Transient infrared characterization of oxygen storage capability of Ce-Pd/Al2O3 catalyst during NO-CO reaction

Article abstract of DOI:10.1021/jp027859i

Oxygen storage and its effects on the adsorbates during NO-CO reaction have been studied by sequential pulsing of CO/O₂/CO into an NO-CO reactant stream over Pd/Al₂O₃ and Ce-Pd/Al₂O₃ catalysts. The ad

Full text of DOI:10.1021/jp027859i

4834
J. Phys. Chem. B 2003, 107, 4834-4843
Transient Infrared Characterization of Oxygen Storage Capability of Ce-Pd/Al2O3 Catalyst
during NO-CO Reaction
Scott A. Hedrick, Steven S. C. Chuang,* Khalid Almusaiteer, and Robert W. Stevens, Jr.
The UniVersity of Akron, Department of Chemical Engineering, Akron, Ohio 44325-3906
ReceiVed: December 30, 2002
Oxygen storage and its effects on the adsorbates during NO-CO reaction have been studied by sequential
pulsing of CO/O
2
/CO into an NO-CO reactant stream over Pd/Al
O
2 3
2 3
and Ce-Pd/Al O catalysts. The
adsorbates and reaction effluent were monitored concurrently by in situ infrared spectroscopy and mass
spectrometry, respectively. It was found that the presence of ceria minimized the impact of O and CO on the
adsorbates over the Ce-Pd/Al catalyst at 673 K. This effect diminished at 573 and 473 K along with its
2
2 3
O
storage capacity. Storage capacity is directly proportional to temperature in the range 473-673 K. Oxygen
storage occurs differently in the presence of NO-CO than in an inert helium/argon environment, which has
been used widely as the medium to study O
2
storage on ceria-based catalysts. The presence of the reactants
and CO to adsorb and react on the ceria surface prior to
2
pulses at 573 and 473 K, unlike those at 673 K, decreased linear/
and their adsorbates on the catalyst surface leads O
2
0
the Pd surface at 673 K. In addition, O
0
+
bridged CO and linear NO on Pd without producing Pd -NO, indicating that oxygen blocks adsorption
0
+
sites on Pd without oxidizing Pd to Pd . The active adsorbates for NO-CO decomposition are linear NO
0
and linear/bridged CO on Pd .
Introduction
takes place, and (ii) to determine the effect of oxygen storage
on the adsorbates’ reactivity during the NO-CO reaction. This
The automobile three-way catalytic converter uses Pd, Pt, and
Rh as key components in catalyzing NO reduction by CO,
hydrocarbon oxidation, and CO oxidation. These catalysts give
approach involves pulsing CO and O2 into the NO-CO reactant
flow and then determining the responses of adsorbates by in
situ infrared (IR) spectroscopy and the responses of gaseous
reactants and products by mass spectrometry (MS). Pulsing CO
into the NO-CO flow over the catalyst creates a reducing
environment for the catalyst, causing the withdrawal of stored
oxygen from the catalyst to form CO2; pulsing O2 into the NO-
CO flow produced an oxidizing environment, causing oxidation
of the reductants (i.e., CO and HC) and the uptake of oxygen
by the catalyst. The amount of stored oxygen available for the
reaction with CO in the reducing environment can be quantified
by measuring CO2 formation as well as CO/O2 conversion
during CO and O2 pulsing. The effect of oxygen storage on the
adsorbates and their reaction pathways can be elucidated from
MS responses of gaseous reactants/products and IR responses
of adsorbates. This paper addresses the role of ceria in altering
the effect of oxygen on the adsorbates and the catalyst activity;
this paper also underscores the importance of oxygen storage
measurement under practical reaction conditions.
1
an optimum conversion for NO, CO, and hydrocarbons (HC)
for the flue gas produced from stoichiometric combustion at an
2
air-to-fuel ratio (A/F) of 14.6. However, this air-to-fuel ratio
oscillates under practical conditions. Use of excess fuel in
combustion produces flue gas containing more HC relative to
O2, providing a reducing (fuel-rich) environment for the catalyst,
increasing NO conversion but decreasing HC and CO conver-
sion. The presence of excess air from fuel-lean combustion in
the flue gas produces an oxidizing environment, decreasing NO
conversion but giving a high conversion for CO and HC
oxidation.
One method that widens the A/F ratio window is the addition
of ceria to the catalyst. Ceria has long been known to exhibit
oxygen storage capability, allowing release of oxygen in a
reducing environment and uptake of oxygen in an oxidizing
3
environment. The key to this oxygen storage ability lies in the
ease in which ceria is able to undergo changes in oxidation state,
4
+
3+
from CeO2 (Ce ) in an oxidizing environment to Ce2O3 (Ce )
in a reducing environment and vice versa as shown by XPS
studies.4 The benefits of oxygen storage include (i) enhance-
ment of Pd catalyst effectiveness for the removal of NO, CO,
and HC under oscillating A/F conditions, and (ii) promotion of
Experimental Section
Catalyst Preparation and Characterization. A 2 wt % Pd/
,5
Al2O3 catalyst was prepared by incipient wetness impregnation
2
of γ-Al2O3 support (Alfa Chemicals, 100 m /g) with a PdCl2
(Alfa Chemicals) solution at a pH of 2.8 and a temperature of
6
the reaction of CO and/or HC with H2O in the fuel-rich mode.
