Influence of ceria and lanthana promoters on the kinetics of NO and N2O reduction by CO over alumina-supported palladium and rhodium

Article abstract of DOI:10.1006/jcat.1999.2780

Automobile exhaust catalysts are designed to reduce emissions of CO, NO_x, and uncombusted hydrocarbons. These catalysts contain noble metals, e.g., Pd, Pt, and Rh. The principal function of rhodium is to control NO_x emissions. In addition to ceria, other rare earth oxides, e.g., lanthana, are added to the alumina washcoat of automotive catalysts for thermal stabilization. Rhodium supported on lanthana has been shown to have higher activity and apparent activation energy for the NO + CO reaction than rhodium on alumina, and palladium has been identified as a possible alternative for rhodium since it is more plentiful and more durable at higher reaction temperatures. The kinetic parameters of the NO + CO and N₂O + CO reactions over alumina-, ceria/alumina-, and lanthana/alumina-supported Rh and Pd were obtained at 425-625 K. For NO + CO, the presence of ceria and lanthana affected activity, apparent activation energy, denitrogen selectivity, and reaction orders to differing degrees over Rh and Pd. The turnover frequency for the ceria and lanthana facilitated NO dissociation on Pd and Rh, but the effect of ceria on Pd was the most intense. For N₂O + CO, ceria promoted the reaction rate on Pd and Rh, while lanthana did not promote the reaction on either metal. Again, the ceria-promoted Pd catalyst was the most active. The turnover frequencies for N₂O + CO reaction were ~ 1 order of magnitude lower than those for the NO + CO reaction under comparable conditions. The rare earth components had little effect on the activation energies and reaction orders for the N₂O + CO reaction. In situ IR spectroscopy revealed the presence of isocyanate surface species on Rh/Al₂O₃ but not on Pd/Al₂O₃, indicating two different reaction mechanisms for the N₂O + CO reaction. No synergistic effect on NO_x reduction was observed for a Pd/Rh bimetallic sample.

Full text of DOI:10.1006/jcat.1999.2780

Journal of Catalysis 190, 247–260 (2000)
doi:10.1006/jcat.1999.2780, available online at http://www.idealibrary.com on
Influence of Ceria and Lanthana Promoters on the Kinetics of NO
and N2O Reduction by CO over Alumina-Supported
Palladium and Rhodium
Joseph H. Holles, Matthew A. Switzer, and Robert J. Davis1
Department of Chemical Engineering, University of Virginia, Charlottesville, Virginia 22904-4741
Received June 24, 1999; revised December 8, 1999; accepted December 8, 1999
dioxide, but previous studies of this reaction have reported
the formation of nitrous oxide as a side product (2–4). Ni-
The kinetic parameters of the NO + CO and N2O + CO reactions
over alumina-, ceria/alumina-, and lanthana/alumina-supported
Rh and Pd were determined from 425 to 625 K. For NO + CO,
the presence of ceria and lanthana affected activity, apparent acti-
vation energy, dinitrogen selectivity, and reaction orders to varying
degrees over both Pd and Rh. The turnover frequency (based on
H2 chemisorption) for the ceria-promoted Pd catalyst was at least
an order of magnitude greater than those for all other catalysts.
Lanthana-promoted Pd was the next most active catalyst. Changes
in kinetic parameters indicated that both ceria and lanthana facil-
itated NO dissociation on both Pd and Rh, but the effect of ceria
on Pd was the most profound. For N2O + CO, ceria promoted the
reaction rate on both Pd and Rh whereas lanthana did not pro-
trous oxide can also undergo further reaction with carbon
monoxide to produce the desired products dinitrogen and
carbon dioxide (2, 5). The overall reaction scheme can be
represented by the following three reactions:
1
CO + NO → N2 + CO2
[1]
2
CO + 2NO → N2O + CO2
CO + N2O → N2 + CO2.
[2]
[3]
Alumina- and silica-supported rhodium catalysts have
mote the reaction on either metal. The ceria-promoted Pd catalyst been thoroughly investigated as catalysts for NOx reduc-
was again the most active. However, the turnover frequencies for
the N2O + CO reaction were about an order of magnitude lower
than those for the NO + CO reaction under similar conditions. The
rare earth components had very little effect on the activation ener-
gies and reaction orders for the N2O + CO reaction. In situ infrared
spectroscopy showed the presence of isocyanate surface species on
Rh/Al2O3 but not on Pd/Al2O3, which suggested two different reac-
tion mechanismsfor theN2O + COreaction. Nosynergisticeffect on
c
tion (3, 4, 6, 7). The low-temperature activity of Rh/Al2O3
increased as the metal particle size increased, and the sup-
port was found to have little effect on selectivity (8). Ceria
is added to automotive exhaust catalysts as a promoter and
is postulated to enhance the water gas shift reaction and to
function as an oxygen storage component (1). Ceria affects
the NO + CO reaction on Rh by suppressing N2O forma-
tion, decreasing the apparent activation energy, and shifting
the rate dependence on NO partial pressure to positive or-
der (3). Since N2O can be an important side product, the
NOx reduction was observed for a Pd/Rh bimetallic sample.
2000
Academic Press
Key Words: NO; N2O; NOx; CO; Pd (palladium); Rh (rhodium);
lanthana; ceria; kinetics; reduction; NO + CO reaction; N2O + CO N2O + CO reaction over Rh(111) single-crystal surface and
reaction; chemisorption.
alumina-supported rhodium particles has also been inves-
tigated (5, 9).
In addition to ceria, other rare earth oxides such as lan-
thana are added to the alumina washcoat of automotive
catalysts for thermal stabilization (1). Lanthanum ions can
also be incorporated directly into the ceria promoter to in-
crease its oxygen storage capacity through extrinsic defects
(10). Pure La2O3 catalyzes the NO + CO reaction under
high-temperature conditions (11), but it is much less active
than Rh for the reaction. Interestingly, rhodium supported
on lanthana has been shown to have higher activity and
apparent activation energy for the NO + CO reaction than
Rh on alumina (7). In related work, Pd supported on lan-
thana has been examined as a catalyst for simulated engine
exhausts (12, 13).
INTRODUCTION
Automobile exhaust catalysts are designed to reduce
emissions of carbon monoxide, nitrogen oxides, and un-
combusted hydrocarbons. These catalysts typically contain
noble metals such as Pt, Pd, and Rh. The principal func-
tion of the rhodium is to control emissions of nitrogen ox-
ides (1). Removal of NOx is accomplished by reaction with
carbon monoxide to form mostly dinitrogen and carbon
1 To whom correspondence should be addressed. Fax: (804) 982-2658.
E-mail: rjd4f@virginia.edu.
247
0021-9517/00 $35.00
c
Copyright
2000 by Academic Press
All rights of reproduction in any form reserved.

