Low-temperature catalytic carbon monoxide oxidation over hydrous and anhydrous palladium oxide powders

Article abstract of DOI:10.1016/j.jcat.2006.09.016

Low-temperature CO oxidation was carried out over hydrous and anhydrous PdO powders to examine the effects of surface water species. Hydrous PdO exhibits 100% CO conversion even at room temperature, whereas hydrous PdO pretreated in He at 400 °C and anhydrous PdO do not exhibit significant CO oxidation activity even at 100 °C. Approximately one-half of the CO is oxidized catalytically over hydrous PdO by gas-phase oxygen, whereas oxygen contained in the solid oxidizes the rest of the CO by a gas-solid reaction. X-ray photoelectron spectroscopy (XPS) data indicate that the water is present in hydrous PdO as hydroxyl groups and not molecular water. Without the presence of these hydroxyl groups, CO oxidation does not occur. During CO oxidation, metallic Pd forms and hydroxyl groups may react to form water, resulting in significant activity decay. The metallic Pd is inactive for catalytic CO oxidation at the temperatures used in this study. As the reaction temperature is increased from 30 to 100 °C, subsurface hydroxyl groups migrate to the surface, partially restoring the catalytic activity.

Full text of DOI:10.1016/j.jcat.2006.09.016

Journal of Catalysis 245 (2007) 35–44
www.elsevier.com/locate/jcat
Low-temperature catalytic carbon monoxide oxidation over hydrous and
anhydrous palladium oxide powders
Seung-Hoon Oh, Gar B. Hoflund ∗
Department of Chemical Engineering, University of Florida, Gainesville, FL 32611, USA
Received 13 July 2006; revised 11 September 2006; accepted 17 September 2006
Available online 19 October 2006
Abstract
Low-temperature CO oxidation was carried out over hydrous and anhydrous PdO powders to examine the effects of surface water species.
◦
Hydrous PdO exhibits 100% CO conversion even at room temperature, whereas hydrous PdO pretreated in He at 400 C and anhydrous PdO do
◦
not exhibit significant CO oxidation activity even at 100 C. Approximately one-half of the CO is oxidized catalytically over hydrous PdO by
gas-phase oxygen, whereas oxygen contained in the solid oxidizes the rest of the CO by a gas–solid reaction. X-ray photoelectron spectroscopy
(XPS) data indicate that the water is present in hydrous PdO as hydroxyl groups and not molecular water. Without the presence of these hydroxyl
groups, CO oxidation does not occur. During CO oxidation, metallic Pd forms and hydroxyl groups may react to form water, resulting in significant
activity decay. The metallic Pd is inactive for catalytic CO oxidation at the temperatures used in this study. As the reaction temperature is increased
◦
from 30 to 100 C, subsurface hydroxyl groups migrate to the surface, partially restoring the catalytic activity.
© 2006 Elsevier Inc. All rights reserved.
Keywords: Catalytic CO oxidation; Low-temperature CO oxidation; Hydrous PdO; Anhydrous PdO; X-ray photoelectron spectroscopy; Stoichiometric CO and O
2
1. Introduction
The purpose of this study was to simplify the catalytic
process as much as possible to gain information about the cat-
alytically active phase of Pd, the influence of surface water and
hydroxyl groups, and the mechanism of CO oxidation. In a pre-
vious study, H2-reduced Pd metal powder was studied for CO
oxidation [25]. The reaction feed contained both CO and O2 and
thus was both reducing and oxidizing. With a slightly O2-rich
(∼150 ppm) feed stream, PdO formed at the surface at 50–
300 ◦C, according to X-ray photoelectron spectroscopy (XPS)
data. This PdO was the catalytically active phase for CO oxi-
dation under the conditions used. Because reduction in H2 was
performed, Pd hydride may have been the starting material, and
a high concentration of hydroxyl groups likely was present on
these surfaces. Catalytic CO oxidation did not occur over Pt and
Pd at low temperatures because CO covered the surface, leaving
no sites for chemisorption of oxygen. In the case of platinized
tin oxide, Hoflund et al. [26] found that a Pt/Sn alloy provided a
two-site surface over which low-temperature CO oxidation oc-
curred. They proposed a mechanism in which bicarbonate or
formate species are the reaction intermediates formed by reac-
tion with surface hydroxyl species [27]. These hydroxyl species
Palladium-containing catalysts are widely used for oxidation
reactions including the complete oxidation of methane, auto-
motive exhaust treatment, and CO oxidation [1–9]. Although
oxide–supported platinum and gold catalysts [10–19] are gen-
erally more active for low-temperature (<100 ◦C) CO oxida-
tion than oxide–supported palladium catalysts, Pd-containing
catalysts are used for this reaction [4,5,20–23]. The supported
catalysts used for low-temperature CO oxidation are quite com-
plex; thus, determining how they function has been difficult,
even though CO oxidation is a relatively simple catalytic reac-
tion. Pd supported on different oxide powders gives different
catalytic behavior for both CO oxidation [12] and methane oxi-
dation [24], indicating that the support–metal or support–oxide
interaction is important for oxide–supported Pd catalysts in ox-
idation reactions.
*
Corresponding author. Fax: +1 352 392 9513.
E-mail address: garho@hotmail.com (G.B. Hoflund).
