Low-temperature ammonia oxidation on platinum sponge studied with positron emission profiling

Article abstract of DOI:10.1016/j.jcat.2004.04.035

The low-temperature deactivation of a platinum sponge catalyst used for ammonia oxidation was studied with the positron emission profiling technique (PEP). Evidence that irreversibly adsorbed nitrogen species deactivate the catalyst is presented. Two reactivity regimes are distinguished. Initial fast N₂ production at low surface coverage and a relatively slow N ₂ and N₂O production at steady state when the surface is fully covered. The fast deactivation of the platinum sponge is mainly caused by adsorbed nitrogen species. The formation of PtO is relatively slow compared to surface nitride. The fast initial deactivation of platinum sponge by nitrogen and oxygen species is greatly retarded at temperatures above 388 K. Temperature-programmed reaction together with temperature-programmed desorption experiments show reactivation of the catalyst above this temperature.

Full text of DOI:10.1016/j.jcat.2004.04.035

Journal of Catalysis 225 (2004) 466–478
www.elsevier.com/locate/jcat
Low-temperature ammonia oxidation on platinum sponge studied
with positron emission profiling
a
a
a
b
b,∗
a
D.P. Sobczyk, J. van Grondelle, P.C. Thüne, I.E. Kieft, A.M. de Jong, and R.A. van Santen
a
Schuit Institute of Catalysis, Laboratory of Inorganic Chemistry and Catalysis, Eindhoven University of Technology, P.O. Box 513, 5600 MB Eindhoven,
The Netherlands
b
Department of Applied Physics, Centre of Plasma and Radiation Physics, Eindhoven University of Technology, P.O. Box 513, 5600 MB Eindhoven,
The Netherlands
Received 12 November 2003; revised 26 April 2004; accepted 26 April 2004
Available online 1 June 2004
Abstract
The low-temperature deactivation of a platinum sponge catalyst used for ammonia oxidation was studied with the positron emission
profiling technique (PEP). Evidence that irreversibly adsorbed nitrogen species deactivate the catalyst is presented. Two reactivity regimes
are distinguished. Initial fast N production at low surface coverage and a relatively slow N and N O production at steady state when the
2
2
2
surface is fully covered. The fast deactivation of the platinum sponge is mainly caused by adsorbed nitrogen species. The formation of PtO
is relatively slow compared to surface nitride. The fast initial deactivation of platinum sponge by nitrogen and oxygen species is greatly
retarded at temperatures above 388 K. Temperature-programmed reaction together with temperature-programmed desorption experiments
show reactivation of the catalyst above this temperature.

2004 Elsevier Inc. All rights reserved.
13
15
Keywords: NH ; O ; Positron emission profiling; Ammonia oxidation; Deactivation; Platinum; Platinum sponge; Low temperature
3
2
1
. Introduction
TPD, and TPR, showed the formation of NO, N2, and
H2O [5–8]. The selectivity toward nitrogen products de-
pends mainly on the temperature and ammonia/oxygen ra-
tio. NO(a) plays a key role in the product formation. How-
ever, as Bradley et al. stated [8], at temperatures below 400 K
a route for the N2 formation is opened, which does not in-
volve an NO(a) intermediate. Other reported species present
on the surface are NH , O , and OH . The reaction path-
The study of low-temperature ammonia oxidation to ni-
trogen and water has become of increasing interest due to
the need to clean agricultural and industrial waste streams.
The need for purification of ammonia slip streams from in-
dustrial processes, like the SCR [1] and soda process, was
an impulse to start efforts to convert ammonia into harm-
less products. Platinum-based catalysts are promising in the
conversion of gaseous ammonia to harmless N2 and H2O at
relatively low temperatures. They have a high activity and
selectivity for N2 formation [2–4]. However, the most im-
portant drawback of platinum-based catalysts is their fast
deactivation.
(a)
(a)
(a)
ways change greatly with the pressure. At atmospheric pres-
sure and low temperature the main products are N and H O
2
2
with N2O as a by-product. NO is formed at a temperatures
of 573 K and higher [2]. IR experiments done by Matyshak
et al. suggested that the platinum surface is mainly cov-
ered with nitrogen species [9]. TPD, TPR, and XPS studies
showed that after a steady-state condition was reached NH
and OH species were the main intermediates present at the
catalyst’s surface [3]. At atmospheric pressure, the role of
NO(a) is only attributed to the production of the by-product,
N2O. The mechanism of fast initial deactivation of the plat-
inum catalysts has not been studied extensively. In the 1970s
Ostermaier et al. proposed that the initial deactivation of
Gas-phase ammonia oxidation over platinum catalysts at
low temperatures has been extensively studied before. Ex-
periments done at ultralow pressure on platinum crystals
with different techniques, like SIMS, AES, LEED, EELS,
*
Corresponding author.
E-mail address: a.m.de.jong@tue.nl (A.M. de Jong).
0021-9517/$ – see front matter  2004 Elsevier Inc. All rights reserved.
doi:10.1016/j.jcat.2004.04.035