248
HOLLES, SWITZER, AND DAVIS
The demand for additional rhodium in automotive ex- from Air Products, and NO (99% ) from MG Industries.
haust catalysts to meet more stringent emissions controls Gas flow rates were controlled using mass flow controllers
1
for new automobiles accounts for a large portion of the (Brooks Series 5850C) and varied from 5 to 250 ml min
.
Western-world consumption of rhodium (1). Palladium has The total pressure in the reactor was near atmospheric, and
been identified asone possible alternative for rhodium since partial pressures of the reactants were varied by chang-
it is more plentiful and has been found to be more durable ing the individual flow rates while simultaneously adjusting
at higher reaction temperatures (14, 15). As a result, the the flow rate of pure He. Reactant and product analyses
NO + CO reaction has been studied over Pd single crystals involved a combination of gas chromatography (HP 5890
(16, 17), planar model supported Pd/Al2O3/Ta(110) cata- Series II with an Alltech CTR I column) and mass spec-
lysts, and high surface area Pd/Al2O3 powder catalysts (15, trometry (Dycor MA 100 model). The standard catalyst
18–20).
pretreatment consisted of heating at approximately 523 K
Typical three-way catalysts do not promote NOx reduc- for 1 h under flowing He. Product analysis was performed
tion except in a narrow temperature window in the vicin- after a period of 45 min on stream under each set of reac-
ity of catalyst lightoff (1). This is a concern during cold tion conditions to allow the reaction to reach steady state.
start since the catalyst has not yet warmed up to operating Arrhenius plots were determined using 5.07 kPa each of
temperatures and a significant amount of total emissions NO and CO for the NO + CO reaction and 4.05 kPa each
are released during this period. Thus the present investi- of N2O and CO for the N2O + CO reaction. For the de-
gation examines the effect of ceria and lanthana on the ki- termination of reaction orders, the partial pressures of NO
netics of NOx reduction over rhodium and palladium cata- or CO, and N2O or CO, were held constant at 4.05 kPa
lysts in the relatively low temperature range of 425–625 K. while the pressure of the other gas was varied from 1.01 to
The catalysts include Pd/Al2O3, Rh/Al2O3, Pd/CeOx/Al2O3, 16.2 kPa.
Pd/La2O3/Al2O3, Rh/CeOx/Al2O3, Rh/La2O3/Al2O3, and
The percentage of the metal atoms exposed was deter-
Pd/Rh/Al2O3. This approach allows the comparison of mined using H2 chemisorption (99.999% from BOC gases
results between the different metals as a function of through a Supelco OMI purifier) in a Coulter Omnisorp
promoter.
100CX system. Adsorption isotherms were obtained at
303 K. The hydrogen uptake was calculated from the dif-
ference between total chemisorption and reversible chemi-
sorption. Elemental analysis was performed by Galbraith
Laboratories Inc, Knoxville, TN.
EXPERIMENTAL METHODS
The Pd/Al2O3 samples were prepared by slurrying
-
The FT-IR spectra were collected using a Bio-Rad FTS-
60A in the transmission mode. The sample holder consisted
of a stainless steel cylinder with KBr windows. Entrance
and exit feedthroughs allowed for flow of gases through the
cell. The temperature was monitored with a thermocou-
ple placed in contact with a self-supporting sample pellet.
Scans were recorded from 400–4000 wavenumbers with a
resolution of 2 cm 1. Samples were heated to 698 K in flow-
ing dihydrogen to first reduce the metal and then cooled
to 573 K in flowing helium. Scans were then recorded at
573 K with reactant gases flowing (5% N2O and 5% CO in
He). After flow was halted, the cell was purged with helium,
and scans were recorded. Finally, scans were obtained after
evacuation of the cell to at least 10 mPa using a Pfeiffer
TPU-260 turbomolecular vacuum pump.
Al2O3 (Alfa Aesar, 99.97% ) and palladium(II) acetylacet-
onate (acac) (Aldrich, 99% ) in toluene for 2 h at 353 K and
drying under vacuum in a rotary evaporator followed by
24 h in air at 473 K. Samples were then calcined in flowing
1
air (BOC gases) by heating to 673 K at 0.5 K min and
then remaining at 673 K for 4 h. Subsequent reduction of
the catalyst occurred at 673 K for 2 h in flowing dihydro-
gen (99.999% from BOC gases, passed through a Matheson
Model 8371V purifier). The Rh/Al2O3 samples were pre-
pared in the same way using rhodium(III) acetylacetonate
(Aldrich, 97% ). The palladium/rhodium bimetallic cata-
lyst was prepared using an appropriate mixture of Pd(acac)
and Rh(acac). For catalysts supported on CeOx/Al2O3 and
La2O3/Al2O3, the promoter was first deposited on the alu-
mina by slurrying cerium(III) acetylacetonate hydrate or
lanthanum(III) acetylacetonate hydrate (Aldrich) and alu-
mina in toluene for 2 h at 353 K and then drying and calcin-
ing as described above. The metal (palladium or rhodium)
was then deposited on the CeOx/Al2O3 or La2O3/Al2O3
RESULTS
Results from elemental analysis are presented in Table 1.
using the same procedure as that used for the alumina- The loadings of rhodium were typically lower than those
supported samples. of palladium due to the geometric form of the precursor.
The reaction experiments were performed in a quartz re- The Pd(acac)2 molecule is essentially planar which allows
actor containing approximately 50 mg of catalyst diluted in easy contact between the central Pd atom and the sub-
chromatographic silica gel (Fisher) supported on a quartz strate. However, Rh(acac)3 is three-dimensional and re-
frit. Feed gases included 5.00% NO/5.07% CO/He, N2O quiresbreakdown ofthe molecule duringcalcination before
(99.97% ), and He (99.999% ) from BOC Gases, CO (UHP) the Rh atom can efficiently contact the substrate (21). As

NOx REDUCTION ON CeO2- AND La2O3-PROMOTED Pd AND Rh
249
TABLE 1
with CeO2 and La2O3 indicated that the promoters were
well dispersed on the alumina.
Elemental Analyses and Chemisorption Results
Results from H2 chemisorption on the supported met-
als are included in Table 1. The presence of ceria did
not significantly affect Rh dispersion but more than dou-
bled the Pd dispersion. Dihydrogen adsorption isotherms
for Rh/Al2O3, Pd/Al2O3, and Pd/Rh/Al2O3 are presented
in Figs. 1a–1c. The isotherms for Rh/Al2O3 are typical
for chemisorption of dihydrogen on a metal particle. The
isotherms for Pd/Al2O3 exhibit a large uptake of dihydro-
gen at about 3kPa due to the formation ofa -phase hydride
which is stable at 303 K at pressures greater than 2.3 kPa
(23). The isotherms for Pd/Rh/Al2O3 do not show the large
uptake associated with -hydride formation in Pd particles.
This result likely indicates that the Pd/Rh/Al2O3 catalyst
contains bimetallic particles instead of separate particles of
pure Rh and Pd on the support. Another possible explana-
Wt% cerium
or lanthanum
Catalyst
Pd/Al2O3
Pd/CeOx/Al2O3
Pd/La2O3/Al2O3
Rh/Al2O3
Rh/CeOx/Al2O3
Rh/La2O3/Al2O3
Pd/Rh/Al2O3
Wt% metala
H/Mb
4.92
4.12
4.59
1.85
4.46
—
0.18
0.46
0.10
0.40
0.36
0.15
0.18
10.66
14.79
—
10.57
14.55
—
3.65
1.85c/1.41d
a Quantitative analysis by Galbraith Laboratories Inc.
b Total chemisorption–reversible chemisorption at 303 K.
c Weight percent palladium.
d Weight percent rhodium.
a result, an appreciable amount of Rh(acac)3 was removed tion that we consider unlikely due to the catalyst synthe-
during the calcination process.
sis conditions is that separate particles of Pd are too small
The loadings of cerium and lanthanum were on the order to have significant internal volume to form the -phase
of that needed for a theoretical monolayer ( 9 wt% Ce, hydride.
based on the Ce surface density of 4.83 Ce/nm2 derived
from the CeO2(110) plane and 26 wt% La from Ref. (22)). were N2, N2O, and CO2. No NO2 was detected in any
For the NO + CO reaction, the only products observed
Very small and broad X-ray diffraction peaks associated reaction run. For the N2O + CO reaction, the products
FIG. 1. Dihydrogen chemisorption isotherms for (a) Rh/Al2O3, (b) Pd/Al2O3, and (c) Pd/Rh/Al2O3. Lines indicate extrapolated values of total and
reversible dihydrogen uptake.

