CO + O2 and CO + NO reactions over Pd/Al2O3 catalysts

Article abstract of DOI:10.1021/jp971262z

The kinetics of the CO + NO and CO + O₂ reactions have been studied over several different Pd/Al₂O₃ powder catalysts covering a wide range of average Pd particle sizes. The structure-insensitive nature of the CO + O₂ reaction over Pd has been exploited to determine the relative dispersions in several Pd/Al₂O₃ powder catalysts by measuring the rate of that reaction and normalizing against surface area. This method, assuming particles of hemispherical shape, yields average particle sizes that are consistent with observations using TEM. For the CO + NO reaction over the same catalysts, a pronounced structure sensitivity is evident that results in higher reaction activities over larger Pd particles. Differences between CO oxidation rates for the CO + O₂ reaction prior and subsequent to exposure to the CO + NO reaction are suggestive of the formation of an inhibiting, site-blocking species during the latter reaction. These results are discussed with reference to surface science and kinetics studies over single crystal and model planar Al₂O₃ supported Pd catalysts which have indicated that the structure-sensitive formation and stabilization of an inactive atomic N species plays a significant role in determining the reaction activity over a particular crystal plane or particle size.

Full text of DOI:10.1021/jp971262z

J. Phys. Chem. B 1997, 101, 10769-10774
10769
CO + O2 and CO + NO Reactions over Pd/Al2O3 Catalysts
D. R. Rainer,† M. Koranne,‡ S. M. Vesecky,§ and D. W. Goodman*,†
Department of Chemistry, Texas A&M UniVersity, College Station, Texas 77843-3255, W. R. Grace and Co.,
Washington Research Center, Columbia, Maryland 21044, and Monsanto Company, 800 N. Lindbergh BlVd.,
St. Louis, Missouri 63167
ReceiVed: April 11, 1997; In Final Form: October 6, 1997X
The kinetics of the CO + NO and CO + O2 reactions have been studied over several different Pd/Al2O3
powder catalysts covering a wide range of average Pd particle sizes. The structure-insensitive nature of the
CO + O2 reaction over Pd has been exploited to determine the relative dispersions in several Pd/Al2O3 powder
catalysts by measuring the rate of that reaction and normalizing against surface area. This method, assuming
particles of hemispherical shape, yields average particle sizes that are consistent with observations using
TEM. For the CO + NO reaction over the same catalysts, a pronounced structure sensitivity is evident that
results in higher reaction activities over larger Pd particles. Differences between CO oxidation rates for the
CO + O2 reaction prior and subsequent to exposure to the CO + NO reaction are suggestive of the formation
of an inhibiting, site-blocking species during the latter reaction. These results are discussed with reference
to surface science and kinetics studies over single crystal and model planar Al2O3 supported Pd catalysts
which have indicated that the structure-sensitive formation and stabilization of an inactive atomic N species
plays a significant role in determining the reaction activity over a particular crystal plane or particle size.
1. Introduction
more strongly on the surface than molecular O2, the surface of
the catalyst is predominantly populated with adsorbed CO
Due to the important implications for automobile emissions
control, the CO + O2
under reaction conditions, which inhibits the dissociative ad-
sorption of oxygen. Thus, the rate-limiting step under typical
conditions is believed to be the desorption of CO. This is
consistent with the negative first-order dependence on CO
concentration and positive first-order dependence on O2 con-
centration generally observed for the reaction except at ratios
far from stoichiometric.2,6-8
Another important aspect of the CO/O2 reaction is its structure
insensitivity at and near stoichiometric reaction conditions;
excellent agreement in the rate of the reaction over Pd single
crystal and supported Pd/SiO2 catalysts has been observed,1 as
well as over a range of particle sizes in supported catalysts.2
This particular characteristic of the reaction was very significant
for the CO + NO studies over Pd/Al2O3 powder catalysts
described in this work, as it allowed the various powder catalyst
dispersions to be estimated using the CO oxidation rate as a
measure of Pd surface area.