D.P. Sobczyk et al. / Journal of Catalysis 225 (2004) 466–478
467
platinum supported on alumina is caused mainly by PtO for-
2.2. Steady-state experiments
mation [10,11].
In this paper new insights into the ammonia oxidation
mechanism are presented. A previous paper reported on the
ammonia adsorption step in the ammonia oxidation [12].
This study is also focused on the deactivation of platinum.
First, the conversion and product formation in the ammonia
oxidation over the platinum sponge catalyst are presented
as a function of temperature. Here, we will demonstrate
with positron emission profiling (PEP) experiments that be-
low 413 K the catalyst deactivates due to poisoning of the
catalyst surface, mainly by nitrogen species. In addition,
the nature of the deactivating species is further investigated
with X-ray photoelectron spectroscopy (XPS), temperature-
programmed desorption (TPD), and oxidation (TPO). Ex-
periments with NO and N2O pulses, and with preoxidized
platinum sponge provide complementary information on the
mechanism of the ammonia oxidation. Finally, we discuss a
reaction mechanism of the ammonia oxidation over platinum
at low temperatures.
A reaction mixture was used consisting of 2.0 vol% of
ammonia and 1.5 vol% of oxygen in helium with a total flow
of 46.5 cm /min. Conversion rates were calculated from
the measured concentrations, (Cin − Cout)/Cin. The steady-
state experiments were performed in the temperature range
of 413–573 K.
3
2
.3. Positron emission profiling
The Eindhoven 30 MeV cyclotron was used to produce
N and O nuclei. N was produced via irradiation of a
1
3
15
13
water target with highly energetic protons of 16 MeV. The
irradiation time was 10 min and a typical beam current of
5
00 nA was used. The target was a flowthrough water target,
with a total volume of 7 ml containing a dual foil (Duratherm
1
3
6
[
00, thickness 15 µm). In this way formed [ N]nitrate and
1
3
13
N]nitrite were subsequently reduced to NH3, using De-
Varda’s alloy method [18,19]. The production method of
1
3
gaseous pulses of [ N]NH3 is described elsewhere [20].
1
5
16
15
To produce labeled oxygen ( O O written as [ O]O2),
a nitrogen gas target (75 ml/min) at a pressure of 4 bar was
continuously irradiated with 9.2 MeV deuterons (500 nA)
2
. Experimental
[
21,22]. The irradiated effluent was remotely transported
2
.1. Catalytic reactor
from the target vault into the PEP laboratory. The labeled ef-
fluent was transported to a GC passing first sodalime and ac-
tivated charcoal absorbers to remove by-products of the irra-
The platinum sponge was acquired from Johnson Mat-
1
3
diation. A pulse time of 3–10 s was used to inject [ N]NH3
they. The sponge sample was of > 99.9% purity. The par-
ticle size of the sponge was between 250 and 350 µm and
the size of the small nonporous particles about 1.0–5.0 µm.
The amount of platinum sites calculated from BET measure-
1
5
or [ O]O2 into the reactant stream. The concentration of
radiolabeled molecules varies in each pulse and thus the
absolute concentration of radiolabeled molecules in the var-
ious experiments cannot be directly compared. However, the
labeled species are either introduced in a large flow of nonla-
beled ammonia or as trace amounts in ammonia-free flows.
As such, these variations will not affect the results between
1
8
ments is 3.0 × 10 sites/g and via hydrogen chemisorption
1
8
1
.3×10 sites/g. The metal surface area is determined with
2
BET to be 0.099 m /g. Catalytic tests were performed in
a fixed-bed reactor setup equipped with a quadrupole mass
spectrometer (Balzers Instruments Omnistar GSD 3000),
which was calibrated for on-line analysis of reactants and
products. A quartz tube with an internal diameter of 4 mm
was used as the reactor. For the experiments a sample of
the various experiments.
The 13N or 15O nuclei emit a positron upon decay. Posi-
tron emission profiling is based on the detection of the two
5
11 keV gamma photons which originate from the annihi-
lation of this positron with an electron. These two 511 keV
gamma photons are simultaneously emitted in opposite di-
rections and they travel typically a few centimeters in solid
matters. Coincidental detection of the two photons by scin-
tillation detectors (BGO) provides the position of the annihi-
lation. In practice, the tubular reactor is horizontally placed
between two arrays (upper and lower) of nine scintillation
detectors. In both arrays the detectors were tightly placed,
which results in the spatial resolution of 2.9 mm. The av-
erage concentration of all radiolabeled molecules within a
certain volume (length 2.9 mm) is measured at time inter-
vals of 0.5 s. The measurement of radiolabeled molecules
is simultaneous over the total detection length (5 cm), thus
within the 17 volume segments. In this way the concentra-
tion distribution of radiolabeled molecules can be measured
as a function of position and time [23,24].
1
.8 g of pure platinum sponge was used in a catalytic bed
with a length of 4.0 cm. The experiments were done at tem-
peratures between 323 and 673 K.
Before each experiment the platinum sponge was in situ
reduced by heating the sample in a 10 vol% H2/He flow
3
(
40 cm /min) from 298 to 673 K. Subsequently the sample
was kept at this temperature for 2 h. Then the catalyst was
flushed with He for 20 min before the reaction temperature
was set.
A preoxidized platinum sponge is obtained in the follow-
ing way. The reduced catalyst was pretreated with a 1 vol%
O2/He flow (46.5 cm /min) for 1 h at 373 K. Oxygen ad-
sorbs dissociatively at this temperature [13–17]. Then the
catalyst was flushed with He (46.5 cm /min) for 1 h before
the reaction temperature was set.
3
3