250
HOLLES, SWITZER, AND DAVIS
FIG. 2. Plot of conversion for the (a) NO + CO and (b) N2O + CO reactions as a function of space time (catalyst volume/volumetric flow rate) at
573 and 673 K, respectively, 4.05 kPa NO or N2O, and 4.05 kPa CO for Pd/CeOx/Al2O3. Typical conversion was 3–6% for all catalysts.
were N2 and CO2, and for the N2O decomposition reac- of 553 K yields a ratio of turnover frequencies of 1 : 2.9 : 31.8
tion, the only products were N2 and O2. Figures 2a and for the N2O + CO, N2O decomposition, and the NO + CO
2b show CO conversion for the NO + CO reaction and reactions, respectively. Figure 3 also shows that the three re-
N2O conversion for N2O + CO as a function of reactor actions have different apparent activation energies (109 kJ
space time (volume catalyst/volumetric flow rate) at 573 mol 1 for the N2O + CO reaction, 139 kJ mol 1 for the N2O
and 673K, respectively, over Pd/CeOx/Al2O3. Reactant flow decomposition reaction, and 86.6kJ mol 1 for the NO + CO
rates were varied to change space time. Linear conversion reaction).
(up to 6% ) with space time indicates differential behav-
Figures 4a and 4b are Arrhenius-type plots for the
ior and the absence of significant external transport ef- NO + CO reaction on Rh and Pd catalysts, respectively.
fects on the measured rates. Evaluation of the Weisz–Prater Apparent activation energies derived from these plots are
criterion indicates that no pore diffusion limitations exist. presented in Table 2. Comparing the observed rates at
Conversions for these experiments were kept in the 3–6% 526 K gives relative turnover frequencies of 1, 3.2, 5.6, 7.5,
range.
7.8, 34.8, and 235 for Pd/Rh/Al2O3, Pd/Al2O3, Rh/Al2O3,
Rates for the three different reactions (NO + CO, N2O + Rh/CeOx/Al2O3, Rh/La2O3/Al2O3, Pd/La2O3/Al2O3, and
CO, and N2O decomposition) are compared in Fig. 3, Pd/CeOx/Al2O3, respectively, for the NO + CO reaction.
which shows Arrhenius-type plots for each reaction over Clearly the ceria-promoted Pd catalyst is much more active
Rh/Al2O3. Comparing the curves at a common temperature than any of the other catalysts. Ceria also promoted the Rh
catalyst while lanthana promoted only the Pd catalyst. For
this reaction, the nitrogen-containing products are N2 and
N2O. Figures 5a and 5b compare the dinitrogen selectivity
for the Rh and Pd catalysts, respectively, as a function of
temperature. Dinitrogen selectivity is defined as the rate of
N2 production divided by the rate of N2 and N2O produc-
tion. For both the Pd and Rh catalysts, the presence of ceria
decreased the dinitrogen selectivity. In fact, the very active
Pd/CeOx/Al2O3 catalyst had the lowest N2 selectivity (22% ,
independent of temperature). Lanthana had a minimal ef-
fect on the N2 selectivity over Rh and a slightly deleterious
effect over Pd.
The influence of reactant partial pressures on the rate
of the NO + CO reaction at 483 K is summarized in Fig. 6
for Rh/CeOx/Al2O3. The plot indicates a negative reaction
order for CO partial pressure and a positive order for NO
partial pressure. The lines represent a linear regression of
the rate data which yield orders of 1.10 0.15 for NO and
0.29 0.15 for CO (based on NO consumption). Results
FIG. 3. Arrhenius-type plots for NO + CO, N2O + CO, and N2O de-
composition reactions over Rh/Al2O3. Conditions were 5.07 kPa each of
NO and CO for the CO + NO reaction, 4.05 kPa each of N2O and CO
for the N2O + CO reaction, and 4.05 kPa N2O for the N2O decomposition
reaction.
for the other catalysts are shown in Table 3. Reaction orders

NOx REDUCTION ON CeO2- AND La2O3-PROMOTED Pd AND Rh
251
TABLE 2
Apparent Activation Energies for NO + CO, N2O + CO, and N2O Decomposition Reactions
1
Apparent activation energy (kJ mol
)
NO + CO
Literature value
N2O + CO
N2O decomp,
This work
Catalyst
Pd/Al2O3
This work
59.8
This work
104
Literature value
154 (15), 95.4 (31),
88.2 (32), 83.6 (33),
87.9 (16), 67.8 (18)
—
—
Pd/CeOx/Al2O3
Pd/La2O3/Al2O3
Rh/Al2O3
Rh/CeOx/Al2O3
Rh/La2O3/Al2O3
Pd/Rh/Al2O3
76.6
100
86.6
49.4
150
—
—
101
98.4
109
139
107
129
—
—
—
—
139
157
229
118
117 (3), 101 (7)
75.3 (3)
126 (7)
—
167 (5), 167 (34)
—
—
—
100
Note. Numbers in parentheses refer to references.
TABLE 3
Orders of Reaction for NO and CO
Reaction order
NO consumption
NO
Literature value
CO
CO consumption
Catalyst
Pd/Al2O3
Pd/CeOx/Al2O3
Pd/La2O3/Al2O3
Rh/Al2O3
Rh/CeOx/Al2O3
Rh/La2O3/Al2O3
Pd/Rh/Al2O3
T (K)
This work
1.02
This work
Literature value
NO
CO
553
463
513
533
483
543
573
0.50 (15)
1.11
0.50
0.85 (15)
—
—
0.86
0.38
—
0.88
0.41
—
—
—
0.37 (7), 0.14 (3)
0.30 (3)
0.03
1.10
0.62
0.02
0.12
0.29
0.01
0.03 (7), 0.11 (3)
0.19 (3)
0.04 (7)
—
0.02
1.05
—
0.04
0.10
—
0.16 (7)
—
0.10
Note. indicates a maximum in the reaction order plot. Numbers in parentheses refer to references.
FIG. 4. Arrhenius-type plots for the NO + CO reaction over (a) Rh/Al2O3, Rh/CeOxAl2O3, Rh/La2O3/Al2O3, and Pd/Rh/Al2O3 and (b) Pd/Al2O3,
Pd/CeOx/Al2O3, Pd/La2O3/Al2O3, and Pd/Rh/Al2O3 for reactant mixture of 5.07 kPa each of NO and CO.