1-9
and CO + NO10-19 reactions have
been extensively studied over a variety of transition and noble
metals. In particular, Rh, Pt, and Pd have received a great deal
of attention due to their successes as three-way catalysts (for
the simultaneous reduction of NO and the oxidation of CO and
unburnt hydrocarbons) in commercial automotive catalytic
converters. There has been a substantial effort to elucidate the
basic reaction pathways and catalytic characteristics of these
systems to facilitate their improved design. While Rh has
typically been perceived as the critical component in these
multimetal catalysts, due largely to its enhanced selectivity for
the complete reduction of NO,20 there has recently been an
impetus to develop a Pd only catalyst suitable for use in
commercial converters. Several factors account for this.
Continuous improvement in gasoline formulation has margin-
alized the negative impact of the inferior resistance to Pb and
S poisoning displayed by Pd relative to the other metals.
Additionally, Pd has demonstrated excellent characteristics for
hydrocarbon oxidation21 and a resistance to sintering at elevated
temperatures. Factoring in the greater availability (and hence
cheaper cost) of Pd relative to Rh20 makes the desire for a Pd
only catalyst very understandable. Here we examine the
catalytic properties for the CO + NO reaction over several Pd-
only catalysts and show how the parallel study of the CO + O2
reaction over the same catalysts complement these efforts.
The CO oxidation reaction has generated considerable interest,
not only due to its importance in terms of emissions control
but also because its relative simplicity makes it an ideal, tractable
heterogeneous catalysis problem. The reaction mechanism over
noble metals has been described by a Langmuir-Hinshelwood
model where molecularly adsorbed CO reacts with atomic
oxygen on the surface, desorbing CO2. Because CO adsorbs
The following mechanism for the CO + NO reaction has
been proposed for Rh:12
CO(g) a CO(a)
NO(g) a NO(a)
NO(a) a N(a) + O(a)
NO(a) + N(a) a N2O(g)
N(a) + N(a) a N2(g)
CO(a) + O(a) a CO2(g)
The fast step in this process is the scavenging of atomic O(a) by
CO(a) to desorb CO2.
* To whom correspondence should be addressed.
† Texas A & M University.
The present study of the CO + NO reaction over Pd/Al2O3
catalysts reveals a pronounced particle size dependence com-
parable to that observed over supported Rh,10 with larger
particles displaying significantly enhanced specific activities
‡ W. R. Grace and Co.
§ Monsanto Co.
X Abstract published in AdVance ACS Abstracts, December 1, 1997.
S1089-5647(97)01262-5 CCC: $14.00 © 1997 American Chemical Society

10770 J. Phys. Chem. B, Vol. 101, No. 50, 1997
Rainer et al.
Figure 1. TEM images of the four different Pd/Al2O3 powder catalysts used in the kinetics studies. The R-Al2O3 is a low surface area support (∼5
m2/g) relative to the γ-Al2O3 (∼155 m2/g), resulting in much larger particles at a similar Pd loading.
relative to smaller ones. This structure dependence is discussed
in relation to UHV experiments over planar model supported
Pd/Al2O3/Ta(110) catalysts that indicate an enhanced activity
for atomic nitrogen formation and stabilization on smaller Pd
particles. Additionally, a brief discussion of the CO + O2
reaction over the powder Pd/Al2O3 catalysts is presented that
emphasizes the utility of this structure-insensitive reaction for
dispersion measurement.
2. Experimental Section
The CO + O2 and CO + NO reactions over Pd/Al2O3
powders were performed in a conventional atmospheric flow

CO + O2 and CO + NO Reactions
J. Phys. Chem. B, Vol. 101, No. 50, 1997 10771
reactor. The microreactor itself consists of a vertical, hollow
quartz tube ∼15 cm long, with an inner diameter of ∼5 mm
along half the length of the inlet side, and tapering down to ∼1
mm, housed in a ceramic heating furnace. The catalyst resides
on a plug of glass wool packed in at the constriction point. This
design accelerates the removal of the product gas eluting from
the catalyst bed, minimizing the exposure time to the heated
region after the catalytic reaction has occurred. The temperature
was monitored by a type “K” thermocouple junction residing
just above (∼1 mm) the catalyst bed, coupled to the electronic
furnace control. The flow rates of the reactant gases were
controlled by Brooks Model 5850 mass flow controllers.
Mixtures of 5% CO, NO, and O2 diluted in He (Matheson,
99.995%) were passed through the quartz microreactor at typical
flow rates of 68 cm3/min. The product gases were analyzed
using a Varian Model 3400 gas chromatograph with a column
Figure 2. Arrhenius plot for the CO + O2 reaction over the 0.1%
1
from HayeSep (Model DB, /8 in. × 30 ft).