0021-9517/$ – see front matter © 2006 Elsevier Inc. All rights reserved.
doi:10.1016/j.jcat.2006.09.016

36
S.-H. Oh, G.B. Hoflund / Journal of Catalysis 245 (2007) 35–44
have been observed using XPS on catalytically active platinized
tin oxide surfaces [13].
peak, with the accepted binding energy (BE) of 336.9 eV. No
charging problems were encountered in this study. Model ex-
periments were performed to ensure that the brief exposure to
air during the transfer from the catalytic reactor to the character-
ization chamber did not alter the Pd chemical state. The Pd pow-
der was fully reduced in the reactor, exposed to air for 1 h, and
analyzed using XPS. Only Pd metal was observed. In another
experiment, both hydrous and anhydrous PdO were partially re-
duced in the UHV system. XPS was performed, the samples
were exposed to air for 1 h, and then XPS was performed again.
The spectra taken before and after the air exposure were iden-
tical, indicating that Pd metal was not oxidized and water was
not adsorbed by this air exposure at room temperature.
Based on these results, this present study of CO oxidation
was carried out over anhydrous PdO powder, anhydrous PdO
powder that had been exposed to humid air for about 1 year,
and hydrous PdO. The reaction and XPS data obtained provide
information about the role of water or hydroxyl groups in low-
temperature CO oxidation, the catalytically active species, and
the mechanism.
2. Experimental
Hydrous palladium (II) oxide (PdO·0.8H2O, Pd 77.81%,
purity 99.9% Alfa Aesar), air-exposed anhydrous palladium
(II) oxide (PdO, Pd 86.80%, purity 99.9%, Alfa Aesar) and
fresh, Ar-sealed anhydrous palladium (II) oxide (PdO, purity
99.998%, Aldrich) were examined. Two anhydrous PdO pow-
ders from different sources were used to ensure that the results
do not depend on sample source or minor differences in im-
purity levels. A specific mass (100 or 400 mg) of palladium
oxide was loaded into a 1/4-inch-diameter, U-shaped quartz
tube and supported by glass wool. The BET surface area was
35.7 m2/g for hydrous PdO, 28.9 m2/g for air-exposed anhy-
drous PdO, and 30.2 m2/g for anhydrous PdO. The tube was
then inserted into a Thermolyne 21100 furnace. A K-type ther-
mocouple was placed around the U-tube at the center of the
catalyst bed and connected to an autotuning PID controller. The
temperature was increased at a rate of 10 ◦C/min from room
temperature to 100 ◦C in flowing He and maintained accurately
inside the furnace within 1 ◦C. Ultra-high-purity test gases
obtained from Praxair were mixed to yield a composition close
to 1.0% CO and 0.50% O2 in helium. Because it is not possible
to maintain a stoichiometric feed stream due to mass controller
fluctuations, the feed stream was maintained slightly oxygen
rich by <200 ppm. The total flow rate was 45 cc/min at STP.
Initially, a pure He flush gas was passed through the catalyst
bed for 5 min. The feed gas mixture was bypassed and analyzed
in a Hewlett Packard 5890A gas chromatograph to determine
the feed concentration. For product analysis, a molecular sieve
5A column was connected to a thermal conductivity detector
(TCD). The oven and TCD temperatures were kept essentially
constant at 80 and 150 ◦C, respectively.
3. Results and discussion
Three different powders were examined in this study: a hy-
drous PdO powder, an anhydrous PdO powder that had been
exposed to humid air for about 1 year, and an anhydrous PdO
powder taken directly from a sealed bottle and stored in a desic-
cator. XPS Pd 3d spectra obtained from these three powders are
shown in Fig. 1, and the corresponding XPS O 1s and Pd 3p3/2
spectra are shown in Fig. 2. In Fig. 1, the primary difference
is in the FWHM of the Pd 3d peaks. The fresh anhydrous PdO
has the smallest FWHM (1.73 eV), and the hydrous PdO has
the largest FWHM (1.96 eV). The larger FWHM obtained from
hydrous PdO indicates an electronic interaction between the lat-
tice water species and PdO. The air-exposed, anhydrous PdO
has a slightly larger FWHM than that of fresh anhydrous PdO,
suggesting that only a small amount of water was adsorbed at
the surface during the long-term exposure to moist air.
The corresponding O 1s and Pd 3p spectra shown in Fig. 2
exhibit much larger differences than the Pd 3d spectra shown in
Fig. 1. However, these spectra are considerably more compli-
cated, because they contain contributions from multiple species.
Surface characterization studies were carried out on the cata-
lysts before and after reaction at various temperatures. The cat-
alyst powders were thoroughly mixed, pressed into aluminum
cups, inserted into the ultra-high-vacuum chamber, and charac-
terized by XPS. The bed lengths were 5 mm for the 100-mg
sample and 20 mm for the 400-mg sample. There may have
been catalyst surface compositional differences across the bed.
By thoroughly mixing the used powders, an average compo-
sition was obtained using XPS. XPS was performed using a
double-pass cylindrical mirror analyzer (DPCMA, PHI model
25-255 AR) and pulse-counting detection [28] by operating the
DPCMA in the retarding mode, with a pass energy of 25 eV
for the narrow-range spectra and 50 eV for the survey spectra.
The X-rays were generated using a MgKα X-ray source. The
raw spectra were shifted to align the predominant PdO Pd 3d5/2
Fig. 1. XPS Pd 3d spectra obtained from fresh hydrous PdO, air-exposed anhy-
drous PdO and argon-sealed anhydrous PdO.

S.-H. Oh, G.B. Hoflund / Journal of Catalysis 245 (2007) 35–44
37
Fig. 3. XPS O ls and Pd 3p difference spectrum obtained by subtracting the
anhydrous PdO spectrum shown in Fig. 2 from the hydrous PdO spectrum also
shown in Fig. 2.