468
D.P. Sobczyk et al. / Journal of Catalysis 225 (2004) 466–478
2
.4. PEP pulse experiments
373 K. The formation of NO was not observed. The detec-
tion of the water signal by the mass spectrometer is delayed
due to readsorption of water on platinum. Figs. 1A and 1B
show two regions, in which the selectivity changes with
time. The start of the ammonia oxidation shows a selective
formation of nitrogen and water. The second region in the N-
The reduced platinum catalyst was kept under a flow of
2
.0 vol% of ammonia and 1.5 vol% of oxygen in helium
3
with a total flow of 46.5 cm /min at 273–573 K (gas hourly
−
1
space velocity (GHSV) = 5600 h ). Subsequently, a pulse
1
3
15
of [ N]NH3 or [ O]O2 was injected in this reaction flow.
The radiolabeled pulse experiments over the platinum and
preoxidized platinum sponge were performed in a similar
manner.
product selectivity begins when N O evolves. In Fig. 1A this
2
is after 40 s, subsequently N production decreases and N O
2
2
selectivity sharply increases to 16% and decreases in time
to 7%. At 373 K the catalyst is highly active longer and the
N2O formation is observed after about 10 min. Concentra-
tion profiles for N2 and N2O are similar in the temperature
range between 323 and 473 K. Initially, only N2 is formed
in this temperature range. At higher temperatures the du-
ration of N2 formation at a high stable conversion level is
greatly extended. It is thought that as the catalyst becomes
covered with the reaction intermediates, N2O evolves (low
quantities) and the production of N2 is strongly reduced. Ta-
ble 1 presents the initial deactivation time, i.e., the time till
the N2 concentration sharply decreases, in the temperature
range 323–473 K. The platinum catalyst possesses a similar
high initial activity level at all these temperatures. The ini-
tial deactivation time strongly depends on temperature, from
2
.5. XPS
XPS data are obtained with a VG Escalab 200 spectrom-
eter, equipped with an Al-Kα source and a hemispherical
analyzer connected to a five-channel detector. Measurements
are done at 20 eV pass energy. XP spectra were taken of
freshly reduced platinum, and two samples that were used
in the ammonia oxidation experiment performed at 323 K,
one for 20 s and one for 2 h. All samples were transported to
the XP spectrometer in an oxygen-free environment. It was
not possible to get a clear O(1s) signal for oxygen on plat-
inum due to the interference with the oxygen signal of the
indium foil, which was used to fix the platinum particles to
the sample holder.
0
.6 min at 323 K till 13 min at 373 K. At higher temper-
atures, however, the deactivation of the catalyst is further
retarded. At 398 K the catalyst is deactivated after 16.5 h
and at 423 K not even after 24 h time on stream.
2
.6. Temperature-programmed experiments
To investigate more closely the influence of temperature
on the initial deactivation a temperature-programmed reac-
tion experiment is performed (Fig. 2). After the catalyst is
deactivated at 373 K, the activity of the catalyst increases
at approximately 383 K. The N2O formation shows an inter-
Temperature-programmed experiments were performed
after the catalyst was deactivated at 323 K. First the catalyst
3
was flushed with He (20 cm /min) for 1 h. Next, the tem-
3
perature was raised with 10 K/min (20 cm /min) under He
(
(
TPD) or 1.0 vol% O2/He (TPO) flow or 0.5 vol% NO/He
TP-NO) flow.
Table 1
Initial deactivation times for the platinum sponge catalyst
3
. Results and discussion
Temperature
23 K
348 K
Initial deactivation time
3
0.6 min
2.1 min
13 min
20 min
> 20 h
3
.1. Ammonia oxidation at 323–473 K
373 K
388 K
423 K
Fig. 1 shows the concentration of the formed products
(
N2, N2O, and H2O) in the ammonia oxidation at 323 and
−1
Fig. 1. Concentration of N , N O, and H O versus time for the ammonia oxidation reaction at (A) 323 K, (B) 373 K (GHSV = 5600 h , NH /O = 2/1.5,
2
2
2
3
2
3
flow = 46.5 cm /min).

D.P. Sobczyk et al. / Journal of Catalysis 225 (2004) 466–478
469
tially adsorbs at the fcc hollow site at low coverage, at higher
coverage NO additionally adsorbs on the atop site in a tilted
geometry [31,32]. Thus, the adsorption of oxygen on plat-
inum should not be completely blocked by ammonia or other
surface species. The additional production of water cannot
explain this conversion of oxygen, because during the depo-
sition of oxygen the water production is already decreased.
We speculate that the formation of subsurface oxygen might
be the cause of the slower decrease of the conversion of oxy-
gen or that oxygen adsorbs on the platinum surface without
subsequent formation of N2O and N2.
Fig. 2. Ammonia oxidation at 373 K till the catalyst is deactivated followed
by a temperature-programmed reaction. TP-reaction part only is shown
3
.2. Positron emission profiling
−1
3
(
10 K/min, GHSV = 5000 h , NH /O = 2/1.5, flow = 46.5 cm /min).
3
2
1
3
The injection of [ N]NH3 in the reactant flow at the very
esting development. First the concentration of N2O increases
together with N2 till 413 K and then decreases very fast to
a very low concentration. The reactivation behavior of the
catalyst is the reverse of the deactivation. Above 413 K pro-
duction of N2 is favored above N2O. Finally, the catalytic
activity of the platinum sponge catalyst is restored to its ini-
tial value and with similar high selectivity for N2. Thus, the
stable activity of the platinum catalyst above 413 K for a
relatively long time suggests that adsorbed species, which
deactivate the catalyst at lower temperatures, are not present
at the surface.
With respect to initial deactivation of the platinum sponge
the conversion of ammonia and oxygen shows an interesting
feature. Fig. 3 presents the conversion of both reactants as
a function of time at 323 and 373 K. Clearly, as the cata-
lyst is very active in the beginning of the reaction (region 1)
both conversions are about 100%. However, as the initial
activity of the catalyst rapidly drops, the N2 concentration
decreases (Fig. 1), the conversion of oxygen does not de-
crease as fast as that of ammonia. In time the conversions of
ammonia and oxygen reach the same low conversion level.
From the literature it is known that oxygen and ammonia
preferentially adsorb on platinum at different sites. Oxygen
molecules occupy hollow sites [25–28] whereas ammonia
molecules occupy on-top sites [29,30]. NH and NH2 adsor-
bates (2-fold); H, O, and N atoms (3-fold) [29]. NO preferen-
beginning of the reaction shows in Fig. 4A that 13N-labeled
species are formed and adsorbed in front of the catalyst bed
(
catalyst bed starts at position 1.5 cm). A part of 13N activ-
ity went through the total catalyst bed, observed as a thin
line of activity during the 10 s of injection time, also at the
last positions of the catalyst bed. Since ammonia is initially
1
3
converted to N2, the measured activity is gaseous [ N]N2
Fig. 1A). Fig. 4A shows that part of the radiolabeled nitro-
(
gen species desorb much slower from the catalyst surface,
as the slow intensity decrease indicates. A substantial part of
the labeled nitrogen species remains adsorbed at the catalyst
surface and does not move through the catalyst bed during
the measurement time of 30 min. Fig. 4B shows that a la-
beled oxygen pulse does not remain, in contrast to Fig. 4A, at
the catalyst surface. The initial deposition of the O species at
the surface is negligible. Oxygen is mainly converted into the
gaseous products (water), as conversion of oxygen is very
high at that moment. The retention time of water is short,
which means that the desorption–readsorption equilibrium
of water is fast. This means that water does not compete with
ammonia for adsorption sites and thus water does not poison
the catalyst surface. This is in line with the results of Van
de Broek [33], which showed that the addition of water to
the reaction flow does not influence the performance of the
catalyst.
−1
Fig. 3. Conversion of NH and O versus time for the ammonia oxidation reaction at (A) 323 K, (B) 373 K (GHSV = 5600 h , NH /O = 2/1.5,
3
2
3
2
3
flow = 46.5 cm /min).