3
33 K. The catalyst was calcined in flowing air at 673 K for 8
Although oxygen storage capacity of ceria has been quantified
h and then reduced in flowing hydrogen at 673 K for 8 h. The
Pd loading on the Pd/Al2O3 catalyst was determined to be 1.75
wt % by inductively coupled plasma (ICP) analysis (Galbraith
Laboratories, Inc.). X-ray diffraction (XRD) yielded an average
Pd crystallite size of 6.4 nm, corresponding to a dispersion of
7
-12
in the literature,
most measurements were done far from
practical operating conditions. We have devised an approach
to (i) determine the oxygen storage capability of Pd/Al2O3 and
Ce-Pd/Al2O3 under conditions where the NO-CO reaction
1
0%. Ce-Pd/Al2O3 (2 wt % Pd, 20 wt % Ce) was prepared by
*
Author to whom correspondence should be addressed.
coimpregnation with a Ce(NO3)3 and PdCl2 solution. XRD data
1
0.1021/jp027859i CCC: $25.00 © 2003 American Chemical Society
Published on Web 04/24/2003

Infrared Study of Oxygen Storage on Ce-Pd
J. Phys. Chem. B, Vol. 107, No. 20, 2003 4835
Figure 1. Experimental approach.
yielded an average Pd crystallite size of 5.2 nm, corresponding
to a dispersion of 12%.
Reaction Studies. The experimental apparatus, including an
in situ IR reactor cell with CaF2 windows, has been reported in
detail elsewhere.13 A schematic of the experimental setup used
in this study is outlined in Figure 1. For each experiment, a
total of 143 mg of fresh catalyst was used with approximately
15 mg of this total pressed into a self-supporting disk and placed
in the path of the IR beam. The remainder of the catalyst was
placed in the vicinity of the disk for the purpose of increasing
conversion. Prior to each experiment, the catalyst was heated
in flowing He at 673 K. The reaction was carried out at 473,
3
5
73, and 673 K and 0.1 MPa at a total flow rate of 30 cm /min
with each run carried out on fresh catalyst. The reactant mixture
consisted of equimolar flows of 1 vol % NO in He and 1 vol %
CO in He. Upon reaching steady-state reaction, a series of 0.5
cm CO and O2 pulses were introduced into the NO-CO
reactant flow by a 6-port valve as depicted in Figure 1. Flow
Figure 2. IR spectra and band assignments for 1% NO/1% CO flow
over Ce-Pd/Al
2 3 2 3
O and Pd/Al O under steady conditions prior to
3
pulsing at 673, 573, and 473 K.
“
A” represents the NO/CO reaction mixture, and flow “B”
TABLE 1: Steady-State NO Conversion
NO conversion (%)
2 3
Ce-Pd/Al Pd/Al O
represents the species being pulsed into A.
The first series of pulses into the NO-CO flow is CO, which
follows the sequence of Phase I as shown in Figure 1. The
objective of this series of CO pulses is to deplete kinetically
accessible oxygen on the catalyst by reacting to produce CO2.
CO pulses were made until all reactants, products, and adsor-
bates gave the same subsequent responses, indicating that all
the kinetically accessible oxygen available for reaction had
temperature (K)
O
2 3
6
5
4
73
73
73
100
75
55
100
95
37
the amount of CO consumed during CO and O2 pulses was
quantified by multiplying the area under the m/e response curve
by its responding factor. As a result of the relatively low
response factor of N2 at m/e ) 14, the amount of N2 produced
during the pulses can be not accurately determined.
3
reacted. After these pulses, 0.5 cm pulses of 100% O2 (Phase
II in Figure 1) were introduced into the reaction mixture to
saturate the surface with adsorbed oxygen and to determine its
effect on adsorbates and catalyst activity. Pulsing continued until
the catalyst was no longer able to uptake oxygen as evidenced
by oxygen breakthrough, which gave the same size O2 peak
for subsequent O2 pulses in the MS profile. CO pulses were
again introduced (Phase III), as shown in Figure 1, to react with
the stored oxygen deposited from the O2 pulses.
Results
Table 1 lists the NO conversion on Pd/Al O and Ce-Pd/
2
3
Al O at 473, 573, and 673 K. The steady-state NO conversion
2
3
increased with temperature on both catalysts. The activities of
both catalysts did not change over the course of each experiment
at each respective temperature. NO conversion on Ce-Pd/Al2O3
is less sensitive to temperature in the range 473-673 K than
on Pd/Al2O3. The results suggest that the reaction over Ce-
Pd/Al2O3 has a lower activation energy than over Pd/Al2O3.
Similarly, addition of ceria to a Rh/Al2O3 catalyst was found
IR spectra were collected during each pulse by a Nicolet
Magna 550 spectrometer equipped with an MCT-B detector at
-
1
a resolution of 4 cm . Gaseous effluent was monitored by a
Balzers QMG 112 mass spectrometer for experiments with Pd/
Al2O3 and by a Prisma QMS 200 mass spectrometer for
experiments with Ce-Pd/Al2O3. We monitored the following
m/e ratios: 4 (He), 12 (CO), 22 (CO2), 28 (CO and N2), 30
1
4
to cause a reduction in activation energy.