Fig. 2. XPS O 1s and Pd 3p spectra obtained from fresh hydrous PdO, air-
exposed anhydrous PdO and argon-sealed anhydrous PdO.
Table 1
PdO is the contribution due to incorporated water. This dif-
ference spectrum, shown in Fig. 3, consists of a fairly narrow
peak with a BE of 531.1 eV. This is lower than the accepted
BEs for water by about 2 eV [30,31] but is typical of hydroxyl
groups [27,30–32]. This suggests that the formula of hydrous
PdO (PdO·H2O) is Pd(OH)2. The actual formula of this pow-
der is PdO·0.8H2O, indicating that some PdO is also present in
this powder. Apparently no information on Pd(OH)2 is avail-
able in the literature. The small feature at 528.7 eV is assigned
as chemisorbed atomic O. A similar species has been observed
in a Ag oxide study [33] and in a study of surface and subsur-
face oxygen on Pd(111) [34].
XPS peak binding energy assignments
Species
Peak
BE (eV)
References
0
Pd
Pd 3d
Pd 3p
335.3
532.3
[25,29]
[25]
5/2
3/2
PdO
Pd 3d
Pd 3p
O 1s
336.9
533.7
530.1
[25,29]
[25,29]
[25,29]
5/2
3/2
Pd(OH)
Pd 3d
Pd 3p
O 1s
336.9
533.7
531.1
This study
This study
This study, [27,30,31]
2
5/2
3/2
H O
2
O 1s
533.0
[30,31]
The catalytic activity of 100 mg of hydrous PdO for CO ox-
idation as a function of time at 100 ◦C is shown in Fig. 4. The
feed gas was 1.0% CO and 0.51% O2. Complete CO conversion
was obtained over 30 min with an oxygen conversion of only
46–52% for this period. For the catalytic reaction between gas-
phase CO and O2, 100% CO conversion required almost 100%
O2 conversion for the feed ratio used. The fact that the O2 con-
version is about 50% indicates that oxygen from the hydrous
PdO reacted with CO to produce CO2 at 100 ◦C. Specifically,
this oxygen must come from the PdO or hydroxyl groups, or
both sources. When pure Pd metal was tested with 1% CO and
0.51% O2 in He at 300 ◦C [25], complete conversion of CO and
nearly complete conversion of O2 were achieved. After 33 min,
the activity of hydrous PdO gradually decreased over the next
60 min and then became negligible. The O2 conversion remains
at approximately 50% of the CO conversion until both conver-
sions become negligible.
Peak BE assignment used in this study are given in Table 1;
they are quite close to the reference literature values. Lines are
drawn at the BEs of Pd 3p3/2 in PdO, Pd 3p3/2 in Pd metal,
and O 1s in PdO. No contributions from Pd metal are present
in these spectra based on the spectra shown in Fig. 1. The spec-
trum obtained from fresh anhydrous PdO is characteristic of
that of anhydrous PdO, containing a well-defined O 1s peak
due to O in PdO and a Pd 3p3/2 peak due to Pd in PdO. The
spectrum obtained from hydrous PdO is very different. Contri-
butions from O in PdO and Pd in PdO are apparent, but a large
contribution due to incorporated lattice water species is present
between 530 and 533 eV. The nature of the incorporated lat-
tice water species is discussed below. The spectrum obtained
from the air-exposed anhydrous PdO exhibits a Pd 3p3/2 peak
due to PdO and an O 1s peak due to PdO. It also exhibits struc-
ture between 530 and 533 eV due to exposure to humid air,
but this structure is much smaller than that of hydrous PdO.
The O 1s and Pd 3p3/2 spectra in Fig. 2 are consistent with the
data in Fig. 1 but are much more sensitive to incorporated sur-
face water species. The difference between the O 1s and Pd 3p
spectra obtained from hydrous PdO and the fresh anhydrous
To determine the effect of hydroxyl groups in hydrous PdO
on catalytic activity, 100 mg of fresh hydrous PdO was pre-
treated by drying in He at 400 ◦C for 1 h, then cooled to
100 ◦C in flowing He. The reaction-gas feed stream for the
400 ◦C—calcined hydrous PdO was slightly O2-rich (1.0% CO

38
S.-H. Oh, G.B. Hoflund / Journal of Catalysis 245 (2007) 35–44
◦
◦
Fig. 4. Comparison of CO and O conversions as a function of reaction time over 100 mg of hydrous PdO at 100 C before and after annealing in He at 400 C for
2
1 h. The feedstreams were slightly oxygen rich (1.0% CO and 0.51% O without annealing in He and 1.0% CO and 0.52% O with annealing in He) at total flow
2
2
rates of 45 cc/min.
Fig. 5. XPS Pd 3d spectra obtained from fresh hydrous PdO, hydrous PdO after
Fig. 6. XPS O 1s and Pd 3p spectra obtained from fresh hydrous PdO, hydrous
◦
◦
drying in He at 150 C for 120 min in the reactor and hydrous PdO after drying
PdO after drying in He at 150 C for 120 min in the reactor and hydrous PdO
◦
◦
in He at 400 C for 120 min in the reactor.
after drying in He at 400 C for 120 min in the reactor.
and 0.52% O2). The initial CO conversion of the dried hydrous
PdO was 20%, whereas the O2 conversion was only 10%. CO
conversion decreased to <5% after 30 min of reaction time.