252
HOLLES, SWITZER, AND DAVIS
FIG. 7. Reaction order determination for the NO + CO reaction for
Pd/CeOx/Al2O3 at 463 K. The closed circles are for a fixed PNO = 4.05 kPa
with a varying PCO. The open circles are for a fixed PCO = 4.05 kPa with a
varying PNO
.
about negative one-half to positive one. In addition, the
reaction order for CO varied from about zero to negative
one. Promotion of the metals by ceria dramatically altered
the reaction orders. For instance, the presence of ceria on
Pd/Al2O3 increased the CO reaction order from 1.11 to
0.50. However, for Rh/Al2O3, ceria increased the reac-
tion order for NO from zero to 1.10 and decreased the CO
reaction order from 0.12 to 0.29. For some catalysts, the
reaction orders were positive at low partial pressures and
negative at high partial pressures, over the range of con-
ditions investigated. Figure 7 illustrates an example of this
behavior for Pd/CeOx/Al2O3. Interestingly, the maximum
in rate occurred near equimolar concentrations of NO and
CO for these samples.
Figures 8a and 8b are the Arrhenius-type plots for the
N2O + CO reaction over the Rh and Pd catalysts, respec-
tively. The ceria-promoted Pd catalyst again had the high-
est activity, but the promotional effect was not as dra-
matic as that seen for the NO + CO reaction. The relative
turnover frequencies are 1, 1.3, 1.9, 7.8, 13.8, 14.2, and 42.2
for Pd/Rh/Al2O3, Rh/La2O3/Al2O3, Rh/Al2O3, Pd/La2O3/
Al2O3, Rh/CeOx/Al2O3, Pd/Al2O3, and Pd/CeOx/Al2O3,
respectively, at 555 K. The presence of lanthana slightly
decreased the reaction rate on both Pd and Rh. The N2O
molecule can also decompose in the absence of CO over
Rh catalysts. The relative turnover frequencies for N2O
decomposition without CO were 1, 2.6, 5.4, and 42.2 for
Pd/Rh/Al2O3, Rh/La2O3/Al2O3, Rh/Al2O3, and Rh/CeOx/
Al2O3, respectively, at 573 K. Figure 9 shows the Arrhenius-
type plots for this reaction. The decomposition of N2O
under these reaction conditions in the absence of CO did
not occur to any appreciable extent on Pd/Al2O3, Pd/CeOx/
Al2O3, and Pd/La2O3/Al2O3 over the 513–573 K tempera-
ture range. It was necessary to increase the temperature to
about 673 K before N2O decomposed on those catalysts.
FIG. 5. N2 selectivity vs temperature for the NO + CO reaction over
(a) Rh/Al2O3, Rh/CeOxAl2O3, Rh/La2O3/Al2O3, and Pd/Rh/Al2O3 and
(b) Pd/Al2O3, Pd/CeOx/Al2O3, Pd/La2O3/Al2O3, and Pd/Rh/Al2O3.
in Table 3 are shown for both NO and CO consumption.
The reaction orders for the different reactants vary slightly
due to different N2 selectivities for the catalysts. Depend-
ing on the catalyst, the reaction order for NO varied from
FIG. 6. Reaction order determination for the NO + CO reaction for
Rh/CeOx/Al2O3 at 483 K. The open circles are for a fixed PNO = 4.05 kPa
with a varying PCO. The closed circles are for a fixed PCO = 4.05 kPa with
a varying PNO
.

NOx REDUCTION ON CeO2- AND La2O3-PROMOTED Pd AND Rh
253
FIG. 10. Reaction order determination for the N2O + CO reaction for
Rh/CeOx/Al2O3 at 543 K. The open circles are for a fixed PN = 4.05 kPa
O
2
with a varying PCO. The closed circles are for a fixed PCO = 4.05 kPa with
a varying PN
.
O
2
in Fig. 10 for Rh/CeOx/Al2O3. The plot indicates a neg-
ative reaction order for CO partial pressure and a posi-
tive order for N2O partial pressure. The lines represent a
linear regression of the rate data which yielded orders of
0.68 0.15 for CO and 0.86 0.15 for N2O. Results for
the other catalysts are shown in Table 4. Reaction orders
for N2O were positive and similar for all catalysts. Reaction
orders for CO were negative for all catalysts, but the Pd
catalysts were less inhibited by CO than the Rh catalysts.
Ceria had little effect on the reaction order for either reac-
tant whereas lanthana slightly decreased the N2O reaction
order on Pd and slightly increased the CO reaction order
on Rh.
FIG. 8. Arrhenius-type plots for the N2O + CO reaction over (a)
Rh/Al2O3, Rh/CeOxAl2O3, Rh/La2O3/Al2O3, and Pd/Rh/Al2O3 and (b)
Pd/Al2O3, Pd/CeOx/Al2O3, Pd/La2O3/Al2O3, and Pd/Rh/Al2O3 for reac-
tant mixture of 4.05 kPa each of N2O and CO.
Infrared spectroscopy was performed on catalysts to
identify adsorbed species for the N2O + CO reaction. The
Pd and Rh catalysts were exposed to N2O, NO, and CO
individually to determine the adsorption bands for each of
these adsorbates. The reaction of N2O + CO was then per-
formed on the catalysts at 573 K in the IR cell. Spectra for
Rh/Al2O3, Rh/CeOx/Al2O3, Pd/Al2O3, and Pd/CeOx/Al2O3
in vacuum, after N2O + CO reaction, are shown in Fig. 11.
For Pd/Al2O3, the major absorption band in the 1700–
Apparent activation energies for these two reactions
(N2O + CO, N2O decomposition) are presented in Table 2.
The influence of the reactant partial pressures on the
rate of the N2O + CO reaction at 543 K is summarized
1
1
2500-cm range is a small band at 1907 cm and is at-
tributed to bridged carbonyl (24). A similar bridged car-
1
bonyl species at 1887 cm was observed for Pd/CeOx/
Al2O3 (24). For Rh/Al2O3, a strong band at 2024 cm 1 indi-
cates either a dicarbonyl or a triply bonded carbonyl (25).
In contrast to the species observed on the Pd catalysts,
1
there were also bands at 1895 cm for either a cationic
NO or bridged carbonyl species (25), and a strong band at
2251 cm 1 attributed to isocyanate (–NCO) on the alumina
support (26–28). Finally, for Rh/CeOx/Al2O3, the spectrum
is simpler with a bridged carbonyl at 1855 cm 1, and a dicar-
bonyl or tricarbonyl species at 2005 cm 1 (exact assignment
ofthispeak isunclear;however, it isclearlya CO peak based
on reference spectra).
FIG. 9. Arrhenius-type plots for the N2O decomposition reaction
over Rh/Al2O3, Rh/CeOx/Al2O3, Rh/La2O3/Al2O3, and Pd/Rh/Al2O3 for
4.05 kPa N2O.

254
HOLLES, SWITZER, AND DAVIS
TABLE 4
Orders of Reaction for N2O and CO
Reaction order
N2O
Literature value
CO
Literature value
Catalyst
Pd/Al2O3
Pd/CeOx/Al2O3
Pd/La2O3/Al2O3
Rh/Al2O3
Rh/CeOx/Al2O3
Rh/La2O3/Al2O3
Pd/Rh/Al2O3
T (K)
This work
This work
543
553
583
593
543
643
613
0.69
0.62
0.46
0.72
0.86
0.76
0.80
—
—
—
0.32
0.31
0.35
0.86
0.68
0.48
0.37
—
—
—
0.65 (5), 1.14 (34)
1.0 (5), 1.18 (34)
—
—
—
—
—
—
Note. Numbers in parentheses refer to references.
DISCUSSION
a much smaller increase in the turnover frequency, about
50% . For lanthana, the specific reaction rate increased by
1 order of magnitude on supported Pd, yet the rate on sup-
ported Rh was barely affected at our reaction conditions.
The promotional effect of the ceria (and lanthana) can-
not be attributed to the increased metal dispersion since
activity for both Rh and Pd has been shown previously
to decrease with increasing dispersion (8, 15). Although
alumina-supported ceria catalyzed the NO + CO reaction
with no metal present, the reaction rate, on a per gram ba-
sis, was a factor of 4 lower than that with any other cata-
lyst studied in this work. In addition, reaction rates on pure
lanthana/alumina were an order of magnitude below those
on ceria/alumina. Therefore, the enhancement in observed
rates is due to either a modification of the metal particles
or creation of new sites at the metal/promoter interface.
The promotional effect of ceria (and lanthana) cannot
be explained simply by the additional promoter surface
area.
NO + CO Reaction
The kinetic results clearly showed that the presence of
ceria increases the reaction rate, decreases the dinitrogen
selectivity, and alters the NO + CO reaction orders over
both Pd and Rh. Even though the reaction orders over Pd
and Rh were affected by the presence of lanthana, the spe-
cific activity was promoted only in the case of Pd. The im-
plications of these results are discussed below and a kinetic
model for the reaction is developed.
The turnover frequency for the ceria-promoted Pd was
approximately 2 orders of magnitude greater than that for
the unpromoted Pd. Promotion of Rh with ceria produced
The effects of ceria and lanthana on apparent activation
energy, N2 selectivity, and reaction order for Rh were sim-
ilar to those previously reported in the literature (3, 7, 29,
30). Literature valuesfor apparent activation energyand re-
action order are shown in Tables 2 and 3, respectively (16,
31–33). The only difference is that Oh found similar N2 se-
lectivity over both Rh/Al2O3 and Rh/CeOx/Al2O3 whereas
we observed a decrease in selectivity with addition of ceria.
For unpromoted Pd, apparent activation energy, N2 selec-
tivity, and reaction order were similar to those reported
by Rainer et al. (15, 18). The effects of ceria and lanthana
promoters on N2 selectivity and reaction order for Pd were
similar;in contrast, the apparent activation energyon Pd de-
creased with ceria promotion and increased with lanthana
promotion.
Proposed reaction sequence. For the reaction of NO +
CO, our experimental results and previous works (3, 4, 6,
FIG. 11. Infrared spectra of Pd/Al2O3, Rh/Al2O3, Pd/CeOx/Al2O3,
and Rh/CeOx/Al2O3 under vacuum at 573 K after the N2O + CO reaction. 18, 20, 29, 30, 34) suggest the reaction can be represented