(
NO), 32 (O2), 44 (CO2 and N2O), and 46 (NO2). The
Figure 2 compares the IR spectra of the steady-state NO-
CO reaction, corresponding to the NO conversion listed in Table
1. The steady-state reaction at 473 K produced linear CO at
contribution of CO2 and N2O to m/e ) 44 can be resolved by
CO2 calibration which shows that the area of the 44 peak is
4
pulse. The contribution of CO and N2 to the m/e ) 28 was
resolved by calibrating the intensity ratio of m/e ) 28 to m/e )
1
-
1
-1
0
1.6 times the area of the m/e ) 22 peak for any given CO2
2075 cm , bridged CO at 1902 cm , and linear NO on Pd at
-
1
1750 cm on both catalysts as well as a very strong bidentate
-
1
carbonate band on Ce-Pd/Al2O3 at 1595 cm . The prominent
4 for N2. The amount of CO2 and N2O produced as well as
bands, i.e., linear CO, bridged CO, linear NO, and carbonate,

4836 J. Phys. Chem. B, Vol. 107, No. 20, 2003
Hedrick et al.
Figure 3. CO/O
with the Phase II O
2
pulsing into 1% NO/1% CO flow over Ce-Pd/Al
and Phase III CO pulses, respectively.
O
2 3
at 673 K. The amount of oxygen stored and CO
2
produced are listed along
2
2
Figure 4. CO/O
2
pulsing into 1% NO/1% CO flow over Pd/Al
pulses, respectively.
O
2 3
2
at 673 K. The amount of oxygen stored and CO produced are listed along with
the Phase II O and Phase III CO
2
2
produced at 473 K were not observed on either catalyst at 573
K, indicating that the rates of conversion of these species are
at m/e ) 22 and 44. A positive peak indicates an increase in
the concentration of each species with respect to the steady-
state concentration. The positive peak at m/e ) 30, correspond-
ing to an increase in NO concentration, indicates a decrease in
the NO conversion as a result of the CO pulse. The amounts of
CO2 and N2O produced as well as the amount of CO consumed
for each CO and O2 pulse are listed in Table 2. The amount of
CO2 estimated from the peak area of the m/e ) 22 curve should
be considered as a lower limit as a result of the high baseline
of CO2 produced from the NO-CO reaction. The shift in the
CO baseline in Phase I in Figure 3 may lead to an overestimation
of CO consumption. The second and third CO pulses produced
nearly identical MS responses to the first pulse, suggesting that
the amount of oxygen available for the reaction with each of
these CO pulses was the same.
-
1
greater than their formation rates. The Al-NCO at 2239 cm ,
-
1
-1
bidentate nitrato at 1631 cm , carbonate at 1565 cm , and
-
1
chelating nitrato at 1473 cm emerged on both catalysts at 573
K. The commonality of the carbonate and nitrato bands on both
catalysts indicates that these species are most likely adsorbed
on the support. These band assignments are well established in
the literature.1
5-24
At 673 K, the Ce-Pd/Al2O3 catalyst produced linear CO at
-
1
-1
2
094 cm and a new band at 2008 cm which has been
25
assigned as a compressed bridged CO species. Because the
reaction achieved nearly complete conversion at this tempera-
ture, the presence of these species suggests that these species
are in excess with respect to adsorbed NO on the catalyst
surface.
The first O2 pulse in the Phase II sequence resulted in peaks
at m/e ) 22, 28, 30, and 44, indicating increases in the formation
of CO2/N2O as well as a decrease in NO conversion. Subsequent
pulses (i) produced smaller amounts of CO2 as indicated by
decreases in the m/e ) 22 and 44 peaks, (ii) decreased the rate
of NO conversion as shown by increases in the NO peaks, and
Figures 3 and 4 show the MS transient responses for the entire
experimental run at 673 K on Ce-Pd/Al2O3 and Pd/Al2O3,
respectively, providing an overall picture of CO and O2 pulses
and their product responses. Figure 3 shows that the first CO
pulse in Phase I produced CO2 and N2O, as evidenced by peaks

Infrared Study of Oxygen Storage on Ce-Pd
J. Phys. Chem. B, Vol. 107, No. 20, 2003 4837
TABLE 2: Quantification of CO
2
, N
2
O, CO, and O
2
for CO
Figure 4 shows that Pd/Al2O3 took a significantly lower
number of O2 pulses in Phase II to saturate the catalyst surface
with oxygen than Ce-Pd/Al2O3; Pd/Al2O3 also took less CO
to restore the initial CO response than Ce-Pd/Al2O3. This
observation revealed that the ceria component on Ce-Pd/Al2O3
is capable of enhancing oxygen uptake during O2 pulses.