Pd/γ-Al2O catalyst.
3
Four different Pd/Al2O3 powder catalysts were prepared using
a standard incipient wetness impregnation technique from a Pd
nitrosyl nitrate solution. Three different loadings of Pd were
added to a γ-Al2O3 powder (surface area ∼155 m2/g) in weight
percentages of 0.1, 1.0, and 5.0%. These amounts were chosen
to obtain three catalysts exhibiting distinctly different average
particle sizes for kinetics comparisons for the CO + NO
reaction. Additionally, a low surface area R-Al2O3 powder
(surface area ∼5 m2/g) was loaded with 2% Pd in order to ob-
tain a very low-dispersed catalyst with a large average particle
size. The catalyst was dried in air for ∼16 h at 80 °C, ground
into a powder, and reduced in flowing pure hydrogen for 4 h at
700 K.
TABLE 1: CO + O2 Reaction Rate Data and Dispersion
Estimates
dispersion approx avg
rate
at 450
based on
CO + O2
particle size
based on
Ea
K (TOF) rate (7%) dispersion (Å) (kcal/mol)
a
a
a
a
0.1% Pd/γ-Al2O3
1.0% Pd/γ-Al2O3
5.0% Pd/γ-Al2O3
0.013
e
e
100 f
19.2
8.9
<30g
60
125
30.1
29.3
31.3
30.7
24.6
26.0-33.1
27.0
2.0% Pd/R-Al2O3
e
0.9
1200
b
5.0% Pd/SiO2 [3]
0.031
0.03
0.04
Pd(110)c[1]
Pd(100)d[9]
a 4.4 Torr of CO, 2.3 Torr of O2. b 9.8 Torr of CO, 4.9 Torr of O2.
c 16 Torr of CO, 8 Torr of O2. d 1 Torr of CO, 0.5 Torr of O2. e Rates
normalized against 0.1% Pd/γ-Al2O3 catalyst. f Assumed. g From TEM
images.
The micrograph images of the catalyst were acquired using
a JEOL JEM 2010 model transmission electron microscope
(TEM) at a beam voltage of 2 kV, achieving a maximum
magnification of 400 000×.
The planar model supported catalysts used in the temperature
programmed desorption (TPD) experiments were obtained by
vapor depositing Pd onto a previously prepared Al2O3 thin film
supported on a Ta(110) substrate in ultrahigh vacuum (UHV).
The UHV system was of a type previously described in the
literature,22,23 equipped for Auger electron spectroscopy (AES),
infrared reflection-absorption spectroscopy (IRAS), and TPD.
The Pd was evaporated from a source comprised of a high-
purity Pd wire tightly wrapped around a resistively heated Ta
filament. The preparation of the Al2O3 thin film has been
described previously.24,25 The sample temperature was moni-
tored using a type “C” thermocouple junction spot-welded to
the edge of the single-crystal substrate, and the TPD experiments
were carried out with a linear heating rate of 5 K/s.
dispersion. Since the reaction has been shown to be structure
insensitive,1,2 the turnover frequencies (molecules reacted per
site per second), TOF, were assumed to be equal for each
catalyst. The rates of the remaining three catalysts were then
normalized against the 0.1% catalyst to yield the dispersion of
each. This approach to metal surface area measurement for
supported Rh catalysts has yielded results consistent with those
acquired by H2 chemisorption titration techniques.10
There is some degree of error inherent in this approach; based
on a simple model for spherical (or hemispherical) particles,
the dispersion for 20 Å Pd particles is approximately 50%. If
all the particles in the 0.1% catalyst were 20 Å in diameter, the
dispersion assumption would be off by only a factor of 2. This
should be the upper limit of the error introduced by this
assumption since no particles larger than ∼20 Å were imaged.
Furthermore, no error is introduced into the relative dispersions
between the four catalysts since they are all based on the same
assumption of 100% dispersion for the lightly loaded catalyst.
The average particle sizes as determined from this method were
compared with TEM micrographs of each catalyst. While it is
difficult to get a statistically meaningful size distribution from
the TEM data due to the large range of particle sizes involved,
the two methods appear to be qualitatively consistent.