NOx REDUCTION ON CeO2- AND La2O3-PROMOTED Pd AND Rh
255
by the following set of elementary steps:
the rate-determining step on Pd but instead the removal of
thermally stable atomic nitrogen species is more important
(15). The results in Table 3 indicate all of the above cases.
This variety of reaction orders indicates that different ele-
mentary reactions may be rate determining under different
conditions. Thus, we develop a rate expression without as-
suming a single rate-determining step. The overall rate of
NO consumption is given by
CO + S ⇔ COa
NO + S ⇔ NOa
NOa + S → Na + Oa
NOa + Na → N2O + 2S
Na + Na → N2 + 2S
COa + Oa → CO2 + 2S,
[4]
[5]
[6]
[7]
[8]
[9]
rNO = 2rN O + 2rN
,
2
[11]
2
where “S” denotes a surface site and the subscript “a” de-
notes an adsorbed species.
An additional elementary reaction step,
where rN and rN O denote the rates of N2 and N2O forma-
2
2
tion depicted in reactions [7] and [8]. The rates rN and rN O
2
2
can be expressed in terms of the fractional surface cover-
ages ( ) of the adsorbed species as
NOa + Na → N2 + Oa + S,
[10]
2
hasbeen proposed previouslyto provide an alternative path
for N2 formation at low temperatures (6). However, Belton
et al. have shown that NOa and Na react on Rh to form
only N2O and that N2 formation at low temperature occurs
only by N atom recombination (34). For Pd, Rainer et al.
(18) and Almusaiteer and Chuang (20) have also proposed
mechanisms without this alternative N2 formation pathway.
Therefore, we have not included this step in our mechanism.
The development of a kinetic model assumes that the
presence of ceria or lanthana did not change the basic re-
action mechanism, but instead altered the rates of the indi-
vidual elementary steps. We have shown that both alumina-
and ceria/alumina-supported Pd and Rh are also capable of
further reducing N2O with CO to N2 and CO2. However,
under our conditions, the rate of the N2O + CO reaction
was at least 1 order of magnitude lower than the rate of the
NO + CO reaction, and the actual amount of N2O produced
was small due to the low levels of conversion. Additionally,
experiments with isotopically labeled N2O over Rh(111)
(35) and varying concentrations of N2O in the feed over
Pt10Rh90(111) (36) have not provided any evidence of N2O
readsorption and reaction to form N2. Consequently, steps
involving readsorption and surface reaction of N2O have
not been incorporated into the proposed mechanism.
A kinetic analysis of the NO + CO reaction over silica-
supported Rh by Hecker and Bell (4) indicates that when
the dissociation of NO (reaction [6]) is the rate-determining
step, the catalyst is predominately covered with molecularly
adsorbed NO. Thus, a negative first-order dependence on
NO partial pressure is anticipated. After studying Rh sup-
ported on alumina and ceria/alumina, Oh (3) concluded
that weakly negative or moderately positive reaction or-
ders in NO suggest that the NO dissociation reaction is not
rate determining. Additional work by Oh et al. (6) demon-
strated that when the NO dissociation rate is reasonably
rapid, the catalyst surface is largely covered with both ad-
sorbed nitrogen atoms and molecularly adsorbed NO. Fi-
rN = k8
[12]
[13]
2
N
rN O = k7
,
N
NO
2
where k7 and k8 are the rate constants for reactions [7]
and [8], respectively. Substitution of Eqs. [12] and [13] into
Eq. [11] yields
2
N
rNO = 2 k7
+ k8
.
[14]
NO
N
Equilibrium adsorption relationships are used to determine
and
:
CO
NO
= K5 PNO
= K4 PCO
[15]
[16]
NO
CO
V
V
.
The overall site balance is given by
1 =
+
+
+
,
V
[17]
NO
N
CO
where V is the fraction of vacant sites. Determination of
N
requires a nitrogen surface balance and the pseudo-steady
state hypothesis:
2
rN = k6
k7
NO N
k8 = 0
[18]
NO
V
N
Equation [18] can then be solved for
:
N
= N
[19]
N
V
with
1/2
k7 K5 PNO + [(k7 K5 PNO )2 + 4k8k6 K5 PNO
2k8
]
N =
. [20]
Substitution of [15] through [20] into [14] yields the rate
expression for NO consumption:
k7 K5 PNO N + k8 N2
rNO = 2
.
[21]
(1 + K5 PNO + K4 PCO + N)2
nally, for a positive one-half reaction order in NO, Rainer Even though the resulting expression is complex, the impor-
et al. have concluded that the dissociation of NO is not tant consequence is that the observed NO reaction order

