaY (IE) (b).
of 60%) to a large extent, from 81 to 27% for the Pd/NaY
(DI)-zeolite sample reduced at 573 K. It is observed that
the reaction time to reach a conversion level of 60% for
the bimetallic sample is reduced, in accordance with the
loss of activity already discussed. The inhibited formation
of N2O on Pd-W/NaY (DI)-zeolite may be attributed to the
resistance of palladium to oxidation or to oxygen retention
in the presence of reduced tungsten species, as found on a
previous study carried out on Pd-W/␥-Al2O3 [29]. It was
demonstrated by Rottlander et al. [30] that surface oxygen
limits the dissociation of adsorbed NO and consequently the
NOad + Nad reaction is favored. Another interpretation for
the change in selectivity may be the promotion of oxygen
desorption. It was shown by Chuang and Tan [31] that the
addition of terbium oxide to a Pt/␥-Al2O3 catalyst facilitates
the desorption of oxygen during NO decomposition. Unfor-
tunately, we were not able to follow the evolution of oxy-
gen due to the use of FTIR analysis to follow the reaction
progress.
Turning our attention to the behavior of the Pd/NaY
(DI)-zeolite sample reduced at 973 K, Table 2 exhibits a
sharp decline in activity, 60% conversion in 25 min, but also
a very low selectivity to N2O. This is a surprising result be-
cause it is expected that large palladium particles favor the
formation of N2O [32]. In our opinion the reason for this
behavior is the slow response of the FTIR cell to sudden
changes in N2O concentration. Consequently, the effect of
the high reduction temperature on the reaction selectivity
of the Pd/NaY (DI)-zeolite sample is apparent.
3.3.2. Pd/NaY(IE)-zeolite and Pd-W/NaY(IE)-zeolite
catalysts
The monometallic sample exhibited a high and stable ac-
tivity for a period of 7 h when reduced at 573, as shown in
Fig. 7. When tungsten was incorporated there was a period
of 100% conversion of approximately 100 min followed by
a marked deactivation. Therefore, here again it seems that
the presence of tungsten in close association with palladium
120
100
80
60
40
20
0
(a)
(b)
0
50
100
150
200
250
300
350
400
Reaction Time (min)
Fig. 8. NO decomposition at 573 K. Conversion (%) vs. time on stream for samples reduced at 973 K. Pd/NaY (IE) (a) and Pd-W/NaY (IE) (b).

S.B.C. Pergher et al. / Journal of Molecular Catalysis A: Chemical 209 (2004) 107–115
115
particles inhibited the NO reaction. However, when the re-
References
duction temperature was 973 K, a more comparable perfor-
mance between the mono and the bimetallic catalyst was
obtained. As observed in Fig. 8 the Pd-W/NaY(IE)-zeolite
catalysts exhibited a conversion of 100% for about 3 h, fol-
lowed by a slow deactivation. This behavior is attributed
to the presence of reduced tungsten species. No change in
palladium particle size should be expected, because of the
very large particles on the fresh catalyst obtained by the IE
preparation method.
Turning the attention to the reaction selectivity, the bene-
ficial effect of tungsten addition on catalysts prepared by IE
was important for samples reduced at 573 and 773 K. The
results in Table 2 indicate a clear reduction on the unde-
sirable formation of N2O; the selectivity to this compound
decreased from 87 to 60% and from 75 to 52% depend-
ing on the reduction temperature. Here again several in-
terpretations could be offered to explain these results; the
presence of reduced tunsgten species may act as oxygen
scavenger or may facilitate the desorption of oxygen. More
specific studies are needed to discriminate between these
possibilities.
[1] H.S. Ghandi, H.C. Yao, H.K. Stepien, in: A.T. Bell, L.L. Hegedus
(Eds.), Catalysis Under Transient Conditions, A.C.S. Symposium
Series, 1982, 178, p. 143.
[2] K.M. Adams, H.R. Ghandi, Ind. Eng. Chem. Prod. Res. Dev. 22
(1983) 207.
[3] H.K. Plummer, S. Shinozaki, K.M. Adams, H.S. Ghandi, J. Mol.
Catal. 20 (1983) 251.
[4] H. Muraki, H. Shinjoh, H. Sobukawa, K. Yokota, Y. Fujitani, Ind.
Eng. Chem. Res. Dev. 25 (1986) 202.
[5] I. Halasz, A. Brenner, M. Shelef, K.Y. Simon Ng, Appl. Catal. A
82 (1992) 51.
[6] I. Halasz, A. Brenner, M. Shelef, Appl. Catal. B 2 (1993) 31.