470
D.P. Sobczyk et al. / Journal of Catalysis 225 (2004) 466–478
1
3
15
13
Fig. 4. PEP image of [ N]NH and [ O]O pulse injection into the reaction stream of NH /O /He in the first 2 s of the ammonia oxidation. (A) [ N]NH
3
3
2
3
2
15
−1
3
PEP image, (B) [ O]O PEP image (T = 323 K, GHSV = 5600 h , NH /O = 2/1.5, flow = 46.5 cm /min). Color intensity represents the concentration
2
3
2
13
15
of [ N]NH or [ O]O (dark, high concentration).
3
2
Table 2
Percentage of the labeled species remaining at the catalyst after a radiola-
beled pulse was injected in the reaction flow at the start of the ammonia
oxidation
1
3
15
Temperature
K)
N remained adsorbed
O remained adsorbed
(
(%)
(%)
323
348
373
423
27 ± 3
16 ± 2
17 ± 2
< 1
3 ± 1
2 ± 1
4 ± 1
< 1
1
3
Fig. 5. Normalized sum of radioactivity versus time for [ N]NH and
3
species. The high activity of the catalyst is maintained due
to the movement of the reaction front to the next positions
in the catalyst bed. This is observed, for example, at 323 K
15
[
O]O pulse injection into the reaction stream of NH /O /He in the
2
3
2
−1
first 2 s of the ammonia oxidation (T = 323 K, GHSV = 5600 h
,
3
NH /O = 2/1.5, flow = 46.5 cm /min).
3
2
1
3
when [ N]NH3 is injected at the moment that the reac-
tion was already 20 s on stream (not shown). Thus, in time
this deactivation front moves to the end of the catalyst bed,
the catalyst is covered with reaction species and the deac-
tivation of the catalyst is observed. An experiment with a
half amount of the catalyst also supports this reaction front
movement. This experiment showed in the same way the
formation and concentration of the products; however, the
catalyst remained active for half the time of the normally ap-
plied catalyst bed. Thus, below 413 K the catalyst remains
initially active because the reaction zone moves to the sub-
sequent bed positions, after the previous positions became
fully covered with the adsorbed reaction species.
Fig. 5 shows the total radioactivity of the nitrogen- and
oxygen-containingspecies adsorbed at the catalyst surface at
3
23 K. The oxygen containing species do not stay adsorbed
1
3
at the catalyst, in contrast to the nitrogen species. The
N
profile shows a very fast decrease of the radioactivity in the
first 20 s, which is assigned to N2 formation. A relatively
slow decrease of the 13N profile is observed during the 20
to 400 s that reaction is on stream. Fig. 5 clearly shows the
1
3
injection moment of NH3 and that after 400 s on stream,
a stable amount of 13N activity is left.
To calculate the amount of irreversibly adsorbed N and
O species the sum of radioactivity at 5 positions is used.
Table 2 shows the amount of the irreversibly adsorbed radi-
olabeled nitrogen and oxygen species at 323–423 K. With
increasing temperatures less nitrogen is deposited on the
platinum catalyst and the amount of oxygen deposition is
rather stable up to 373 K. The PEP images do not differ too
much up to 373 K. As already shown in Fig. 2 the catalyst
is very active above 413 K and all ammonia and oxygen are
converted to nitrogen and water.
1
3
15
Injection of a [ N]NH3 or [ O]O2 pulse after the ini-
tial deactivation of the catalyst confirmed that the platinum
surface is fully covered and that conversion of ammonia and
oxygen is low (not shown). Certainly, no significant amount
of the nitrogen or oxygen species remains adsorbed at the
catalyst surface. A similar behavior was seen after 2 h on
1
5
stream. This result for [ O]O2 pulses contrasts Fig. 3, as
1
5
we expected to observe a significant deposition of [ O]O2
on the surface. We cannot explain this obvious difference,
especially because both measurements are reliable and re-
producible.
As already stated discussed Fig. 4A, the ammonia oxi-
dation reaction proceeds at the first positions of the catalyst
bed. These positions are mainly deactivated with nitrogen