256
HOLLES, SWITZER, AND DAVIS
can be either positive, negative, or zero depending on the Enhanced NO dissociation should decrease the steady-state
relative surface coverages of intermediates. NOa surface coverage and increase the NO reaction order,
For high surface coverage of NO, the rate will exhibit a which is observed for Rh/CeOx/Al2O3.
negative NO reaction order. However, if the surface cover-
For Pd/CeOx/Al2O3 under stoichiometric conditions, en-
age of adsorbed N atoms is high, the rate will exhibit a NO hanced NO dissociation increases the coverage of N atoms.
reaction order of about zero. Intermediate surface cover- Moreover, additional oxygen atoms from dissociated NO
ages of NO and Na or high CO coverage results in a positive contribute to higher CO oxidation rates and lower surface
NO reaction order. This model can also easily explain the coverages of CO, thus increasing the CO reaction order.
resultson Pd/CeOx/Al2O3, which indicate a maximum in the With fewer adsorbed CO molecules present, the coverage
reaction order plot, by noting that high CO coverage at low of adsorbed NO increases, which results in higher Na cover-
NO partial pressure results in a positive NO reaction order. age due to enhanced dissociation and a decreased reaction
Increasing NO surface coverage as the NO partial pressure order. The result that N2 and N2O production rates are
increases causes the reaction order to become negative.
both enhanced by the presence of the promoter is consis-
We found that the selectivity to N2 generally decreased tent with the postulate that the interface produces more N
as NO partial pressure (and thus NO surface coverage) in- atoms.
creased. Clearly, the relative surface coverage of NO and
The presence of ceria did not lead to a dramatic increase
N depends on the ability of the catalyst to dissociate NO. in activity of Rh since NO dissociation is readily accom-
The affinity of Pd and Rh to dissociate NO can be con- plished by the Rh metal itself. Even though ceria did indeed
trasted by comparing reaction energetics obtained on sin- promote the reaction over Rh, the combination of metal
1
gle crystals at low coverages. For Rh, Edis = 73.2 kJ mol
and promoter may dissociate NO too well, thus decreasing
the coverage of adsorbed NO to the point where there is
1
1
with A = 2.1 1010
s
and, for Pd, Edis = 115.7 kJ mol
with A = 4 1011
s
(34, 37). Although Edis depends on not enough adsorbed NO to react with the Na. The surface
1
NO surface coverage, in both cases NO dissociation rates is effectively blocked by adsorbed N atoms. This situation
decrease similarly as NO coverage increases. Therefore, at is consistent with our results on Rh/CeOx/Al2O3 since the
500 K, NO dissociates approximately 103 times faster on Rh NO reaction order was 1.10 and the CO reaction order was
than Pd. This high NO dissociation rate should correspond slightly negative.
to higher relative surface coverages of adsorbed N atoms on
For the lanthana-promoted catalysts, the reaction was en-
Rh. Examination of our experimental results on Rh/Al2O3 hanced on Pd and was essentially unchanged on Rh. Thus,
shows that the NO + CO reaction is essentially zero order an important interaction between the Pd and lanthana is in-
in both NO and CO, indicating that there are moderate to dicated. We propose a mechanism similar to that described
high coverages of both NO and Na. For Pd, where dissocia- above. First, NO adsorbs at the metal/oxide interface.
tion is less energetically favored, reaction orders are + 1.02 In this case, an interaction between the oxygen of NO
in NO and 1.11 in CO. The negative CO reaction order in- with oxophilic Lewis acid sites on the surface such as the
dicates that the surface coverage is dominated by adsorbed La3+ cation may be responsible for weakening the N–O
CO, which leads to a low activity for NO reactions on Pd. bond. This explanation is qualitatively the same as that pro-
Again, the similar low surface coverages of both NO and posed by Alvero et al. for enhanced CO dissociation over
Na lead to similar amounts of N2 and N2O formed, which a lanthana-promoted rhodium/alumina catalyst (38). Any
agrees with the results of 50% N2 selectivity at 550 K.
improvement in N–O bond dissociation provided by the
When ceria is added to the support, the turnover fre- support would quickly result in rapid product formation as
quency is increased for both Rh and Pd catalysts. An under- discussed earlier.
standing of the interaction among ceria, transition metal,
and adsorbed species is required to explain this increased
activity. One possible explanation of the promotional effect
N2O + CO Reaction
has been proposed by Oh (3). In his mechanism, NO ad-
sorbs near the metal/ceria interface with the nitrogen end
of the molecule attached to the metal and the oxygen end
attached to an oxygen-deficient cerium species. The concur-
rent interaction of both ends of the NO molecule with the
catalyst surface should weaken the N–O bond and facilitate
dissociation. The increased activity of the ceria-promoted
catalysts may also be partially due to the ability of ceria
to store this removed oxygen in the lattice. This explana-
tion of enhanced NO dissociation in the presence of ceria is
consistent with the measured reaction orders in this work.
For both Pd and Rh, the presence of ceria increased the
reaction rates by about 0.5 order of magnitude when com-
pared to the unpromoted catalysts. Similar to the NO + CO
reaction, ceria-promoted Pd exhibited the highest activity.
When compared to the NO + CO reaction under similar
conditions, rates for the N2O + CO reaction on the same
catalyst (Pd or Rh) were lower by 0.5 to 2 orders of magni-
tude. Additionally, for the Rh catalysts, inhibition ofthe rate
of N2O reaction by CO clearly demonstrates that CO occu-
pied a large fraction of the available surface sites and sup-
pressed N2O adsorption and decomposition. In contrast to

NOx REDUCTION ON CeO2- AND La2O3-PROMOTED Pd AND Rh
257
results with ceria, no promotional effect was observed with
Since the evidence in this work indicates that both ad-
lanthana; in fact, reaction rates on lanthana-promoted Pd sorbed nitrogen atoms and NO are present on the surface
and Rh were lower by a factor of 2. This may be attributed for Rh/Al2O3 and not on the Pd/Al2O3 surface, the poten-
to the lower metal dispersion of the lanthana-promoted tial exists for the reaction to occur by two separate mecha-
catalysts. Alumina-supported ceria (without any transition nisms over the two different metals. If adsorbed N2O were
metal) also catalyzed the N2O + CO reaction, but the re- to dissociate into NOa and Na, the high CO surface cover-
action rate on a per gram basis was lower by a factor of 2 age would allow for ready formation of isocyanate, which
than that for any other catalyst. Therefore, similar to the would then migrate to the alumina support. The high sur-
NO + CO reaction, the observed rates are due to either the face area alumina support acts as an isocyanate trap, which
metal or ceria in contact with the metal.
allows for easy interrogation by IR spectroscopy. The pres-
The presence of ceria and lanthana did not have as dra- ence of isocyanate therefore serves as a marker for the N2O
matic an effect on the kinetic parameters for the N2O + CO reaction into NOa and Na. The lack of a similar isocyanate
reaction as for the NO + CO reaction. With the exception marker on the Pd/Al2O3 catalyst indicates that this N2O de-
of ceria-promoted Rh, the promoter had very little effect composition path is not occurring to any significant extent.
on apparent activation energy. Literature values for unpro- For the ceria-supported catalysts, the lack of the isocyanate
moted Rh are given in Table 2 (5, 9). Likewise, the N2O marker for both metals is expected since high NO coverage
and CO reaction orders are in general agreement with the is required for isocyanate formation (25). The high NO cov-
literature values in Table 4. We have not been able to find erage required for isocyanate formation is not achieved due
analogous studies on Pd catalysts and promoted Rh cata- to the presence of CO. Therefore, for the ceria-promoted
lysts. It is interesting to note that ceria had little effect on catalysts, infrared spectroscopy does not provide conclusive
reaction orders but the CO reaction order was significantly evidence for either N2O decomposition mechanism. How-
different on the two metals.
ever, when the IR spectroscopy results are combined with
In an attempt to gain insight into the elementary steps reaction kinetics, we conclude that, although the two un-
involved in the N2O + CO reaction, we performed in situ promoted metals may have different reaction mechanisms
IR spectroscopy on samples previously exposed to the re- for N2O decomposition, the presence of ceria did not ap-
actant gases. As expected for weakly adsorbing N2O and preciably change the observed reaction orders and likely
strongly adsorbing CO, the majority of the observed ab- did not change the mechanism.
sorption bands were attributed to carbonate on the support
and carbonyls on the metal. No N2O absorption bands were
Proposed reaction sequence. Based on our experimen-
tal results and those in the literature (5, 9, 41), models for
observed. For the Rh/Al2O3 catalyst, the isocyanate band
N2O + CO reaction are developed below. In the mechanism
1
at 2251 cm indicates the presence of adsorbed nitrogen
given below, N2O dissociates in two successive steps to give
atoms on the surface. Interestingly, adsorption and decom-
adsorbed N atoms, which recombine to N2:
position of N2O without CO on this catalyst did not reveal
N2O(g) + S ⇔ N2Oa
N2Oa + S ⇔ Na + NOa
NOa + S → Na + Oa
2Na → N2(g) + 2S.
[22]
the presence of NO on the surface. Isocyanate (–NCO) is
formed by the reaction between an adsorbed nitrogen atom
and CO (39, 40). The presence of isocyanate on Rh/Al2O3
indicates that N2O decomposes in the presence of CO to
form adsorbed NO and N atoms. This mechanism has re-
cently been proposed by Uner to describe the kinetics of the
N2O + CO reaction over Rh surfaces (41). Previous stud-
ies have shown that for alumina-supported noble metals
(including Pd and Rh), the isocyanate species forms from
NO and CO on the metal and subsequently migrates to the
support (28, 40). Once on the support, the isocyanate is
quite stable and is not readily decomposed. These studies
have also indicated that the surface coverage of isocyanate
is greatest under conditions of high CO and low NO surface
coverages as would be the case during the N2O + CO reac-
tion. Additionally, isocyanate forms at similar rates under
similar conditionsfrom NO and CO on both Pd and Rh (28).
For ceria-supported catalysts, formation of isocyanate re-
quires different surface conditions. In that case, isocyanate
formation occurs when the surface is first covered by NO
and then exposed to CO (25).
[23]
[24]
[25]
Using the above set of elementary reactions, the overall rate
expression is (see Ref. (41), for a complete development of
the expression):
2/3
1/2
N2
2/3
N2O
diss
NO
1
CO
rN O+CO = k
k
K
Kdiss
K
P
PCO , [26]
N2O
2
N2O
diss
where k
and kN are the rate constants for Eqs. [24] and
2
NO
diss
N2O
[25], respectively, and K
is the equilibrium constant for
Eq. [23].
Similar to the NO + CO reaction, the promotional ef-
fects of the supports can be understood by contrasting the
experimental results for the different catalysts and compar-
ing results from the NO + CO reactions. For the two-step
mechanism, the rate-determining step is assumed to be the
dissociation of adsorbed NO. Comparing the two promot-
ers for this reaction, ceria increased activity while lanthana