and O
2
Pulses on Ce-Pd/Al
2
O
3
and Pd/Al O at 673 K
2 3
CO
2
N
2
O
CO
O
2
O
2
produced produced consumed consumed stored
pulse
(µmol)
(µmol)
(µmol)
(µmol)
(µmol)
Ce-Pd/Al
2 3
O
Phase I
To gain insight into the reaction involving CO and O2 pulses,
the IR spectra collected during the first CO pulse in Phase I
were plotted in Figure 5. Prior to the CO pulse, the NO-CO
1
2
3
st CO
nd CO
rd CO
1.1
0.8
0.5
0.5
0.7
0.4
2.1
1.5
0.3
-
1
Phase II
reaction produced Al-NCO at 2250 cm , linear CO at 2093
-
1
-1
1
2
3
4
5
6
7
st O
nd O
rd O
2
3.3
1.8
0.8
0
0
0
1.4
0.5
0.8
1.1
0.7
0.7
0.7
- -
20.5
20.5
20.5
20.5
14.4
0
18.8
19.5
20.0
20.5
14.4
0
cm , compressed bridged CO at 2012 cm , bidentate nitrate
- -
- -
- -
- -
- -
- -
-1
-1
2
at 1630 cm , and bridged carbonate at 1560 cm . As described
earlier, the presence of adsorbed CO on Pd surface indicates
2
0
th O
th O
th O
th O
2
these species are in excess with respect to adsorbed NO on the
catalyst surface. This excess adsorbed CO appears to keep the
catalyst surface in the reduced state. Further addition of CO by
pulsing did not produce any appreciable change in the intensity
of the adsorbates. The intensity of a number of prominent bands
was plotted as a function of time along with the MS profiles in
Figure 5b,c. The major effect of the CO pulse was to increase
the formation of CO2 and N2O. The subsequent CO pulses gave
the same responses as the first CO pulse.
2
2
2
0
3.1
3.1
Phase III
1
2
3
4
5
6
7
st CO
nd CO
rd CO
th CO
th CO
th CO
th CO
4.5
6.7
6.5
3.8
2.6
2.2
1.5
1.7
1.8
1.8
0.2
0
0.2
0.5
7.2
8.0
7.2
4.0
2.8
2.5
1.5
The MS profiles of reactants/products as well as IR profiles
of the specific adsorbates for the first CO pulse in Phase I
Total
96.3
2 3
Pd/Al O
(Figure 5b,c) on Ce-Pd/Al2O3 were plotted along with selected
Phase I
O2 pulses in Phase II and CO pulse in Phase III in Figures 6
and 7 to highlight the effect of oxygen storage on the adsorbates’
reactivity. The first O2 pulse led to the formation of a significant
amount of CO2 and N2O, but little changes in adsorbate
intensity. Figures 6 and 7 show that the IR intensities of
adsorbates after the first O2 pulse are virtually the same as those
following CO pulses in Phase I. The significant changes in the
adsorbate intensities were observed during the second O2 pulse,
1
2
st CO
nd CO
0
0
0.4
0.4
0
7.3
Phase II
1
2
3
4
5
st O
nd O
rd O
2
1.6
0
0
0
0
0.7
0.5
0.6
0.2
0.2
20.5
20.5
15.5
3.0
19.7
20.5
15.5
3.0
2
2
th O
th O
2
2
0
0
Phase III
+
-1
1
2
3
st CO
nd CO
rd CO
4.3
0.5
0
1.6
0.3
0.6
8.0
4.8
1.1
which caused emergence of linear NO on Pd at 1795 cm
(
(
Figure 6) and a decrease in the nitrate and carbonate intensities
Figure 7). The second O2 pulse also dramatically reduced the
Total
58.7
intensities of linear CO, compressed bridged CO, and Al-NCO.
Subsequent O2 pulses, as represented by the seventh O2 pulse
(
iii) eventually led to oxygen breakthrough at the fifth O2 pulse.
+
in Figures 6 and 7, caused (i) an increase in linear NO on Pd ,
The oxygen introduced to the catalyst in a pulse should follow
one of three paths: (i) oxygen storage, (ii) reaction with CO to
form CO2, or (iii) departure from the reactor as part of the
effluent. Thus, the amount of oxygen stored from each O2 pulse
can be obtained from the following mole balance:
0
(
ii) further diminution of linear CO (i.e., Pd -CO), compressed
bridged CO, and Al-NCO, and (iii) decreases in NO conversion
as evidenced by increases in the NO profile. The observation
0
+
of the depletion of Pd -O and the formation of Pd -NO is a
strong indication that all the Pd surface atoms are oxidized as
the catalyst surface becomes saturated with oxygen.
Mol O stored ) Mol O in - Mol O out -
2
2
2
The first CO pulse in Phase III following the saturation of
the catalyst surface with oxygen produced significantly more
CO2 than the first CO pulse in Phase I. This enhanced CO2
production is a clear indication that the stored oxygen on the
catalyst surface is active for reaction with the CO coming from
the CO pulse. However, like the first O2 pulse in Phase II, the
first CO pulse had no impact on the adsorbates. The significant
variation of adsorbate intensity was observed during the second
1
/2
Mol CO produced
2
These quantities are listed in Table 2 and Figures 3 and 4
along with the O2 MS profiles. The sixth and seventh O2 pulses
produced the same MS responses, suggesting that the catalyst
surface had become saturated with oxygen. Upon saturation,
the O2 pulses resulted in an inhibition of the NO-CO reaction
as indicated by the emergence of the prominent NO peak on
the fifth and subsequent pulses.