D.P. Sobczyk et al. / Journal of Catalysis 225 (2004) 466–478
471
Fig. 6. XP N(1s) spectra of a platinum sponge catalyst. (A) fresh; after 20 s ammonia oxidation on stream (Pt sponge at initial state); and after 2 h on stream
Pt sponge at steady state); (B) enlargement of N(1s) spectrum of Pt sponge at initial state; (C) enlargement of N(1s) spectrum of Pt sponge at steady state.
(
3
.3. Characterization of the adsorbed nitrogen/oxygen
With higher coverage the symmetric peak was observed at
397.7 eV and the shoulder at 399.5 eV. This result shows
similar features to earlier XPS experiments on deactivated
platinum sponge performed by Van den Broek et al. [3].
According to Sun et al. [34], the peak at lower binding
energy of 398.0 eV is assigned to NH and the peak at higher
binding energy to NH2 species. The BE peaks representing
NO(a) [35,36], NH3 [34,37], N2 [38], N2O [39], and atomic
nitrogen [40] were not observed in our XP spectra. With
respect to the NO(a) this indicates that NO(a) can be only
present at very low quantities at the catalyst surface.
species
3
.3.1. X-ray photoelectron spectroscopy (XPS)
In this X-ray photoelectron spectroscopy experiment the
ammonia oxidation reaction was stopped after respectively
0 s and 2 h on stream and then the reactor was flushed with
2
He. Therefore the gaseous products of the ammonia oxida-
tion are expected to be desorbed from the catalyst surface
and should not be present in the XP spectra. The first sam-
ple was taken the moment the catalyst started to deactivate
showing N2 and N2O as the products at a still relatively high
conversions. The second sample was taken when the catalyst
is totally deactivated showing N2 and N2O as the products at
very low conversions. Fig. 6A shows the XP-spectra of these
samples and a clean background for N(1s) for a reduced plat-
inum sponge.
3.3.2. Temperature-programmed desorption (TPD)
In the TPD spectrum three products are observed. N2,
N2O, and H2O. Fig. 7 shows that N2O desorbs first, already
at 388 K. We assume that N2 and N2O are not molecularly
bound on the surface at the start of the TPD experiment be-
cause it is thermodynamically favorable for N2 and N2O to
desorb from platinum [41,42]. Then the desorption of N2O
indicates that the reaction of NO(a) with N(a) took place.
The amount of produced N2O is relatively low. This explains
the XPS results in which NO was not detected on the plat-
inum surface. Further, Fig. 7 shows that the peak of N2 has
two shoulders with peak maxima at 433 K and at 483 K.
Again, water desorption is concomitant with nitrogen de-
sorption. The mass spectrometer signal of water is somewhat
delayed due to the longer residence time of water on plat-
inum caused by the readsorption of water as will be shown
At the beginning of the N2 and N2O formation a broad
asymmetric peak is observed, which is centered around
3
98 eV. The intensity of the N(1s) signal of the deactivated
catalyst is higher. The relative number of N(1s) species per
platinum area, normalized to indium, is about 35% lower in
the initial state of the oxidation than after the total deactiva-
tion of the catalyst. The shape of the peak at a binding energy
of 398 eV does not change significantly for both experiments
as presented in Figs. 6B and 6C. For the N(1s) spectrum at
low coverage a symmetric peak was found at a binding en-
ergy of 398.0 eV with a small shoulder around 399.4 eV.

472
D.P. Sobczyk et al. / Journal of Catalysis 225 (2004) 466–478
Fig. 7. Formation of N , N O, and H O measured by online mass spectrom-
Fig. 8. Formation of N2, N2O, and H2O measured by online mass spec-
trometry in a temperature-programmed oxidation experiment after ammo-
2
2
2
etry in a temperature-programmed desorption experiment after ammonia
3
nia oxidation at 323 K (10 K/min, 1.0 vol% O /He flow of 40 cm3/min).
oxidation at 323 K (10 K/min, He flow of 40 cm /min).
2
Nitrogen is formed at 383 K followed by the water pro-
duction with a peak maximum at 403 K. The production of
N2O is clearly much higher than in the TPD experiment.
This can be explained by the reaction of NH with O(a) or
OH(a), which gives atomic nitrogen and water as in the TPD
experiment:
in Section 3.6. This simultaneous desorption of nitrogen and
water indicates that surface reactions of both oxygen- and
nitrogen-containing species occur. The presence of atomic
nitrogen on the surface is not likely, because recombination
of atomic nitrogen would lead directly to N2 without water
formation. Absence of atomic nitrogen on the surface is also
in agreement with the XPS measurements. At this relatively
high surface coverage some O(a) and OH is left on the sur-
face, trapped in between NHx species. For that reason the
production of water through the reaction of two hydroxyls is
only partly responsible for the water formation,
NH(a) + OH(a) → N(a) + H2O(a),
NH(a) + O(a) → N(a) + OH(a).
O2(g) will adsorb and dissociate on every vacant site
which becomes available. This will lead to a production of
NO(a):
2
OH(a) ↔ H2O + O(a) + ∗.
∗
NH(a) + 2O(a) → NO(a) + OH(a) + ,
The atomic oxygen reacts further with NHx species ab-
∗
N(a) + O(a)NO(a) + .
stracting hydrogen. Two possible routes for the OH in-
volvement in the formation of N2 can be proposed, either
from NH or NH2. At low temperature the reaction of NH
with OH to form water is favored with reaction energy of
Thus, at the surface there will be a competition between
reactions toward N2 or N2O:
∗
2
N(a) → N2 + 2 ,
−
11 kcal/mol [29]. The twostep reaction of NH2 (via NH)
with OH has a reaction energy of −26 kcal/mol:
∗
NO(a) + N(a) → N2O + 2 ,
∗
NO(a) + NH(a) → N2 + OH(a) + ,
NH(a) + OH(a) → N(a) + H2O(a),
∗
NO(a) + NH(a) → N2O + H(a) + .
∗
2
N(a) → N2 + 2 .
The TPO spectrum shows that the maximum of the N2
peak appears at lower temperatures than the maximum of
the N2O peak. This indicates that the production of N2O is
dependent on the formation of NO(a) from NH(a) species or
on the availability of free sites for oxygen adsorption and
dissociation.
At higher temperatures, one additional peak evolves with
a maximum at 493 K, which is assigned to NO. This suggests
that at a certain moment the sole species originating from
NHx on the platinum surface is NO(a). Apparently, oxygen
strongly promotes the NO formation and blocks the N2 and
N2O production.
At higher temperatures NH2 reacts to form N2:
NH2(a) + OH(a) → NH(a) + H2O(a),
NH(a) + OH(a) → N(a) + H2O(a),
∗
2
N(a) → N2 + 2 .
The third nitrogen peak with a maximum at 483 K is most
probably formed via an NO(a) intermediate, already formed
at much lower temperatures. A relatively high surface cov-
erage promotes the NO formation, but desorption of NO is
slow at these relatively low temperatures [8,43,44].
3
.3.3. Temperature-programmed oxidation
The most important results from the TPO experiment are
3.3.4. Temperature-programmed reaction with NO
(TP-NO)
the evolution of one N2 peak, together with one H2O peak
and additional NO evolution at higher temperatures (Fig. 8).
Fig. 9 shows the products of the reaction between the sur-
face species left after the ammonia oxidation and NO from