258
HOLLES, SWITZER, AND DAVIS
did not. Since ceria produced a similar promotional effect Cunningham et al. have attributed the increased activity
for the NO + CO reaction where NO dissociation was the with ceria to sites with excess electrons at the metal/ceria
important step, a parallel explanation for the N2O + CO re- interface. This provides storage sites for the dissociated O
action can be provided. As discussed above, ceria provides atoms until another O atom is available to scavenge the oxy-
metal/interface sites for NO adsorption and dissociation. gen and complete the redox cycle (45). Again, the ceria may
Results from IR spectroscopy indicate that this mechanism also store this oxygen internally. For the N2O + CO reac-
may occur on Rh catalysts. Consistent with the NO + CO tion, the CO scavenges the stored oxygen. Based on the lack
results, lanthana demonstrated no promotional effect for ofan isocyanate marker in the IR spectroscopy, we conclude
the N2O + CO reaction, again indicating no ability to en- that the one-step mechanism is occurring on the Pd cata-
hance NO dissociation on Rh.
lysts. It should also be noted that the one-step mechanism
An alternative sequence to form N2 by N2O dissociation may also contribute to some activity on the Rh catalysts.
in one step is represented by the following set of reactions: Similar to Rh, lanthana did not promote the N2O + CO re-
action over Pd. This is a clear example of the ability of ceria
N2O(g) + S ⇔ N2Oa
[27]
[28]
[29]
[30]
to promote a reaction by virtue of its ability to easily cycle
between oxidation states. Lanthana is not capable of easily
changing oxidation states under these conditions.
N2Oa → N2(g) + Oa
CO + S ⇔ COa
COa + Oa → CO2(g) + 2S.
We also studied the N2O decomposition reaction in the
absence of CO and found that (1) N2O will not decompose
over the Pd/Al2O3, Pd/CeOx/Al2O3, and Pd/La2O3/Al2O3
catalysts at the reaction temperatures used here (513–
573 K) and (2) the presence of ceria increased the rate of
N2O decomposition while lanthana slightly decreased the
rate of N2O decomposition on Rh containing catalysts.
Consistent with our results on the inability of Pd to cata-
lyze the N2O decomposition, Isa and Saleh reported that,
after initial oxidation, Pd did not tend to adsorb further
N2O between 303 and 523 K (46). Li and Armor have
also shown that for similar conversion, N2O decomposition
on Pd/Al2O3 requires a temperature approximately 150 K
higher than that of Rh/Al2O3 (47). Our results show that a
temperature of greater than 663 K was required for a 3%
conversion of N2O. This observation agrees with that of
Redmond, who determined a rate expression for N2O de-
composition at elevated temperatures (1013–1217 K) and
demonstrated that the reaction is poisoned by adsorbed
oxygen (48). For our Pd catalysts, the fresh sample appears
to decompose enough N2O to oxidize the surface and then
[33] no further reaction of N2O occurs.
From these reactions we can see that the rate of the N2O +
CO reaction is given by
rN O+CO = kdiss
= kr
,
CO O
[31]
N2O
2
where kdiss and kr are the rate constants for Eqs. [28] and
[30], respectively. The appropriate terms for the fractional
surface coverage of N2O are developed using the equilib-
rium relationships for the adsorption of N2O and CO, an
overall site balance, and the pseudo-steady-state hypoth-
esis. Assuming that surface coverage of atomic oxygen is
small, the fractional surface coverage of N2O is
K
P
N2O N2O
=
.
[32]
N2O
1 + KN O PN O + KCO PCO
2
2
This is substituted into Eq. [31] to yield the overall reaction
rate expression
k
diss K
P
N2O N2O
rN O+CO
=
.
2
1 + KN O PN O + KCO PCO
2
2
For the Rh catalysts, the rate of consumption of N2O was
faster in the absence of CO. The N2O decomposition reac-
If the decomposition step described by [28] requires an ad-
jacent vacant site, the form of the rate expression is similar
to [33] except the denominator is squared.
The proposed N2O + CO one-step reaction mechanism
does not contain the same NO dissociation reaction as
the two-step mechansim. Therefore, a different explana-
tion of the ceria promotional effect is required. The pro-
posed mechanism requiressimple N2O adsorption with sub-
sequent dissociation to leave behind an adsorbed O for CO
oxidation. Thisissimilar to the N2O decomposition reaction
observed on Rh(111) in the absence of CO (42), Cu(111) CO reaction, the turnover frequency on Pd/Rh/Al2O3 was
(43), Ni(110) (44) and isconsistent with our IR spectroscopy lower than that on either Pd/Al2O3 or Rh/Al2O3, which
results. The important step in this mechanism is the N2O demonstrates that there is no synergistic effect when two
decomposition reaction (Eq. [28]) and it likely involves metals are combined. Interestingly, the apparent activation
the ceria promoter. For N2O decomposition without CO, energy and reaction orders on Pd/Rh/Al2O3 were similar
tion, in the absence of CO, cannot simply be ascribed to the
presence of alumina or ceria/alumina support since tem-
peratures of 753 and 713 K, respectively, were necessary to
begin N2O decomposition on these supports without tran-
sition metals. Similar to both the NO + CO and N2O + CO
reactions, the presence of ceria served to increase the rate
of N2O decomposition over Rh by approximately 1 order
of magnitude.
NOx reactions on the bimetallic catalyst. For the NO +