Although the conversions were very low after the drying treat-
ment, the O2 conversion was still about one-half of the CO
conversion.
XPS Pd 3d spectra obtained from fresh hydrous PdO and hy-
drous PdO pretreated in He at 400 ◦C for 2 h are shown in Fig. 5.
In both cases, the predominant peaks are due to PdO or hydrox-
ide, and no feature due to metallic Pd is apparent. The spectrum
obtained from the fresh, hydrous PdO is slightly broader than
those obtained from the 400 ◦C annealed sample, indicating that
annealing removes hydroxyl groups from hydrous PdO by a
solid-phase reaction that forms PdO and H2O from Pd(OH)2.
This is consistent with the catalytic results shown in Fig. 4.
These peaks are not as narrow as those obtained from anhy-
drous PdO, most likely because a defect structure remains after
annealing at 400 ◦C. The corresponding high-resolution XPS
O 1s and Pd 3p3/2 spectra are shown in Fig. 6. Comparing these
spectra with those in Fig. 2 indicates that annealing results in
desorption of water from the hydrous PdO. After annealing at
400 ◦C in He for 2 h, the spectrum approaches that obtained
from fresh anhydrous PdO but with some surface and/or subsur-
face hydroxyl groups still present, because the valley between
two predominant peaks is not as deep.

S.-H. Oh, G.B. Hoflund / Journal of Catalysis 245 (2007) 35–44
39
◦
◦
Fig. 7. Comparison of CO and O conversions as a function of reaction time at 25 C over 400 mg of hydrous PdO, air-exposed anhydrous PdO and at 100 C
2
over 100 mg of argon-sealed anhydrous PdO. The feedstreams were slightly O rich (1.0% CO and 0.51% O for hydrous PdO and 1.0% CO and 0.502% O for
2
2
2
air-exposed anhydrous PdO and argon-sealed anhydrous PdO) at a total flow rate of 45 cc/min.
The CO and O2 conversions obtained at room temperature
from hydrous PdO and air-exposed anhydrous PdO are com-
pared in Fig. 7. Fresh hydrous PdO (0.40 g) was exposed to
a flow rate of 45 cc/min of 1.0% CO and 0.52% O2 in He at
25 ◦C. Complete CO conversion was maintained over 75 min
while the O2 conversion ranged from 41 to 44% of the CO con-
version over the same time period. After this period, the CO
activity decreased with reaction time until a CO conversion of
15% was reached after 217 min. Air-exposed anhydrous PdO
(0.40 g) was also exposed to a flow rate of 45 cc/min of 1.0%
CO and 0.51% O2 in He at 25 ◦C. For a short period (<12 min),
air-exposed anhydrous PdO also exhibited high CO conversion
(initially 100%). However, it rapidly deactivated and yielded
13% CO conversion after 19 min. The activity then decayed
slowly to a few percent. The oxygen conversion was initially
55% and decreased to 36% at 12 min, which are about one-
half of the corresponding CO conversions. The unusually high
initial activity of air-exposed anhydrous PdO is due to accu-
mulation of hydroxyl groups at the surface during exposure to
moist air over a long period. Fresh anhydrous PdO exhibited no
CO oxidation activity even at 100 ◦C, as shown in Fig. 7.
The XPS spectra shown in Figs. 8 and 9 were obtained from
hydrous PdO after reaction for 30 and 220 min, respectively
(0.40 g, slightly O2-rich feed and 25 ◦C). Metallic Pd features
form during reaction. After 30 min of reaction, they are small
shoulders on the low-BE side of the Pd 3d peaks, but after
220 min of reaction, they are quite prominent. At this time, al-
most 40% of the near-surface Pd(OH)2 and PdO are converted
to Pd metal. The O 1s feature obtained after reaction for 30 min
(Fig. 9) is similar to that obtained from fresh hydrous PdO, as
expected, but it is significantly decreased after 220 min of reac-
tion. These spectra are normalized to allow comparison of their
structures. As oxygen is removed during reaction, the Pd fea-
ture becomes predominant in the normalized spectra. Increased
structure is present in the metallic Pd 3p3/2 region after 30 min
Fig. 8. XPS Pd 3d spectra obtained from fresh hydrous PdO, hydrous PdO after
◦
exposure to 1.0% CO and 0.51% O for 30 min in the catalytic reactor at 25 C
2
and hydrous PdO after exposure to 1.0% CO and 0.51% O for 220 min in the
2
◦
catalytic reactor at 25 C.
of reaction and even more so after 220 min of reaction as PdO
is converted to metallic Pd. These spectra are complex because
the lattice hydroxyl group content also decreases with reaction
time.