-
1
CO pulse which gave rise to linear CO at 2055 cm , bridged
-1
CO at 1904 cm , Al-NCO, carbonate, and nitrate species. The
IR spectra for adsorbed CO here are vastly different from those
given for the initial CO pulse in Phase I, suggesting oxygen
treatment severely modified the catalyst surface. The major
Following saturation of the catalyst with stored oxygen, CO
pulses were again introduced to react with the stored oxygen
(Phase III). The stored oxygen reacted with the CO pulse,
+
effects of the CO pulses are to eliminate the Pd -NO species
producing CO2. The initial CO pulse produced a significant
amount of additional CO2 relative to the Phase I CO pulses.
By the fifth CO pulse of Phase III, the amount of CO2 decreased
to the same level as the CO pulses introduced during Phase I,
indicating that all stored oxygen introduced from the O2 pulses
in Phase II has been depleted.
0
and to promote the formation of Pd -CO and bridged CO
+
0
species, indicating that Pd is converted to Pd .
Figures 8 and 9 show the MS and IR profiles for CO and O2
pulses on Pd/Al2O3 and are analogous to Figures 6 and 7. The
CO pulses in Phase I caused an increase in N2O and Al-NCO

4838 J. Phys. Chem. B, Vol. 107, No. 20, 2003
Hedrick et al.
2 3
Figure 5. IR spectra and MS profiles for the first CO pulse in Phase I over Ce-Pd/Al O at 673 K.
Figure 6. MS profiles and IR adsorbate intensities as a function of time for selected CO/O
2
2 3
pulses in each phase over Ce-Pd/Al O at 673 K.
-
1
Adsorbates shown are those above 1700 cm
.

Infrared Study of Oxygen Storage on Ce-Pd
J. Phys. Chem. B, Vol. 107, No. 20, 2003 4839
Figure 7. IR adsorbate intensities as a function of time for selected CO/O
2
2 3
pulses in each phase over Ce-Pd/Al O at 673 K. Adsorbates shown
-
1
are those below 1700 cm
.
2 2 3
Figure 8. MS profiles and IR adsorbate intensities as a function of time for selected CO/O pulses in each phase over Pd/Al O at 673 K. Adsorbates
-
1
shown are those above 1700 cm
.
formation. The absence of an increase in CO2 formation suggests
the lack of excess oxygen on catalyst surface. The first O2 pulse
in Phase II caused increases in CO2, N2O, and Pd -NO
concentration, but decreased IR intensity of Al-NCO and nitrate
species. This is in contrast to the first O2 pulse over Ce-Pd/
Al2O3 which did not lead to any appreciable change in intensity
of adsorbates. Subsequent O2 pulses gave a similar IR response,
caused a decrease in NO conversion, but little variation in CO2
concentration.
The same experimental sequences were carried out over both
catalysts at 573 and 473 K. Figure 10 shows the IR spectra,
MS profiles, and relative adsorbate peak heights as a function
of time for the first CO pulse in Phase I on Ce-Pd/Al2O3 at
473 K. This CO pulse elicits substantial increases in both linear
and bridged CO with a concomitant reduction in linear NO.
All other adsorbates including nitrate and carbonate species
remained constant during the pulse. The MS profiles show that
CO pulsing produced CO2/N2O and caused a small decrease in
NO conversion.
+
The first CO pulse in Phase III produced substantially more
CO2 formation than the CO pulses during Phase I. This CO
The MS profiles of reactants/products as well as IR profiles
of the specific adsorbates for the first CO pulse in Phase I
(Figure 10b,c) were plotted along with selected O2 pulses in
Phase II and CO pulse in Phase III in Figure 11. The major
-
1
pulse also produced linear CO at 2050 cm and bridged CO
-
1
+
at 1910 cm as well as eliminated Pd -NO, reflecting the
+
0
reduction of Pd to Pd .

4840 J. Phys. Chem. B, Vol. 107, No. 20, 2003
Hedrick et al.
Figure 9. IR adsorbate intensities as a function of time for selected CO/O
2
pulses in each phase over Pd/Al
O
2 3
at 673 K. Adsorbates shown are
-
1
those below 1700 cm
.
Figure 10. IR spectra and MS profiles for the first CO pulse in Phase I over Ce-Pd/Al
O
2 3
at 473 K.
-
1
differences between CO and O2 pulses at 473 and 673 K are
the absence of oxygen storage and the lack of variation in nitrate
and carbonate intensities at 473 K as shown in Figure 12.
Because the conversion of NO is at a level of 55%, the Pd
surface is populated with adsorbed CO and NO, reflecting that
the rate of adsorption of NO and CO is higher than the rate of
conversion of these species. The major effect of the CO pulse
is to displace linear NO with linear and bridged CO, causing a
decrease in NO conversion as evidenced by an increase in NO
concentration during CO pulses.