[7] T. Sekiba, S. Kimura, H. Yamamoto, A. Okada, Catal. Today 22
(1994) 113.
[8] Y.J. Mergler, A. Van Aalst, J. van Delft, B.E. Nieuwenhuys, J. Catal.
161 (1996) 310.
[9] M.A.S. Baldanza, L.F. de Melo, A. Vannice, F.B. Noronha, M.
Schmal, J. Catal. 192 (2000) 64.
[10] I.M. Baibich, J.H. dos Santos, F. Stedile, V.E. Santarosa, I. Baumvol,
Polyhedron 16 (11) (1997) 1937.
[11] I.M. Baibich, J.H. dos Santos, V.C. da Silveira, C.E. Gigola, A.M.
Sica, Canadian J. Anal. Sci. Spectrosc. 43 (1) (1998) 26.
[12] A.M. Sica, J.H. dos Santos, I.M. Baibich, C.E. Gigola, J. Mol. Catal.
A 137 (1999) 287.
[13] A.M. Sica, I.M. Baibich, C. Gigola, J. Mol. Catal. A 195 (2003) 225.
[14] R.M.J. de Almeida, S.B.C. Pergher, C.E. Gigola, I.M. Baibich, Cana-
dian J. Anal. Sci. Spectrosc. 48 (1) (2003) 21.
[15] R.M. Dallago, PhD Thesis, Instituto de Qu´ımica, UFRGS, Brazil,
2002.
4. Conclusions
It has been demonstrated that the photochemical reaction
of tungsten hexacarbonyl with Pd/NaY-zeolite catalysts
of high (large particles) and low (small particles) metal
content enables the preparation of bimetallic samples.
The method of preparation leads to a palladium-tungsten
interaction that modifies the catalytic activity and the se-
lectivity of the Pd/NaY-zeolite catalysts for the NO de-
composition reaction. For samples reduced at high tem-
perature, the presence of tungsten prevents the sintering
of small palladium particles. However, the main effect
of tungsten is a reduction on the selectivity to N2O. In
other words, tungsten suppresses the undesirable formation
of N2O.
[16] S. Özar, G.A. Ozin, K. Moller, T. Bein, J. Am. Chem. Soc. 112 (26)
(1990) 9575.
[17] G.A. Ozin, S. Özar, H.O. Pastore, A.J. Pöe, E.J.S. Vichi, J. Chem.
Soc. Chem. Commun. (1991) 141.
[18] R. Jelinek, S. Özar, H.O. Pastore, A. Malek, G.A. Ozin, J. Am.
Chem. Soc. 115 (1993) 563.
[19] H.O. Pastore, G.A. Ozin, A.J. Pöe, J. Am. Chem. Soc. 115 (1993)
1215.
[20] C.J. Loughran, D.E. Resasco, Appl. Catal. B 7 (1995) 113.
[21] C. Descorme, P. Gélin, C. Lécuyer, M. Primet, J. Catal. 177 (1998)
352.
[22] P. Cañizares, A. de Lucas, F. Dorado, A. Durán, I. Asencio, Appl.
Catal. A 169 (1998) 137.
[23] J.L. Benson, H.S. Hwang, M. Boudart, J. Catal. 30 (1973) 146.
[24] R.M. Dallago, J. Schifino, I.M. Baibich, R.C. Vezes, PhD Thesis,
Instituto de Qu´ımica, Universidade Federal do Rio Grande do Sul,
2002.
[25] S. Ichikawa, H. Poppa, M. Boudart, J. Catal. 91 (1985) 1.
[26] S.T. Homeyer, W.M.H. Sachtler, Appl. Catal. 54 (1989) 189.
[27] M. Boudart, H.S. Hwang, J. Catal. 39 (1975) 44.
[28] M. Xu, J. Zaera, J. Vac. Sci. Technol. A 14 (1996) 415.
[29] A.M. Sica, C.E. Gigola, Appl. Catal. A 239 (1–2) (2003) 121.
[30] C. Rottlander, R. Andorf, C. Plog, B. Krutzsch, M. Baerns, J. Catal.
169 (1997) 400.
Acknowledgements
The authors acknowledge the financial support pro-
vided by CNPq (Brazil), FAPERGS (Brazil), COPESUL
(Brazil), CONICET (Argentina) and to Dr. John Spencer
(John Black foundation, England) for proof reading this
article.
[31] S.S.C. Chuang, Cher-Dip Tan, J. Phys. Chem. B 101 (1997) 3000.
[32] X. Xu, D.W. Goodman, Catal. Lett. 24 (1994) 31.