The effect of reaction temperature for CO oxidation over
hydrous PdO is shown in Fig. 10. At 100 ◦C, complete CO
conversion and nearly 42% O2 conversion are maintained for
170 min. At 25 ◦C, complete CO conversion is maintained for
only about 65 min, and again the O2 conversion is slightly less
than one-half of the CO conversion over the time period stud-
ied. This behavior implies that less than one-half of the oxygen

40
S.-H. Oh, G.B. Hoflund / Journal of Catalysis 245 (2007) 35–44
in the feed gas is participating in CO oxidation, so the rest of
the required oxygen comes from PdO and/or Pd(OH)2. Accord-
ing to the XPS data shown in Figs. 8 and 9, when hydrous PdO
is no longer active for CO oxidation, most of the near-surface
PdO and/or Pd(OH)2 is converted into Pd metal, and the near-
surface hydroxyl groups are depleted. The conversion remains
high longer at 100 ◦C than at 25 ◦C. This implies that subsur-
face hydroxyl groups migrate to the surface faster at 100 ◦C
to maintain the high activity. The corresponding Pd 3d fea-
tures and Pd 3p/O 1s features are shown in Figs. 11 and 12,
respectively, before and after reaction at 25 ◦C for 220 min and
100 ◦C for 205 min. These data indicate that Pd metal forms
more rapidly at higher reaction temperatures even though the
catalyst still exhibits 60% CO conversion before the XPS data
are obtained. The O 1s and Pd 3p3/2 spectra shown in Fig. 12
also indicate that oxygen is removed from the near-surface re-
gion by the decreased size of the O 1s peak. This decrease is
larger after reaction at 100 ◦C than after reaction at 25 ◦C, con-
sistent with the Pd 3d spectra shown in Fig. 11. The feature due
to O in PdO is diminished simultaneously as the peak due to
metallic Pd grows.
Next, 400 mg of fresh hydrous PdO was reacted with 1% CO
(no O2 in the feed gas) at 25 ◦C. The data shown in Fig. 13 indi-
cate that hydrous PdO reacts with CO molecules even at 25 ◦C,
exhibiting a maximum CO conversion of ∼97% for 20 min
before gradually decreasing. The same experiment was also car-
ried out with 0.51% O2 in the feed. Comparing the two conver-
sions versus time curves, the conversion without O2 deactivates
faster and exhibits lower CO conversion over the entire reac-
tion time. The area between the two curves is proportional to
the amount of O2 that reacts catalytically from the gas phase.
The conversions of CO and O2 over 1.0% CO and 0.503%
O2 in He (45 cc/min) over 400 mg of hydrous PdO at increasing
reaction temperatures are shown in Fig. 14. At 25 ◦C, hydrous
PdO exhibits 100% CO conversion and 44% O2 conversion for
60 min before gradually deactivating to CO conversion of 15%
and O2 conversion of 5% at 220 min. The reaction tempera-
ture was then increased to 70 ◦C. At this temperature, the initial
CO conversion is 100%, whereas the O2 conversion is <13%.
The low O2 conversion at 70 ◦C indicates that O2 in the feed
gas does not react to a significant extent, so more oxygen from
the solid reacts with CO for a short period. At 100 ◦C, an ini-
tial 36% CO conversion is found, with no reaction of oxygen in
the feed stream. The CO conversion decreases rapidly to a few
percent at about 65 min. As the reaction proceeds at each tem-
perature, the surface becomes more reduced, and consequently
the conversion decreases. When the temperature is increased,
more hydroxyl groups diffuse to the surface from the subsurface
Fig. 9. XPS O 1s and Pd 3p spectra obtained from fresh hydrous PdO, hydrous
PdO after exposure to 1.0% CO and 0.51% O for 30 min in the catalytic reactor
2
◦
at 25 C and hydrous PdO after exposure to 1.0% CO and 0.51% O for 220 min
2
◦
in the catalytic reactor at 25 C.
◦
◦
Fig. 10. CO and O conversions as a function of reaction time over 400 mg of hydrous PdO at 100 C for 205 min and 25 C for 217 min. The feedstreams contained
2
◦
◦
1.0% CO and 0.52% O at 100 C and 1.0% CO and 0.51% O at 25 C. The total flow rate was 45 cc/min.
2
2

S.-H. Oh, G.B. Hoflund / Journal of Catalysis 245 (2007) 35–44
41
Fig. 11. XPS Pd 3d spectra obtained from fresh hydrous PdO, hydrous PdO
Fig. 12. XPS O 1s and Pd 3p spectra obtained from fresh hydrous PdO, hydrous
after exposure to 1.0% CO and 0.51% O for 220 min in the catalytic reactor at
PdO after exposure to 1.0% CO and 0.51% O for 220 min in the catalytic
2
2
◦
◦
25 C and hydrous PdO after exposure to 1.0% CO and 0.52% O for 205 min
2
reactor at 25 C and hydrous PdO after exposure to 1.0% CO and 0.52% O for
2
◦
◦
in the catalytic reactor at 100 C.
205 min in the catalytic reactor at 100 C.
region, yielding a high CO conversion. The high conversion pe-
riod becomes shorter as the near-surface region becomes more
fully reduced to Pd metal. The Pd metal at the surface is cov-
ered by chemisorbed CO at these temperatures, so gas-phase
oxygen cannot adsorb and react.
The CO conversion comparison for hydrous PdO (0.40 g)
with no O2 added to the feed is also shown in Fig. 14. The CO
conversion without O2 in the feed exhibits lower CO conver-
sion and faster deactivation for the entire reaction time at all
temperatures. Without O2, hydrous PdO yields a 9% CO con-
version after 216 min at 25 ◦C. The activity increases to CO
conversions of 86% at 70 ◦C and 36% at 100 ◦C. The area be-
tween the curves (open circles and open squares) represents the
amount of O2 from the feed stream that is reacted.
XPS spectra obtained from hydrous PdO reacted with and
without O2 in the feed stream for more than 6 h at 25, 70, and
100 ◦C are shown in Figs. 15 and 16. The XPS Pd 3d spectra
indicate that the chemical state of Pd in hydrous PdO changes to
metallic Pd during reaction in both cases. This alteration occurs
more rapidly without the presence of O2 in the feed stream, as
expected. The XPS O 1s and Pd 3p3/2 features (Fig. 16) behave
in a manner consistent with the XPS Pd 3d spectra; the O 1s
feature in PdO is diminished in size, and the Pd 3p3/2 feature is
◦
Fig. 13. CO conversion as a function of reaction time at 25 C over 400 mg of hydrous PdO in 1.0% CO and 0.51% O in He and 1.0% CO only (no O ) in He both
2
2
at a total flow rate of 45 cc/min.