and linear NO at 1752 cm . Unlike the results on Ce-Pd/
+
-1
Al2O3 at 673 K, no formation of the Pd -NO at 1790 cm
0
was observed, although a slight, temporary blue shift in the Pd -
NO peak was observed, indicating that the Pd sites briefly
carried a slightly positive charge. This indicates that the O2
pulses on Ce-Pd/Al2O3 at 473 K act to block adsorption sites
for NO and CO rather than oxidize the surface. The NO site-
blocking effect brought about by O2 decreased with consecutive
O2 pulses. By the thirteenth O2 pulse, the O2 pulse inhibited
0
the conversion of Pd -NO more than blocked the NO adsorp-
tion sites, resulting in a slight increase in Pd -NO intensity
during the O2 pulse. CO pulses in Phase III produced adsorbates
similar to those in Phase I, suggesting the O2 pulse did not lead
to significant modification of the catalyst surface at this
0
The first O2 pulse at 473 K in Figure 11, unlike its counterpart
at 673 K, caused an immediate decrease in NO conversion, as
evidenced by the peak at m/e ) 30. This O2 pulse also caused
-1
-1
a decrease in linear CO at 2092 cm , bridged CO at 1934 cm ,

Infrared Study of Oxygen Storage on Ce-Pd
J. Phys. Chem. B, Vol. 107, No. 20, 2003 4841
Figure 11. MS profiles and IR adsorbate intensities as a function of time for selected CO/O
2
2 3
pulses in each phase over Ce-Pd/Al O at 473 K.
-
1
Adsorbates shown are those above 1700 cm
.
Figure 12. IR adsorbate intensities as a function of time for selected CO/O
2
2 3
pulses in each phase over Ce-Pd/Al O at 473 K. Adsorbates shown
-
1
are those below 1700 cm
.
temperature. Although adsorbed CO and NO on Pd/Al2O3
exhibited different intensities from those on Ce-Pd/Al2O3, their
dynamic behaviors during CO and O2 pulses are very similar
on both catalysts.
that oxygen storage and its impact on the catalysis of the NO-
CO reaction is a kinetic phenomenon which should be measured
under practical reaction conditions. It is expected that oxygen
may interact and react with the adsorbed species and catalyst
surface at different rates under different reaction conditions. A
critical question to be addressed is what is the reaction pathway
for oxygen uptake and withdrawal. As a result of the inability
of in situ infrared spectroscopy to directly observe the behavior
Discussion
The significant difference in adsorbate and oxygen storage
behaviors during O2/CO pulses at 473 and 673 K demonstrates

4842 J. Phys. Chem. B, Vol. 107, No. 20, 2003
Hedrick et al.
of adsorbed oxygen and its pathway, the reaction pathway of
oxygen has to be elucidated from the variation in IR-observable
adsorbates and reactant/product profiles during O2 and CO
pulses.
TABLE 3: Percent Decreases in NO Conversion during
Phase I CO Pulses
NO conversion decrease (%)
temperature (K)
Ce-Pd/Al
O
2 3
2 3
Pd/Al O
The Role of Ceria in the NO-CO Reaction. Results in
Figures 3 and 4 show that CO pulses in Phase I withdrew any
kinetically accessible oxygen by reaction to produce CO2. These
CO pulses rendered the Pd surface to the reduced state, which
chemisorbs CO as linear and bridged CO and gives a strong
Al-NCO intensity. A high concentration of CO, which keeps
the catalyst surface in the reduced state, is evidenced by a high
Al-NCO intensity. Isocyanate has been observed in a number
6
5
473
73
73
1.5
2.1
4.7
8.3
0
0
oxygen and makes extrapolation of the results obtained from
ideal conditions to practical conditions unreliable.
Ceria minimizes the impact of not only initial O pulses but
2
also CO pulses on the adsorbates. Significant increases in linear/
2
1-23,26-28
of NO-CO reaction studies on Pd and Rh catalysts.
bridged CO and Al-NCO were observed immediately following
the CO pulse in Phase III over Pd/Al2O3. In contrast, these
species began reemerging following the second and third CO
pulse over Ce-Pd/Al2O3. The initial CO pulse, much like the
initial O2 pulse, produced CO2 without altering the intensity of
adsorbates on the Pd surface of the Ce-Pd/Al2O3 catalyst. This
observation revealed that CO, like O2, interacts with ceria before
adsorbing on the Pd surface at 673 K. This is supported by a
study on Pt/ceria catalysts which showed that ceria has the
The species that are associated with a reducing environment
and reduced catalyst surface are depleted by O2 pulses that
+
produced Pd -NO at 673 K. Thus, the intensity of adsorbed
0
species such as Pd -CO (i.e., linear CO), bridged CO, and Al-
NCO can serve as an index of the Pd surface state.
The absence of variation of linear/bridged CO and Al-NCO
during the first O2 pulse over Ce-Pd/Al2O3 at 673 K shows
that the O2 from this pulse adsorbs on surface sites that are not
associated with these adsorbed species. The substantial decrease
1
9,32
ability to uptake CO.
with oxygen from ceria.
Adsorbed CO on Pd can also react
3
3
+
in intensity of these species and increase in intensity of Pd -
Although the in situ IR technique can be used to probe the
dynamics of adsorbed NO and CO as well as the oxidation state
of the metal, it cannot be used to directly observe the surface
structure or oxidation state of the ceria. We can only infer from
literature data that ceria is the key component for uptaking and
releasing excess oxygen. Studies using techniques such as
NO began with the second O2 pulses, revealing oxidation of
0
+
Pd to Pd . These observations showed that O2 adsorbs and
0
+
reacts with ceria before oxidizing Pd to Pd . This observation
is in contrast to the proposed role of noble metals in assisting
oxygen spillover from the metal to Ce2O3 for the reaction:
1
7,10
Ce2O3 + /2O2 f 2CeO2.