3. Results and Discussion
3.1. CO + O2 Reaction. The CO oxidation reaction was
performed on each of the four different Pd/Al2O3 powder
catalysts in the flow reaction mode. The primary purpose of
these experiments was to determine the Pd dispersion of each
catalyst for reference in the CO + NO reaction study. Because
the reaction is structure insensitive, the total surface area of the
supported Pd will be proportional to the rate of the reaction.
Each catalyst was subjected to TEM study to determine the
approximate size distribution. Figure 1 contains a representative
TEM image from each catalyst. For the lowest loading (0.1%),
very few particles can be detected in the images and none larger
than approximately 20 Å.
CO oxidation was carried out over these catalysts in the
temperature range from 450 to 480 K at a total pressure of 1
atm. The partial pressure of CO and O2 were 4.4 and 2.3 Torr,
respectively, diluted in a He carrier gas. No significant catalyst
deactivation was observed for this reaction; steady state was
reached very quickly in all cases. The Arrhenius plot for the
0.1% catalyst is presented in Figure 2. Table 1 displays the
TOF at 450 K for the 0.1% catalyst, the dispersion and
The CO oxidation rate was determined for all four catalysts
in units of mol/(g of Pd s) at 460 K. For the 0.1% catalyst the
turnover frequency (TOF) was then calculated assuming 100%

10772 J. Phys. Chem. B, Vol. 101, No. 50, 1997
Rainer et al.
Figure 4. Arrhenius plots for CO2 formation for the CO + NO reaction
over four Pd/Al2O3 powders at the indicated partial pressures (1 atm
total pressure with He diluent).
Figure 3. Partial pressure dependence in each reactant for the CO +
O2 reaction over the 1.0% Pd/γ-Al2O3 catalyst at 460 K.
approximate average particle size for each catalyst based on
the oxidation rate normalized against the 0.1% loading, the Ea
for the reaction over each catalyst, and some literature values
for Pd single crystals and other supported Pd catalysts for
comparison. The excellent agreement in the TOF’s for the CO
oxidation rate with literature values for supported and single
crystal Pd catalysts (within a factor of 2-3) indicate the validity
of this approach for dispersion determination.
The activities for the CO + O2 reaction over the 1.0% Pd/
γ-Al2O3 catalyst at 460 K as a function of partial pressure in
each reactant are presented in Figure 3. The reaction is found
to be approximately negative first order in CO and positive first
order in O2, in agreement with other studies3,4 and consistent
with the mechanism that identifies CO desorption as the rate-
limiting step.
3.2. CO + NO Reaction. The CO + NO reaction was
carried out over each of the four catalysts in the temperature
range 540-580 K, with partial pressures of 4.5 and 5.2 Torr in
CO and NO, respectively. Upon initiation of the reaction, a
period of sharp deactivation was observed over which the
activity decreased by a factor of 5 or more. After approximately
8 h, steady state was reached with only very nominal decreases
in the activity observed as a function of time. All the data
presented here were obtained at steady state.
Figure 5. Arrhenius plots for NO reduction for the CO + NO reaction
over four Pd/Al2O3 powders at the indicated partial pressures (1 atm
total pressure with He diluent).
initial reduction than it experiences in the CO + NO run.
Additionally, comparisons of TEM images of reacted and
unreacted catalysts reveal no qualitative changes in the particle
size distributions, although, again, precise particle size distribu-
tions over wide ranges of particle sizes are sometimes difficult
to obtain by TEM. The nature of this deactivation was
investigated by performing the CO + O2 reaction over the 5%
catalyst both before and after the CO + NO reaction was run
and then comparing the rates.