42
S.-H. Oh, G.B. Hoflund / Journal of Catalysis 245 (2007) 35–44
◦
◦
Fig. 14. CO (open square) and O (open circle) conversions as a function of reaction time over 400 mg of hydrous PdO at 25 C (216 min), 70 C (77 min), and
2
◦
100 C (59 min). The feedstream contained 1.0% CO and 0.503% O in He at a total flow rate of 45 cc/min. CO conversion (solid squares) is also shown as a
function of reaction time at 25 C (216 min), 70 C (70 min), and 100 C (66 min). The feedstream contained 1.0% CO only (no oxygen) in He in this case. The
2
◦
◦
◦
total flow rate was 45 cc/min.
Fig. 15. XPS Pd 3d spectra obtained from fresh hydrous PdO, hydrous PdO after
Fig. 16. XPS O 1s and Pd 3p spectra obtained from fresh hydrous PdO, hydrous
exposure to 1.0% CO and 0.51% O in He at the end of the temperature–time
PdO after exposure to 1.0% CO and 0.51% O in He at the end of the temper-
2
2
program described in Fig. 10 and hydrous PdO after exposure to 1.0% CO only
in He at the end of the temperature–time program described in Fig. 14.
ature–time program described in Fig. 10 and hydrous PdO after exposure to
1.0% CO only in He at the end of the temperature–time program described in
Fig. 14.
of gas-phase oxygen reacts. At 100 ◦C, no gas-phase oxygen
reacts with CO, so the CO that is oxidized reacts only with oxy-
gen or hydroxyl groups that migrate to the surface. These data
demonstrate that metallic Pd is not a catalytically active phase
under the reaction conditions used in this study. This is consis-
tent with a previous study [25] demonstrating that metallic Pd
powder exhibits no catalytic activity for CO oxidation below
170 ◦C.
shifted to that of metallic Pd. These changes occur to a greater
extent for the catalyst run without O2 in the feed stream.
An explanation for the conversion data shown in Fig. 14 can
be based on the XPS data shown in Figs. 15 and 16. As reac-
tion occurs at 25 ◦C with oxygen in the feed, PdO and Pd(OH)2
are reduced to Pd metal, and H2O may be lost to the gas phase.
These processes result in decreasing conversion. When the tem-
perature is increased to 70 ◦C, subsurface oxygen migrates to
the surface, where it reacts with CO. The surface is less active
catalytically, as indicated by the fact that only a small amount
The argument can be made that H2O is present in hydrous
PdO as hydroxyl groups based on a connection between the re-

S.-H. Oh, G.B. Hoflund / Journal of Catalysis 245 (2007) 35–44
43
action data and the XPS data. For the conversion data obtained
using hydrous PdO at 100 ◦C (Fig. 4, solid symbols), 9.24 ×
1020 molecules of CO are converted to CO2 over 120 min. This
requires 4.62 × 1020 molecules of O2. The oxygen in the feed
stream contributes 2.14 × 1020 molecules, so 4.96 × 1020
atoms from the solid phase must react, which is 1.4 × 1016
O
O
atoms/cm2 based on a surface area of 35.7 m2/g and a sam-
ple size of 100 mg. If a PdO·H2O composition with a layer
atom density of 1 × 1015 atoms/cm2 were assumed and only
O from H2O reacted with CO, then the XPS Pd 3d features
shown in Fig. 11 would exhibit no metallic character, which
clearly is not the case. If only O in PdO reacts, then metal-
lic Pd forms. The H2O either remains dissolved in metallic Pd
in the top 42 atomic layers, which is not plausible, or it des-
orbs. This is deeper than the detection depth of XPS [35], so
only metallic Pd should be apparent in the XPS Pd 3d spec-
trum. This also is not the case. If the oxygen in both PdO and
assumed H2O reacted, then 21 atomic layers would be reduced
to Pd metal. XPS probes more deeply than this, so underlying
PdO would be apparent in addition to a larger metallic Pd fea-
ture. This is the case shown in Fig. 11. Therefore, both forms of
oxygen are equivalent in terms of CO oxidation. This supports
the finding described above that PdO·0.8H2O is actually a mix-
ture of Pd(OH)2 and PdO rather than PdO with lattice water,
which would not react with CO.
Several reactions must occur at the surface to explain the re-
action and XPS data. O2 from the gas phase must react with CO
to form CO2 (catalytic reaction), lattice oxygen must react with
CO to form CO2 (gas-phase reaction), and surface hydroxyl
groups may combine to form water, which desorbs (solid-phase
reaction). Schematic diagrams of these reactions are shown in
Fig. 17. No attempt is made to accurately represent the geom-
etry or bonding structure in these schematics. The conversion
data given in Figs. 4, 7, and 8 indicate that hydroxyl groups
must be present for catalytic CO oxidation. This is consistent
with the assertion that a bicarbonate species is the reaction in-
termediate, as shown in Fig. 17a. A similar mechanism was
proposed for low-temperature CO oxidation over platinized tin
oxide catalysts [27] based on extensive data. In this mecha-
nism, gas-phase CO chemisorbs at a vacancy near a hydroxyl
group, and gas-phase oxygen chemisorbs at a nearby vacancy
to form a bicarbonate species that decomposes to yield a des-
orbing CO2 molecule and two vacancies. More gas-phase CO
and O2 adsorb at these sites, and the catalytic cycle continues.