D.P. Sobczyk et al. / Journal of Catalysis 225 (2004) 466–478
473
does not take place, because no N2O or N2 are formed. At
73 K some N2O(g) and N2(g) are formed again, which prob-
5
ably originate from the NH2 species. The dissociation of the
NO(g) starts above 623 K (not shown).
3
.4. Experiments on preoxidized platinum
Before ammonia oxidation was started, the platinum
sponge catalyst was first preoxidized as described in the ex-
perimental part. Fig. 10 shows the conversion and selectivity
of the preoxidized catalyst at 373 K. The same selectivity
characteristics as on the reduced platinum sponge catalyst
are observed (Fig. 1). Thus, the high oxygen surface cov-
erage does not favor initial nitrous oxide formation. The
conversion of ammonia and oxygen proceeds now at the
same level; as expected no additional oxygen deposition is
measured. The main difference with the reduced platinum
sponge is the faster deactivation of the preoxidized cata-
lyst at temperatures below 413 K (Table 3). However, above
Fig. 9. Formation of N , N O, and H O measured by online mass spectrom-
2
2
2
etry in temperature-programmed NO experiment after ammonia oxidation
3
at 323 K (10 K/min, 0.5 vol% NO/He flow of 40 cm /min).
the gas phase. Nitrogen evolves at 398 K followed by the
water production with a peak maximum at 413 K. The N2O
production is much higher than in the TPD and TPO exper-
iments. The maximum of the N2O peak appears at 413 K,
higher than the maximum of the N2 peak. This was already
observed in the TPO spectrum and indicates that the N2O
production is dependent on the formation of N(a) from NH(a)
species and the availability of free sites for NO adsorption.
The increased N2O production can be explained by the re-
action of NH with O(a) or OH(a), which gave nitrogen and
water in the TPD experiment. First, atomic nitrogen is pro-
duced:
4
13 K also the preoxidized catalyst keeps a high activity and
selectivity toward nitrogen. Thus, the presence of oxygen at
the platinum surface does not cause a permanent deactiva-
tion of the catalyst. Above 413 K the catalyst is reduced by
ammonia.
The deactivating effect of unreacted oxygen was con-
1
5
firmed by the [ O]O2-labeled PEP experiments. In these
1
5
experiments [ O]O2 was first injected on reduced plat-
inum sponge before the ammonia oxidation was started. In
1
5
NH(a) + OH(a) → N(a) + H2O(a),
NH(a) + O(a) → N(a) + OH(a).
N2 starts to desorb and some vacant sites are created.
NO(g) will adsorb on the available sites which leads to the
production of N2O,
contrast to the [ O]O2 pulse in the reaction flow stream
Table 3
Initial deactivation times for the preoxidized platinum sponge catalyst and
15
the percentage of the O-labeled species remaining at the catalyst when
pulse was injected before the start of the ammonia oxidation reaction
1
5
N(a) + NO(a) → N2O(g) + 2∗.
Temperature
K)
Initial deactivation time
(min)
O preadsorbed and
(
not removed
(%)
Thus, as already stated there is a competition between
reactions toward N2 or N2O. The excess of NO(a) leads to
a higher production of N2O. The production of N2O and
N2 reaches a certain maximum; at 473 K both gases are
no longer produced. This indicates that the NO dissociation
323
0.2
0.8
5
11 ± 1
10 ± 1
12 ± 1
< 1
348
373
423
–
Fig. 10. Ammonia oxidation reaction performed at 373 K on a preoxidized platinum sponge catalyst. (A) concentration of N , N O and H O versus time,
2
2
2
−1
3
(
B) conversion of NH and O versus time (GHSV = 5600 h , NH /O = 2/1.5, flow = 46.5 cm /min).
3
2
3
2

474
D.P. Sobczyk et al. / Journal of Catalysis 225 (2004) 466–478
Fig. 11. Concentration of N and N O versus time for ammonia oxidation reaction at 348 K; during the reaction (A) NO pulses were injected, (B) N O pulses
2
2
2
−1
3
were injected (GHSV = 5600 h , NH /O = 2/1.5, flow = 46.5 cm /min).
3
2
Fig. 12. Concentration of N and N O versus time for ammonia oxidation reaction at 423 K; during the reaction (A) NO pulses were injected, (B) N O pulses
2
2
2
−1
3
were injected (GHSV = 5600 h , NH /O = 2/1.5, flow = 46.5 cm /min).
3
2
(
Table 2) now 10% of pulsed oxygen remains at the catalyst
coverage, the consecutive NO pulses are converted into N2
surface (Table 3). Thus, some oxygen species are formed at
the surface, which are unreactive with respect to water or
nitrous oxide. In a similar experiment the adsorbed [ O]O2
was placed under hydrogen flow (not shown), which resulted
in the formation of water, but also a significant amount of
2
and N O. At high surface coverage, when the catalyst is de-
activated, NO selectively forms N O, which is in line with
2
1
5
our earlier findings [12]:
NHx(a) + NO(a) → N2O(a) + H
,
x(a)
(
unreactive) oxygen remained adsorbed at the surface. The
Hx(a) + O(a) → OHx(a)
.
formation of oxygen islands on platinum provides a possible
explanation for this unreactive adsorbed oxygen phase.
This indirectly supports the interpretation of the XPS
measurement, showing that the NO species are not present at
the catalyst surface otherwise they would react toward N2O.
It should be noted that the water signal does not show any
concentration changes in all of the performed pulse experi-
ments. Fig. 11B shows that the N2O pulses are converted at
the beginning of the reaction selectively into nitrogen. With
the increasing surface coverage less nitrogen is formed and
N2O leaves unconverted the catalyst bed. Thus, at 348 K
N2O decomposes into nitrogen and probably atomic oxy-
gen with the restriction that the surface coverage is relatively
low. This decomposition of N2O is in line with the litera-
ture [45,46]; moreover, it is also reported that the rate of the
N2O decomposition is retarded by oxygen. In Fig. 12 simi-
lar NO and N2O pulse experiments are shown, however, at
a higher temperature (423 K) where the catalyst is still very
active and the surface coverage is low.
3
.5. NO and N2O pulse experiments
In this set of experiments the ammonia oxidation reac-
tion was carried out, wherein diluted NO (Fig. 11A) or N2O
pulses (Fig. 11B) were injected into the reaction stream.
These pulse experiments are compared with He pulses for
correct interpretation of the data, since the injection of a
pulse into the reaction stream causes a visible short decrease
in the product formation as the reaction flow is substantially
diluted at that moment. Thus, the comparison to what extent,
for example, nitrogen is diluted with NO and the He pulses
indicate if nitrogen is formed due to the NO pulse. Fig. 11A
shows NO pulses at 348 K. In the beginning of the reaction
NO adsorbs at the catalyst because the NO pulses are not de-
tected with MS. First, some N2 is formed (first two pulses),
however, also a part of NO remains adsorbed at the surface.
As the catalyst becomes deactivated with increasing surface
Fig. 12A shows production of nitrogen as a consequence
of the NO pulses. At this temperature and low surface cov-