However, an oxygen exchange
3
4
4,5
study using 1 O2, O2, and O O over Pt/Al2O3 and Pt/CeO2
6
18
16 18
XANES/EXAFS and XPS
have directly observed the
changing oxidation state of ceria from Ce2O3 to CeO2 under
oxidizing conditions and from CeO2 to Ce2O3 (or oxygen-
deficient CeO2) under reducing conditions.
at 673 K revealed that oxygen exchange did not proceed via
Pt, but occurred directly on ceria.3 A bifunctional oxygen
-
reaction pathway, which included both direct adsorption of O2
29
Reactivity of Adsorbates. The absence of variation in IR
intensity of carbonate, nitrate, and chelating nitrato species
during CO and O2 pulses at 473 K (shown in Figure 12)
indicates that these species are spectators that are not involved
in catalysis of the NO-CO reaction. These species began
interacting with CO and O2 at 673 K (as shown by variation in
their intensities in Figures 7 and 9); however, they are not
involved in catalytic conversion of NO, as evidenced by lack
of relationship between their IR intensity and NO conversion.
(g) on both ceria sites and metal sites, has also been proposed.
Furthermore, O2 adsorption on metal sites can be strongly
suppressed by adsorbed CO.30
During the NO-CO reaction at 673 K, the presence of
adsorbed NO/CO and their decomposition products on the Pd
surface may lead oxygen to adsorb on the Ce2O3 surface prior
0
to the Pd surface. In the absence of ceria, the O2 pulse into
NO/CO over Pd/Al2O3 led to an immediate alteration of
adsorbates on the Pd surface. Thus, it can be concluded that
the major role of ceria on Ce-Pd/Al2O3 catalyst is to uptake
oxygen and minimize its impact on the adsorbates on the Pd
surface during the initial O2 pulses. This ceria effect is
diminished at 573 and 473 K in which the first O2 pulse alters
the intensity of adsorbates on the Pd surface for both Pd/Al2O3
and Ce-Pd/Al2O3 (see Figure 11). The O2 pulses at 473 K
0
Pd -NO and linear/bridged CO are the species whose intensities
vary with NO conversion during CO and O2 pulses over Pd/
Al2O3 and Ce-Pd/Al2O3. CO pulses into NO/CO caused
0
replacement of Pd -NO by linear/bridged CO as evidenced by
increases in linear/bridged CO with a concomitant decrease in
0
Pd -NO as well as NO conversion; O2 pulses caused decreases
0
in intensity of linear/bridged CO and Pd -NO as well as NO
(
shown in Figure 11) decreased linear/bridged CO and linear
0
conversion. The close relationship between Pd -NO intensity
0
+
NO on Pd without producing Pd -NO, indicating that oxygen
blocks the Pd sites for CO and NO adsorption without oxidizing
Pd to Pd .
0
and NO conversion suggest that Pd -NO is an active adsorbate
involved in NO conversion. This proposition is also supported
0
+
26
by a previous selective poisoning/selective enhancement study.
The replacement of Pd -NO by linear/bridged CO further
revealed that both NO and CO compete for the same Pd
0
The oxygen uptake on ceria has been found to increase with
3
,10,31
0
temperature,
which is consistent with our results. In their
1
7,26
CO/O2 pulsing into He study, Yao and Yu Yao found that
oxygen storage capacity of CeO2/Al2O3 at 573 K was nil and
sites.
The effect of CO pulsing on NO conversion varied by catalyst
and by temperature. Table 3 shows the effect of CO pulsing
(Phase I) on both catalysts at 473, 573, and 673 K. Phase I CO
pulsing on Ce-Pd/Al O caused decreases in NO conversion,
1
0
at 673 K was very limited (0.0003 µmol O2/µmol CeO2).
However, addition of a metal (Pd, Pt, or Rh) to the ceria
facilitated the oxygen storage capacity at temperatures from 573
to 773 K. Most studies typically characterized the oxygen
storage capability of ceria in an ideal, i.e., inert, environment,
typically He, but the presence of NO/CO and their adsorbates
under reaction conditions complicates the reaction pathway of
2
3
indicating that the rate is negative order in CO partial pressure.
This is consistent with another recent study on Pd/ceria
3
5
catalysts. However, on Pd/Al2O3 at 473 and 573 K, Phase I
CO pulses elicited no changes in NO conversion. These different

Infrared Study of Oxygen Storage on Ce-Pd
J. Phys. Chem. B, Vol. 107, No. 20, 2003 4843
behaviors suggest that the kinetics for the NO-CO reaction is
rather complex and that the elementary steps involved in the
reaction are very sensitive to reaction temperature.
(8) Hickey, N.; Fornasiero, P.; Graziani, M.; Blanco, G.; Bernal, S.
Chem. Commun. 2000, 5, 357.
(
9) Monte, R. D.; Fornasiero, P.; Graziani, M.; Ka sˇ par, J. J. Alloys
Compd. 1998, 275-277, 877.
10) Yao, H. C.; Yao, Y. F. Y. J. Catal. 1984, 86, 254.