Under some conditions, the bicarbonate species do not decom-
pose rapidly, and they accumulate on the surface of supported
Pt and Au catalysts, resulting in activity decay [15,36–38]. The
second reaction is a gas–solid reaction in which gas-phase CO
reduces Pd oxide to Pd metal (Fig. 17b). In this case, CO ad-
sorbs near a hydroxyl group and a surface O, again forming a
bicarbonate species that decomposes to a desorbing CO2 and
metallic Pd. Another pathway cannot be eliminated. In this
pathway, gas-phase CO reacts with a hydroxyl group to form
a surface formate intermediate that decomposes to yield a CO2
molecule and a hydrogen that moves to another surface oxygen
to form a surface hydroxyl group. The formation of metallic Pd
during reaction is shown in Figs. 8–10, 11–13, and 15–17.
Fig. 17. Schematic diagram depicting the chemical reactions which occur over
hydrous PdO during catalytic CO oxidation.
Metallic Pd is not catalytically active at these low tempera-
tures [25], so reaction (b) results in the observed activity decay.
The third reaction (c) involves reaction between two hydroxyl
groups to form a water molecule that desorbs and PdO. This re-
action occurs at 400 ◦C as shown in Figs. 4, 5, and 6. It was
also observed at 150 ◦C (not shown). This reaction proceeds
more slowly at lower temperature. Because this reaction elim-
inates hydroxyl groups and produces anhydrous PdO, it also
contributes to activity decay. If this reaction rate is negligible at
lower temperatures, then reactions (a) and (b) proceed at nearly
equal rates, because the O2 reaction rate is about one-half of the
CO reaction rate. Nearly all catalysts undergo decay due to gas–
solid reaction and/or solid-phase reactions. The rate of activity
decay depends on the rate of these reactions. In some cases,
these reactions can benefit the catalytic process, and the activ-
ity will increase with time. In the present study the gas–solid
reaction is rapid, so the catalytic activity decreases rapidly.
Many studies have shown that low-temperature CO oxida-
tion catalysts consisting of Pt, Au, or Pd supported on differ-
ent oxides (noble-metal reducible oxide catalysts [39]) exhibit
widely varying catalytic properties in terms of both activity and
decay. This present study has shown that the activity of hy-
drous PdO powder decays very rapidly. Pd supported on ceria
is known to be a very active catalyst for CO oxidation, more so
than Pd supported on other oxides including zirconia, alumina,
titania, silica, and ZSM-5. It also exhibits essentially no decay
[40] even though the Pd chemical state is PdO [25]. Therefore,
ceria–supported Pd is maintained in the form of PdO by accept-

44
S.-H. Oh, G.B. Hoflund / Journal of Catalysis 245 (2007) 35–44
ing oxygen from the ceria, which is known to have excellent
oxygen storage and transfer capabilities. The activity and de-
cay behavior of any oxide–supported Pd CO oxidation catalyst
may be influenced by how rapidly oxygen is transferred from
the oxide support to the supported PdO phase.
[10] M. Haruta, T. Kobayashi, H. Sano, N. Yamada, Chem. Lett. 15 (1987)
405.
[11] M. Haruta, N. Yamada, T. Kobayashi, S. Iijima, J. Catal. 115 (1989) 301.
[12] S.D. Gardner, G.B. Hoflund, D.R. Schryer, B.T. Upchurch, E.J. Kielin,
Langmuir 7 (1991) 2135.
[13] S.D. Gardner, G.B. Hoflund, D.R. Schryer, B.T. Upchurch, J. Phys. Chem.
95 (1991) 835.
4. Conclusion
[14] G.B. Hoflund, B.T. Upchurch, E.J. Kielin, D.R. Schryer, Catal. Lett. 31
(1995) 133.
[15] G.B. Hoflund, S.D. Gardner, D.R. Schryer, B.T. Upchurch, E.J. Kielin,
Langmuir 11 (1995) 3431.
[16] G.B. Hoflund, S.D. Gardner, D.R. Schryer, B.T. Upchurch, E.J. Kielin,
React. Kinet. Catal. Lett. 58 (1996) 19.
[17] W.S. Epling, G.B. Hoflund, J.F. Weaver, S. Tsubota, M. Haruta, J. Phys.
Chem. 100 (1996) 9929.
[18] M. Daté, M. Haruta, J. Catal. 201 (2001) 221.
[19] M. Haruta, M. Date, Appl. Catal. A Gen. 222 (2001) 427.
[20] Y. Bi, G. Lu, Appl. Catal. B Environ. 41 (2003) 279.
[21] D.I. Kochubey, S.N. Pavlova, B.N. Novgorodov, G.N. Kryukova, V.A.
Sadykov, J. Catal. 161 (1996) 500.
[22] H. Zhu, Z. Qin, W. Shan, W. Shen, J. Wang, J. Catal. 233 (2005) 41.
[23] B.L.M. Hendriksen, S.C. Bobaru, J.W.M. Frenken, Surf. Sci. 552 (2004)
229.
[24] G.B. Hoflund, Z.-H. Li, W.S. Epling, T. Göbel, P. Schneider, H.W. Hahn,
React. Kinet. Catal. Lett. 70 (2000) 97.