Figure 6 shows the CO oxidation rate for the CO + O2
reaction as a function of time on stream (TOS) for a 5% Pd
catalyst over which the CO + NO reaction has previously been
carried out. The initial rate after CO + NO reaction exposure
decreases relative to the rate over the fresh catalyst by a factor
of ∼6. As the TOS for the CO + O2 reaction increases, the
CO oxidation activity increases until it plateaus at about half
the original value at around 400 min. This is suggestive of the
presence of an inhibiting site blocking surface species (or more
than one) that is (are) deposited during the CO + NO reaction
and then reacted off to some degree during the subsequent CO
+ O2 reaction. That only half of the original CO + O2 activity
is recovered implies either the presence of a poison that cannot
Figure 4 shows the Arrhenius plots for each catalyst. The
TOF frequencies are for CO conversion and are based on the
dispersions determined for each catalyst by the CO oxidation
reaction. Arrhenius plots for NO reduction, presented in Figure
5, consistently display TOF’s about 60% higher than for CO
oxidation and yield very similar Ea’s. Clearly, a particle size
dependence exists for this reaction, with a concomitant increase
in the activity with increasing average particle size. The largest
particles, those in the 2% Pd/R-Al2O3 catalyst, display a dramatic
activity enhancement (∼30×) over the particles in the 1% and
0.1% loaded catalyst. A similar particle size dependence for
supported Rh catalysts has been observed by Oh and Eickel,
who have reported a 45-fold increase in the activity for the
largest particles relative to the smallest in an average particle
diameter range from ∼10 to 700 Å.10
The deactivation that occurs upon initiation of the reaction
before steady state is reached could represent particle sintering,
resulting in decreased Pd surface area or possibly the buildup
on the surface of certain inhibiting species until some equilib-
rium is reached. Particle sintering seems unlikely as the catalyst
has been exposed to much harsher temperature conditions during

CO + O2 and CO + NO Reactions
J. Phys. Chem. B, Vol. 101, No. 50, 1997 10773
Figure 6. Effect of exposure to the CO + NO reaction on the rate for
the CO + O2 reaction over the 5% Pd/γ-Al2O3 catalyst. The horizontal
line at the top marks the CO + O2 reaction rate at 460 K before
exposure to the CO + NO reaction at 560 K for 1 day, and the scatter
plot data is the CO + O2 rate after CO + NO exposure. Time on stream
refers to CO + O2 exposure, not CO + NO exposure.
Figure 7. TPD monitoring (a) 15N2 desorption and (b) 15N2O desorption
over several different Pd coverages with the indicated average particle
sizes in a Pd/Al2O3/Ta(110) catalyst after exposure to 15NO at 550 K.
pronounced as in the powder catalysts, and the single-crystal
studies have indicated that close-packed Pd surfaces are more
active than open ones. CO + NO TPD data and postreaction
XPS have shown that the less active surfaces display higher
coverages of atomic N under reaction conditions.
be easily removed in this way or that some sintering is
responsible. Again, the latter seems unlikely based on the harsh
reduction conditions to which the catalyst has been exposed.
An identical experiment was conducted on the 1% Pd/Al2O3
catalyst. The CO + O2 oxidation rate for the catalyst used for
the CO + NO reaction was ∼35% of the rate observed on the
fresh catalyst. At CO + O2 TOS’s up to 1200 min, this rate
did not change appreciably. While it is unclear why this catalyst
did not exhibit the same regenerative behavior for the CO
oxidation reaction, this experiment does demonstrate that the
differences observed in the rates for the 5% and 1% loaded
catalysts are not simply the result of a differing degree of
sintering during the reaction. In fact, immediately following
the CO + NO reaction, the CO + O2 oxidation rate is decreased
more on the 5% relative to the fresh catalyst than on the 1%,
indicating that the 6-fold enhancement in the CO + NO rate
for the 5% relative to the 1% is not due to a higher degree of
sintering among the smaller particles during the CO + NO
reaction.
Regarding the identity of the poison(s), there are several
candidates. Small supported Pd particles have been shown to
be active for CO dissociation under certain conditions.27-30 An
inactive carbon species could well play the role of a site blocker,
and this would be consistent with the observed activity trend,
with the smaller particles subject to the most severe carbon
poisoning. Another consideration is the possibility of the
presence of an inactive adsorbed atomic nitrogen species that
is created and/or stabilized on undercoordinated defect sites.
Such a species has been proposed by Oh and Eickel to explain
the structure sensitivity as a function of particle size observed
for supported Rh catalysts and also differences in the activity
and Ea observed between single crystals and powders.10 For
Pd, based on comparisons with model catalyst systems, there
does indeed seem to be some indication that an inactive form
of adsorbed nitrogen plays an important role in determining the
activity of a catalyst for the CO + NO reaction.