D.P. Sobczyk et al. / Journal of Catalysis 225 (2004) 466–478
475
Fig. 13. Ammonia oxidation experiment with NO traces in the flow (NO/He
3
flow of 0.5 vol% of 10 cm /min) compared to the standard ammonia
−1
oxidation experiment without NO traces at 388 K (GHSV = 5600 h
,
3
NH /O = 2/1.5 flow = 46.5 cm /min).
3
2
erage the formation of nitrogen from NO is plausible, since
NO can dissociate or directly react with the NHx species.
Fig. 12B illustrates that N2O decomposes to nitrogen. How-
ever, if N2O is added continuously to the reaction stream
then the decomposition of N2O ends. This means that the
platinum catalyst is able to decompose only relatively low
amounts of N2O. The oxygen signal does not show any ad-
ditional formation of oxygen, which suggests that formed
atomic oxygen is consumed in a reaction with ammonia.
In order to investigate whether NO dissociates at plat-
inum below 413 K, a steady-state ammonia oxidation exper-
iment is performed with some of NO in the reactant stream
(
Fig. 13). Clearly, NO has a deactivating effect, as the initial
deactivation of the catalyst is much faster. This faster deacti-
vation is also observed at lower reaction temperatures. This
indicates that the decomposition of NO and/or reaction of
NO with the NHx species is much slower than the deposition
of the NHx species at the surface. With increasing surface
coverage more N2O is formed and even after the catalyst is
deactivated NO is fully converted into N2O (also observed
in the NO pulse experiments).
1
5
Fig. 14. PEP image of a pulse injection of [ O]O on Pt sponge
2
3
(
flow = 46.5 cm /min) (A) in 1 vol% NH /He flow at 323 K; (B) in 4 vol%
3
H /He flow at 323 K; (C) in 4 vol% H /He flow at 423 K.
2
2
3
.6. Desorption of water
on the platinum sponge shows a distinct desorption profile,
which is assigned to desorbing water. In this experiment be-
fore oxygen dissociation took place there were no adsorbed
nitrogen species at the surface, which is different compared
to the ammonia oxidation reaction.
As already noted, oxygen adsorbs dissociatively on plat-
inum. The oxygen atoms react with ammonia or hydrogen
toward water, the only oxygen-containing product. In both
experiments the water formation proceeds most probably via
the reaction of hydroxyls. This indicates for the ammonia
oxidation at low temperatures that the water formation via
hydroxyls is not the rate-determining step. The desorbing
water (low quantities) readsorbs on platinum and obviously
with increasing temperature water leaves the catalyst bed
faster (Fig. 14C). The recombination of OH also results in
the formation of very reactive O(a).
The PEP, XPS, and TPD experiments have indicated that
mainly nitrogen-containing species (NH2, NH) cause the de-
activation of the catalyst. The oxygen species, especially
formed hydroxyls, are not causing the profound deactivation
of the catalyst. This suggests that the production of water
through the reaction of two hydroxyls is much faster than
the endothermic reaction between NH and OH:
2
2
OH(a) ↔ H2O(g) + O(a) + ∗ (fast),
NH(a) + OH(a) → N2(g) + H2O(g) + 3
∗
(slow).
The formation of water via hydroxyls and the desorp-
tion of water at low temperatures is investigated with PEP in
1
5
[
O]O2 pulse experiments injected in a hydrogen or ammo-
1
5
nia flow. Figs. 14A and 14B show that a [ O]O2 pulse (very
low concentration) at 323 K in ammonia or hydrogen flow