(
Conclusions
(11) Hori, C. E.; Permana, H.; Ng, K. Y. S.; Brenner, A.; More, K.;
Rahmoeller, K. M.; Belton, D. Appl. Catal. B 1998, 16, 105.
(12) Cho, B. K. J. Catal. 1991, 131, 74.
0
Linear NO and linear/bridged CO on Pd sites are active
adsorbates for the NO-CO reaction. Ceria minimizes the effect
of the oxidizing (O2) and reducing agent (CO) on the adsorbates
on Ce-Pd/Al2O3 at 673 K. This ceria effect diminished at 573
and 473 K. Comparison of oxygen storage during NO-CO
reaction from this study and in O2/CO pulses in helium from
literature results show that the presence of adsorbed NO/CO
and their decomposition products on the Pd surface lead O2 and
(13) Chuang, S. S. C.; Brundage, M. A.; Balakos, M. W.; Srinivas, G.
Appl. Spectrosc. 1995, 49, 1151.
(
(
(
14) Oh, S. H. J. Catal. 1990, 124, 477.
15) Grill, C. M.; Gonzalez, R. D. J. Phys. Chem. 1980, 84, 878.
16) Palazov, A.; Chang, C. C.; Kokes, R. J. J. Catal. 1975, 36, 338.
(17) Almusaiteer, K.; Chuang, S. S. C. J. Catal. 1999, 184, 189.
18) Kikuzono, B. Y.; Kagami, S.; Naito, S.; Onishi, T.; Tamaru, K.
Faraday Discuss. Chem. Soc. 1982, 72, 135.
19) Binet, C.; Badri, A.; Boutonnet-Kizling, M.; Lavalley, J. C. J. Chem.
(
(
0
CO to adsorb and react on the ceria surface prior to the Pd
Soc., Faraday Trans. 1994, 90, 1023.
surface. The O2 pulses at 473 K, unlike those at 673 K,
(20) Raval, R.; Blyholder, G.; Haq, S.; King, D. A. J. Phys. Condens.
Matter 1989, 1, SB165.
decreased linear/bridged CO and Al-NCO without producing
(
(
(
21) Unland, M. L. J. Catal. 1973, 31, 459.
22) Solymosi, F.; Rasko, J. J. Catal. 1980, 63, 217.
23) Hecker, W. C.; Bell, A. T. J. Catal. 1984, 85, 389.
+
Pd -NO, indicating that oxygen blocks the Pd sites for CO
0
adsorption and for Al-NCO formation without oxidizing Pd
+
to Pd .
(24) Nakamoto, K. Infrared and Raman Spectra of Inorganic and
Coordination Compounds; Wiley: New York, 1986.
(
25) Bensalem, A.; Muller, J.-C.; Tessier, D.; Bozon-Verduraz, F. J.
Chem. Soc., Faraday Trans. 1996, 92, 3233.
26) Almusaiteer, K.; Chuang, S. S. C. J. Catal. 1998, 180, 161.
Acknowledgment. This work was supported by the National
Science Foundation under Grant CTS-942111996. We thank Mr.
Toshitaka Tanabe at Toyota Central R&D Lab for helpful
suggestions.
(
(27) Krishnamurthy, R.; Chuang, S. S. C.; Balakos, M. W. J. Catal.
995, 157, 512.
1
(28) Chuang, S. S. C.; Tan, C.-D. J. Catal. 1998, 173, 95.
(
29) Nibbelke, R. H.; Nievergeld, A. J. L.; Hoebink, J. H. B. J.; Martin,
References and Notes
G. B. Appl. Catal. B 1998, 19, 245.
(
30) Rodriguez, J. A.; Goodman, D. W. Surf. Sci. Rep. 1991, 14, 27.
(
(
(
(
(
1) Shelef, M. Catal. ReV.sSci. Eng. 1994, 36, 443.
(31) McCabe, R. W.; Jen, H.-W.; Chun, W.; Graham, G. W.; Haack, L.
2) Taylor, K. C. Catal. ReV.sSci. Eng. 1993, 35, 457.
3) Holmgren, A.; Anderson, B. J. Catal. 1998, 178, 14.
4) Schmieg, S. J.; Belton, D. N. Appl. Catal. B 1995, 6, 127.
5) Putna, E. S.; Gorte, R. J.; Vohs, J. M.; Graham, G. W. J. Catal.
P.; Straccia, A.; Benson, D. Appl. Catal. A 1999, 184, 265.
(32) Holmgren, A.; Andersson, B.; Duprez, D. Appl. Catal. B 1999,
22, 215.
1
998, 178, 598.
(33) Cordatos, H.; Gorte, J. J. Catal. 1996, 159, 112.
(34) Holles, J. H.; Davis, R. J. J. Phys. Chem. B 2000, 104, 9653.
(35) Holles, J. H.; Davis, R. J.; Murray, T. M.; Howe, J. M. J. Catal.
2000, 195, 193.
(
6) Trovarelli, A. Catal. ReV.sSci. Eng. 1996, 38, 439.
(
7) Kakuta, N.; Morishima, N.; Kotobuki, M.; Iwase, T.; Mizushima,
T.; Sato, Y.; Matsuura, S. Appl. Surf. Sci. 1997, 121/122, 408.