This is illustrated in Figure 7, which shows TPD data for
15NO adsorbed on several different Pd loadings in the Pd/Al2O3/
Ta(110) model catalyst corresponding to the indicated average
particle sizes. The desorption spectra for N(a) recombination is
presented in Figure 7a; two features are evident. The higher
temperature desorption peak occurs above 600 K, higher than
the reaction temperatures employed in these experiments, which
suggests that the species giving rise to this feature is inactive
under these conditions. The ratio of the high-temperature feature
to the low-temperature feature is increasing with decreasing
particle size, consistent with the idea of having an enhancement
for the formation of the stable nitrogen species on the smaller
particles. Similar results have been recorded for the single
crystals. The Pd(100) surface, which is ∼5 times less active
than the (111) plane, displays a significantly higher ratio of the
higher temperature to the lower temperature N2 recombination
feature in a similar TPD experiment; additionally, coadsorption
of CO does not appreciably alter these results.19
Another indicator of an increased N(a) coverage on the smaller
particles is the absence of 15N2O in the TPD spectra, evident in
Figure 7b. For the smallest particles studied in the 1.1
equivalent monolayer (ML) loading, 15N2O desorption is beneath
the detection limit, suggestive of a relatively low coverage of
molecularly adsorbed 15NO on the surface. Conversely, the
largest particles in the 22 ML Pd coverage display significant
activity for N2O production. Similar results have been reported
for NO adsorbed on Pd/R-Al2O332 and Pd/SiO2/Mo(110) model
catalysts of the type described here.33
The reaction orders for CO and NO were determined for the
5% loaded catalysts. Plots for activity vs partial pressure for
each reactant are presented in Figure 8. The data reveal a
positive one-half-order dependence on NO concentration and a
negative first-order dependence in CO under the conditions
described. Similar findings have been reported for Pd single
crystals, although the degree of CO inhibition is not as
pronounced.19
Kinetics and UHV surface analysis studies on Pd single
crystals and planar model supported Pd/Al2O3/Ta(110) catalysts
have correlated the activity for the CO + NO reaction with the
formation of inactive atomic nitrogen.19,31 A particle size
dependence similar to that observed for the Pd/Al2O3 powders
has been recorded for the model particles, although not as
The temperature-dependent selectivities between the reaction
pathways for N2 and N2O formation for each catalyst are
presented in Figure 9. No clear correlation between selectivity

10774 J. Phys. Chem. B, Vol. 101, No. 50, 1997
Rainer et al.
to that previously observed for supported Rh catalysts. While
the origin of this particle size dependence has not been entirely
clarified, there has been some indication from model catalyst
studies that the preferred formation and stabilization of an
inactive atomic nitrogen species on the smaller particles, with
their higher step/edge defect densities, plays a role in determin-
ing the reaction rates. The activity loss for the CO + O2 reaction
after the catalyst has been exposed to the CO + NO reaction,
and subsequent recovery with increasing time-on-stream is
consistent with the formation of an inhibiting species on the
surface that is formed during the latter reaction and is removed
by the former.
Acknowledgment. We acknowledge with pleasure the
support of this work by the Department of Energy, Office of
Basic Energy Sciences, Division of Chemical Sciences, and the
Robert A. Welch Foundation.
Figure 8. Partial pressure dependence in each reactant for the CO +
NO reaction over the 5% Pd/γ-Al2O3 catalyst at 550K.
References and Notes
(1) Berlowitz, P. J.; Peden, C. H. F.; Goodman, D. W. J. Phys. Chem.
1988, 92, 5213.
(2) Ladas, S.; Poppa, H.; Boudart, M. Surf. Sci. 1981, 102, 151.
(3) Cant, N. W.; Hicks, P. C.; Lennon, B. S. J. Catal. 1978, 54, 372.
(4) Goodman, D. W.; Peden, C. H. F. J. Phys. Chem. 1986, 90, 4839.
(5) McCarthy, E.; Zahradnik, J.; Kuczynski, G. C.; Carberry, J. J. J.
Catal. 1975, 39, 29.
(6) Szanyi, J.; Goodman, D. W. J. Phys. Chem. 1994, 98, 2972.
(7) Szany, J.; Kuhn, W. K.; Goodman, D. W. J. Phys. Chem. 1994,
98, 2978.
(8) Xu, X.; Goodman, D. W. J. Phys. Chem. 1993, 97, 7711.
(9) Logan, A. D.; Paffett, M. T. J. Catal. 1992, 133, 179.
(10) Oh, S. H.; Eickel, C. C. J. Catal. 1991, 128, 526.
(11) Oh, S. H.; Fisher, G. B.; Carpenter, J. E.; Goodman, D. W. J. Catal.