NOx REDUCTION ON CeO2- AND La2O3-PROMOTED Pd AND Rh
259
to those on Rh/Al2O3. For the N2O + CO reaction, the
REFERENCES
turnover frequency on Pd/Rh/Al2O3 was an order of mag-
nitude lower than that on Pd/Al2O3, but within a factor of 2
of that on Rh/Al2O3. The combined kinetic results of NOx
reactions suggest that the surface properties of the bimetal-
lic sample are dominated by Rh. The lack of formation of
a Pd -hydride phase during dihydrogen chemisorption on
Pd/Rh/Al2O3 is consistent with this hypothesis.
1. Taylor, K. C., Catal. Rev.-Sci. Eng. 35, 457 (1993).
2. Cho, B. K., Shanks, B. H., and Bailey, J. E., J. Catal. 115, 486 (1989).
3. Oh, S. H., J. Catal. 124, 477 (1990).
4. Hecker, W. C., and Bell, A. T., J. Catal. 84, 200 (1983).
5. McCabe, R. W., and Wong, C., J. Catal. 121, 422 (1990).
6. Oh, S. H., Fisher, G. B., Carpenter, J. E., and Goodman, D. W., J. Catal.
100, 360 (1986).
7. Pande, N. T., and Bell, A. T., J. Catal. 98, 7 (1986).
8. Oh, S. E., and Eickel, C. C., J. Catal. 128, 526 (1991).
9. Belton, D. N., and Schmeig, S. J., J. Catal. 138, 70 (1992).
10. Trovarelli, A., de Leitenburg, C., and Dolcettis, G., Chemtech 32
(1997).
CONCLUSIONS
This paper describes the results of a comparative kinetic
study of the NO + CO, N2O + CO, and N2O decomposi-
tion reactions over alumina-, ceria/alumina-, and lanthana/
alumina-supported Pd and Rh catalysts. Since no syner-
gistic effect was found for NOx reduction on a Pd/Rh
bimetallic sample, monometallic catalysts were the fo-
cus of this study. For the NO + CO reaction, the changes
in kinetic parameters between alumina-, ceria/alumina-,
and lanthana/alumina-supported Rh agree qualitatively
with those previously reported in the literature for appar-
ent activation energies and reaction orders. Comparable
studies on promoted Pd catalysts are unavailable. In the
current work, the ceria/alumina-supported Pd catalyst ex-
hibited activity for the NO + CO reaction an order of mag-
nitude greater than that of ceria/alumina-supported Rh.
The lanthana/alumina-supported Pd catalyst was the sec-
ond most active. Additionally, the presence of ceria and
lanthana had a large effect on the kinetic parameters over
both Rh and Pd. However, these effects were not consistent
on both catalysts. The ceria and lanthana promoters appear
to affect the NO + CO reaction by facilitating the dissocia-
tion of NO at the metal/promoter interface. This explana-
tion can account for the altered activities, reaction orders,
and dinitrogen selectivities of the promoted catalysts. In ad-
dition, the ceria/metal interface appeared to dissociate NO
more effectively than the lanthana/metal interface.
11. Huang, S.-J., Walters, A. B., and Vannice, M. A., J. Catal. 173, 229
(1998).
12. Muraki, H., Shinjoh, H., Sobukawa, H., Yokata, K., and Fujitani, Y.,
Ind. Eng. Chem., Prod. Res. Dev. 25, 202 (1986).
13. Muraki, H., Shinjoh, H., and Fujitani, Y., Appl. Catal. 22, 325 (1986).
14. Steel, M. C. F., in “Catalysis and Automotive Pollution Control II”
(A. Crucq, Ed.), Vol. 71, p. 105. Elsevier, Brussels, Belgium, 1991.
15. Rainer, D. R., Vesecky, S. M., Oh, W. S., and Goodman, D. W., J. Catal.
167, 234 (1997).
16. Vesecky, S. M., Chen, P., Xu, X., and Goodman, D. W., J. Vac. Sci.
Technol. A 13, 1539 (1995).
17. Vesecky, S. M., Rainer, D. R., and Goodman, D. W., J. Vac. Sci. Technol.
A 14, 1457 (1996).
18. Rainer, D. R., Koranne, M., Vesecky, S. M., and Goodman, D. W.,
J. Phys. Chem. B 101, 10769 (1997).
19. Almusaiteer, K., and Chuang, S. S. C., J. Catal. 180, 161 (1998).
20. Almusaiteer, K., and Chuang, S. S. C., J. Catal. 184, 189 (1999).
21. Kenvin, J. C., White, M. G., and Mitchell, M. B., Langmuir 7, 1198
(1991).
22. Shi, C., Walters, A. B., and Vannice, M. A., Appl. Catal. B 14, 175
(1997).
23. Benson, J. E., Hwang, H. S., and Boudart, M., J. Catal. 30, 146 (1973).
24. Vannice, M. A., in “CatalysisScience and Technology” (J. R. Anderson
and M. Boudart, Eds.), Vol. 3, p. 139. Springer-Verlag, Berlin, 1982.
25. Keiski, R. L., Harkonen, M., Lahti, A., Maunula, T., Savimaki, A.,
and Slotte, T., in “Catalysis and Automotive Pollution Control III”
(A. Frennet and J.-M. Bastin, Eds.), Vol. 96, p. 85. Elsevier, Amster-
dam, 1995.
26. Unland, M., J. Catal. 31, 459 (1973).
27. Solymosi, F., and Bansagi, T., J. Phys. Chem. 83, 552 (1979).
28. Solymosi, F., Volgyesi, L., and Rasko, J., Z. Phys. Chem. 79, 79
(1980).
29. Cho, B. K., J. Catal. 148, 697 (1994).
30. Peden, C. H. F., Belton, D. N., and Schmeig, S. J., J. Catal. 155, 204
(1995).
31. Butler, J. D., and Davis, D. R., J. Chem. Soc., Dalton Trans. 2249
(1976).
32. Xi, G., Bao, J., Shao, S., and Li, S., J. Vac. Sci. Technol. A 10, 2351
(1992).
33. Graham, G. W., Logan, A. D., and Shelef, M., J. Phys. Chem. 97, 5445
(1993).
34. Belton, D. N., DiMaggio, C. L., Schmieg, S. J., and Ng, K. Y. S., J. Catal.
157, 559 (1995).
Although the presence of ceria increased the rate of the
N2O + CO reaction on both Pd and Rh, it had very little
effect on apparent activation energies and reaction orders.
Lanthana had no effect on activities or activation energies.
Based on infrared spectroscopy and reaction order data, the
N2O + CO reaction occurred via different mechanisms on
Rh and Pd. The reaction on Rh appeared to involve N2O
dissociating into N and NO with subsequent dissociation
of NO. For Pd, N2O appeared to react directly to N2 and
adsorbed oxygen.
35. Permana, H., Ng, S. K. Y., Peden, C. H. F., Schmeig, S. J., Lambert,
D. K., and Belton, D. N., J. Catal. 164, 194 (1996).
36. Ng, S. K. Y., Belton, D. N., Schmeig, S. J., and Fisher, G. B., J. Catal.
146, 394 (1994).
37. Schmick, H.-D., and Wassmuth, H.-W., Surf. Sci. 123, 471 (1982).
38. Alvero, R., Bernal, A., Carrizosa, I., and Odriozola, J. A., Inorg. Chim.
Acta 140, 45 (1987).
ACKNOWLEDGMENTS
This work was supported by the Division of Chemical Sciences, Office
of Basic Energy Sciences, Office of Energy Research, U.S. Department
of Energy. Additional support was provided by the Virginia Academic
Enhancement Program.

260
HOLLES, SWITZER, AND DAVIS
45. Cunningham, J., Hickey, J. N., Cataluna, R., Conesa, J.-C., Soria, J.,
39. Novak, E., and Solymosi, F., J. Catal. 125, 112 (1990).
40. Hecker, W. C., and Bell, A. T., J. Catal. 85, 389 (1984).
41. Uner, D. O., J. Catal. 178, 382 (1998).
42. Daniel, W. M., Kim, Y., Peebles, H. C., and White, J. M., Surf. Sci. 111,
189 (1981).
and Martinez-Arias, A., in “Proceedings, 11th Internation Congress
on Catalysis—40th Anniversary, Baltimore, 1996” (J. W. Hightower,
W. N. Delgas, E. Iglesia, and A. T. Bell, Eds.), Vol. 101, p. 681. Elsevier,
Amsterdam, 1996.
43. Habraken, F. H. P. M., Hieffer, E. Ph., and Bootsma, G. A., Surf. Sci. 46. Isa, S. A., and Saleh, J. M., J. Phys. Chem. 76, 2530 (1972).
83, 45 (1979).
47. Li, Y., and Armor, J. N., Appl. Catal. B 1, L21 (1992).
48. Redmond, J. P., J. Catal. 7, 297 (1967).
44. Hoffman, D. A., and Hudson, J. B., Surf. Sci. 180, 77 (1987).