[25] S.-H. Oh, G.B. Hoflund, J. Phys. Chem. A 110 (2006) 7609.
[26] S.D. Gardner, G.B. Hoflund, M.R. Davidson, D.R. Schryer, J. Catal. 115
(1989) 132.
[27] D.R. Schryer, B.T. Upchurch, B.D. Sidney, K.G. Brown, G.B. Hoflund,
R.K. Herz, J. Catal. 130 (1991) 314.
[28] R.E. Gilbert, D.F. Cox, G.B. Hoflund, Rev. Sci. Instrum. 53 (1982)
1281.
[29] Th. Pillo, R. Zimmermann, P. Steiner, S. Hüfner, J. Phys. Condens. Mat-
ter 9 (1997) 3987.
[30] K. Andersson, A. Nikitin, L.G.M. Petterson, A. Nilsson, H. Ogasawara,
Phys. Rev. Lett. 93 (2004), 196101-1.
[31] J.F. Moulder, W.F. Stickle, P.E. Sobol, K.D. Bowman, Handbook of
X-Ray Photoelectron Spectroscopy, Physical Electronics, Eden Prairie,
MN, 1995.
[32] J.E. Drawdy, G.B. Hoflund, S.D. Gardner, E. Yngvadottir, D.R. Schryer,
Surf. Interface Anal. 16 (1990) 369.
[33] J.F. Weaver, G.B. Hoflund, J. Phys. Chem. 98 (1994) 8519.
[34] F.P. Leisenberger, G. Koller, M. Sock, S. Surnev, M.G. Ramsey, F.P. Met-
zer, B. Klötzer, K. Hayek, Surf. Sci. 445 (2000) 380.
[35] G.B. Hoflund, in: J.C. Riviere, S. Myhra (Eds.), Handbook of Surface and
Interface Analysis: Methods in Problem Solving, vol. 57, Dekker, New
York, 1998.
In this study, low-temperature CO oxidation was examined
over hydrous PdO, anhydrous PdO that had been exposed to
humid air for about 1 year, and fresh anhydrous PdO powders.
The aim was to study the influence of incorporated water of
hydration in PdO without the presence of support effects. XPS
data indicate that the water of hydration is present as hydroxyl
groups, not molecular water. During CO oxidation over hydrous
PdO at 100 ◦C, the activity is initially high and drops to zero
over time as Pd metal forms and hydroxyl groups are depleted.
The oxygen required for CO oxidation originates from two
sources: gas-phase and solid-phase oxygen. Fresh Ar-sealed
anhydrous PdO exhibits no CO oxidation activity under these
same conditions, and both hydrous PdO that had been annealed
in He at 400 ◦C and air-exposed anhydrous PdO briefly exhibit
activity that decays to zero. These results all demonstrate that
the presence of hydroxyl groups is required for low-temperature
CO oxidation over PdO. CO oxidation experiments carried out
at 25 ◦C and then at 70 ◦C and then 100 ◦C with the same hy-
drous PdO sample indicate that subsurface hydroxyl groups
migrate to the surface to maintain high activity.
During CO oxidation, three reactions occur at the surface:
(1) oxidation of CO by O from PdO and hydroxyl groups,
(2) oxidation of CO by chemisorbed gas-phase oxygen, and
(3) formation of water, which desorbs by reaction of two hy-
droxyl groups, leaving PdO at the surface. A model is proposed
in which reactions (1) and (2) occur through a bicarbonate in-
termediate.
References
[1] W.S. Epling, G.B. Hoflund, J. Catal. 182 (1999) 5.
[2] R.J. Farrauto, M.C. Hobson, T. Kennelly, E.M. Waterman, Appl. Catal. A
Gen. 81 (1992) 227.
[3] H.S. Gandhi, G.W. Graham, R.W. McCabe, J. Catal. 216 (2003) 433.
[4] D.S. Stark, A. Crocker, G.J. Steward, J. Phys. E Sci. Instrum. 16 (1983)
158.
[36] M. Haruta, S. Tsubota, T. Kobayashi, H. Kageyama, M. Genet, B. Delmon,
J. Catal. 144 (1993) 175.
[37] B. Schumacher, Y. Denkwitz, V. Plzak, M. Kinne, R.J. Behm, J. Catal. 224
(2004) 449.
[5] D.S. Stark, M.R. Harris, J. Phys. E Sci. Instrum. 16 (1983) 492.
[6] G.C. Bond, L.R. Molloy, M.J. Fuller, J. Chem. Soc. Chem. Commun.
(1975) 796.
[38] M. Azar, V. Caps, F. Morfin, J.-L. Rousset, A. Piednoir, J.-C. Bertolini, L.
Piccolo, J. Catal. 239 (2006) 307.
[39] D.R. Schryer, G.B. Hoflund (Eds.), Low-Temperature CO-Oxidation Cat-
[7] W. Ibashi, G. Groppi, P. Forzatti, Catal. Today 83 (2003) 115.
[8] M. Lyubovsky, L. Pfefferle, Catal. Today 47 (1999) 29.
[9] A.K. Datye, J. Bravo, T.R. Nelson, P. Atanasova, M. Lyubovsky, L. Pfef-
ferle, Appl. Catal. A Gen. 198 (2000) 179.
alysts for Long-Life CO Lasers, NASA Conference Publication 3076,
1990.
[40] M.-F. Luo, Z.-Y. Hou, X.-X. Yuan, X.-M. Zheng, Catal. Lett. 50 (1998)
205.
2