1986, 100, 360.
(12) Simon Ng, K. Y.; Belton, D. N.; Schmieg, S. J.; Fisher, G. B. J.
Catal. 1994, 146, 394.
(13) Cho, B. K. J. Catal. 1994, 148, 697.
(14) Kostov, K. L.; Jakob, P.; Menzel, D. Surf. Sci. 1995, 331-333,
11.
Figure 9. Selectivity for N2 formation vs temperature for the CO +
NO reaction over each powder catalyst.
(15) Alikina, G. M.; Davydov, A. A.; Sazanova, I. S.; Popovskii, V. V.
React. Kinet. Catal. Lett. 1985, 27, 279.
(16) Valden, M.; Aaltonen, J.; Kuusisto, E.; Pessa, M.; Barnes, C. J.
Surf. Sci. 1994, 307-309, 193.
and particle size is evident. At 550 K there is no significant
difference in selectivity among the catalysts. As the reaction
temperature is increased, the selectivity for N2 formation
decreases for the 1% and 5% loaded catalyst, but does not
change appreciably for the other two catalysts. The change is
not large; the highest N2 selectivity observed for any system
was 43% and the lowest 28%.
N2 selectivity has also been found to be independent of
particle size for supported Rh catalysts. The reaction has been
observed to favor N2O formation at low temperatures (75% N2O
at 450 K) and N2 formation at higher temperatures (100% N2
at 550 K).10,34,35
(17) Carballo, L. M.; Hahn, T.; Lintz, H.-G. Appl. Surf. Sci. 1989, 40,
53.
(18) Vesecky, S. M.; Chen, P.; Xu, X.; Goodman, D. W. J. Vac. Sci.
Technol. A 1995, 13, 1539.
(19) Vesecky, S. M.; Rainer, D. R.; Goodman, D. W. J. Vac. Sci.
Technol. 1996, 14 (3), 1457.
(20) Shelef, M.; Graham, G. W. Catal. ReV.sSci. Eng. 1994, 36, 433.
(21) Vesecky, S. M.; Paul, J.; Goodman, D. W. J. Phys. Chem. 1996,
100, 15242.
(22) Campbell, R. A.; Goodman, D. W. ReV. Sci. Instrum. 1993, 63,
172.
(23) Szanyi, J.; Goodman, D. W. ReV. Sci. Instrum. 1993, 64, 2350.
(24) Chen, P. J.; Goodman, D. W. Surf. Sci. Lett. 1994, 312, L767.
(25) Xu, C.; Goodman, D. W. In Handbook Heterogeneous Catalysis;
Ertl, G., Knozinger, H., Weithemp, J., Eds.; VCH: Weinheim, in press,
and references therein.
(26) Rainer, D. R.; Wu, M.-C.; Mahon, D.; Goodman, D. W. J. Vac.
Sci. Technol. A 1996, 14 (3), 1184.
4. Conclusions
The structurally insensitive nature of the CO + O2 reaction
over Pd has been exploited to determine the relative dispersions
in several Pd/Al2O3 powder catalysts by measuring the rate of
the reaction. This method, assuming particles of hemispherical
shape, yields average particle sizes that are consistent with
observations using TEM. For the CO + NO reaction over the
same catalysts, a pronounced structure sensitivity is evident that
results in higher activities over larger Pd particles, comparable
(27) Stara, I.; Matolin, V. Surf. Sci. 1994, 313, 99.
(28) Matolin, V.; Gillet, E. Surf. Sci. 1990, 238, 75.
(29) El-Yakhloufi, M. H.; Gillet, E. Catal. Lett. 1993, 17, 11.
(30) Gillet, E.; Channokhone, S.; Matolin, V. J. Catal. 1986, 97, 437.
(31) Rainer, D. R.; Vesecky, S. M.; Koranne, M.; Oh, W. S.; Goodman,
D. W. , J. Catal. 1997, 167, 234.
(32) Cordatos, H.; Bunluesin, T.; Gorte, R. J. Surf. Sci. 1995, 323, 219.
(33) Xu, X.; Goodman, D. W. Catal. Lett. 1994, 24, 31.
(34) Hecker, W. C.; Bell, A. T. J. Catal. 1983, 84, 200.
(35) Cho, B. K.; Shanks, B. H.; Bailey, J. E. J. Catal. 1989, 115, 486.