476
D.P. Sobczyk et al. / Journal of Catalysis 225 (2004) 466–478
4
. Reaction mechanism
the formation of N(a) via NO(a) require the dissociation of ni-
tric oxide and therefore the recombination of N adatoms is a
more feasible option. Moreover, the formation of NO under
these conditions is apparently not favorable, because the de-
activated catalyst is mainly covered with NHx species (NH
and NH2), instead with NO. Thus, at low surface coverage
NO seems not to be the dominant species, probably due to
the low oxygen surface coverage, which disfavors the for-
mation of NO. It should be noted that the NO dissociation
temperature might be lower for our platinum catalyst due to
possible dislocation preferences and the resulting presence
of steps. The formation of some nitrogen from the NO pulses
in the ammonia/oxygen flow at 348 K demonstrates this.
The formation of N2O is not observed when the catalyst is
active and at low temperatures. The N2O pulse experiments
showed that small amounts of N2O decompose on platinum
to form nitrogen and atomic oxygen, and for that reason the
selectivity toward nitrogen is high.
The dissociation of ammonia is greatly enhanced by the
presence of atomic oxygen on the surface, which has been
reported earlier [12]:
∗
O2(g) + 2 → 2O(a),
∗
NH3(g) + ↔ NH3(a).
At the very beginning of the reaction hydrogen is stripped
from ammonia by dissociatively adsorbed oxygen forming
the NHx and OH species. These reactions are exothermic,
relatively fast, and proceed under all used conditions:
NH3(a) + O(a) → NH2(a) + OH(a),
NH2(a) + O(a) → NH(a) + OH(a),
NH(a) + O(a) → N(a) + OH(a).
Moreover as shown in Section 3.6 the recombination of
OH(a) toward water and active O(a) takes place at all studied
temperatures.
4
.1.2. Inactive regime
As the catalyst deactivates, the NHx species are formed
faster via ammonia adsorption than removed via nitrogen
formation, the selectivity of the catalyst changes, and N2O
starts to be formed. First, NO(a) needs to be formed or it was
already formed in the active stage:
4
4
.1. Temperatures below 398 K
.1.1. Active regime
XPS measurements and temperature-programmed exper-
iments have shown that after the deactivation of the catalyst
the surface of platinum is fully covered with NH and NH2
species. The PEP experiments also indicated that mainly
nitrogen-containingspecies cause the deactivation of the cat-
alyst. This means that the endothermic reactions between
NHx and OH are not proceeding very fast. The production of
water through the reaction of two hydroxyls is much faster:
∗
NH(a) + 2O(a) → NO(a) + OH(a) + ,
N(a) + O(a) → NO(a) + .
∗
NO(a) cannot desorb at these temperatures [7,8,50] and
therefore reacts toward N2O. The NO pulse experiments
show that N2O can also be formed via the reaction of NO
with NHx species:
2
OH(a) ↔ H2O + O(a) + ∗.
N(a) + NO(a) → N2O(g) + 2∗,
In this way formed active oxygen reacts instantly with NHx
species. Atomic nitrogen is not observed in XPS measure-
ments, probably because of the recombination of two N(a)
forms N2:
NHx(a) + NO(a) → N O + Hx(a)
2
(a)
.
Our TPO experiment (Fig. 8) showed that NO desorbs
from platinum from about 423 K, but only at high oxygen
surface coverage. In Fig. 1 a drastic decrease of nitrogen and
N2O formation is observed, which can be explained in terms
of the moving reaction front through the catalyst bed. As the
reaction zone arrives at the last positions, N2O cannot de-
compose any more since there is no fresh platinum surface
left. As the last positions are deactivated, the catalyst’s activ-
ity sharply decreases and the surface remains covered mainly
with NH and NH2. This is supported in the XPS N(1s) mea-
surement, but also indirectly by the NO pulse experiments.
A preoxidized catalyst deactivates much faster than the
reduced platinum sponge. Ammonia adsorption and dissoci-
ation are accelerated by the presence of oxygen. Thus, the
NHx species cover the platinum surface much faster. The
concentration profiles for nitrogen and nitrous oxide do not
change, which indicates that the reaction mechanism is not
changed for the preoxidized catalyst.
2
N(a) → N2 + 2∗.
The formation of atomic nitrogen could proceed via either
NHx(a) or NO(a). The reaction of NH with atomic oxygen is
exothermic and very fast. The reaction of NHx with OH is
less probable since for these reactions a higher energy bar-
rier needs to be overcome. The reaction of NH with OH is
also proposed by Van den Broek et al. [3] to be the rate-
determining step in the ammonia oxidation:
NH(a) + O(a) → N(a) + OH(a),
NH(a) + OH(a) → N(a) + H2O(a).
The formation of nitrogen via an NO(a) intermediate is
not favorable at low temperatures. At temperatures below
3
80 K it has been reported on Pt(100) that the dissociation
of NO is prohibited [8,47–49]. Both proposed reactions for

D.P. Sobczyk et al. / Journal of Catalysis 225 (2004) 466–478
477
4
.2. Temperatures above 398 K
At high surface coverage, when the catalyst deactivates,
the selectivity of the catalyst changes. Next to N2 also N2O
is formed. The intermediate NO(a) seems to be mainly in-
Deactivation of the catalyst is not observed at these tem-
peratures and only N2 and H2O are formed. The change
in the reaction mechanism is indicated by the reactivation
experiment (Fig. 2), in which N2O formation rapidly de-
creases, suggesting a change in the reaction mechanism de-
pending on the surface coverage. The N2O decrease cannot
only be explained by the decomposition of N2O at platinum,
as described in the N2O pulse experiments. Certainly, only
low amounts of N2O can be decomposed:
volved in the formation of a very low amount of N O. The
2
higher consumption of O than of NH suggests that oxygen
2
3
is involved in a relatively slow PtO formation. The initial
deactivation of the platinum catalyst is obviously a case of
self-poisoning. The fact that the catalyst is regenerated after
a reduction step with hydrogen to its initial activity supports
this.
Above 388 K nitrogen and water are formed and the cat-
alyst maintains its high initial activity. The NO pulse experi-
ments indicate that small quantities of NO can be selectively
converted to nitrogen. Also N2O decomposes to nitrogen,
which explains the high selectivity toward nitrogen. How-
ever, the reaction route for the nitrogen formation at higher
temperatures is still unclear.
N2O(a) → N2 + O(a).
Above 398 K the deposition of the NHx species on plat-
inum does not take place, as shown in the PEP experiments.
Thus, the NHx species are now much more reactive. How-
ever, also the NO pulse experiments showed that NO can
be selectively converted into nitrogen. The TPD, TPO, and
TP-NO experiments showed that above 423 K NO is present
at the surface leading to N2, N2O, or NO depending on the
surface coverage. To conclude, three reaction pathways are
responsible for the nitrogen formation, via the exothermic
reaction of NHx species with O(a), endothermic reactions of
NHx with OH, or via the NO and NHx reaction.
References
[
[
[
1] R.I. Heck, Catal. Today 53 (1999) 519.
2] N.I. Il’chenko, Russ. Chem. Rev. 45 (1976) 1119.
3] A.C.M. van den Broek, J. van Grondelle, R.A. van Santen, J. Catal. 185
(1999) 297.
[
[
4] L. Gang, B.G. Anderson, J. van Grondelle, R.A. van Santen, Catal.
Today 61 (2000) 179.
5] Y.M. Fogel, B.T. Nadykto, V.F. Rybalko, V.I. Shvachko, I.E. Korob-
chanskaya, Kinet. Katal. 5 (1964) 431.
5
. Conclusion
[
[
[
6] J.L. Gland, V.N. Korchak, J. Catal. 53 (1978) 9.
7] W.D. Mieher, W. Ho, Surf. Sci. 322 (1995) 151.
8] J.M. Bradley, A. Hopkinson, D.A. King, J. Phys. Chem. 99 (1995)
17032.
Below 388 K, two different product distributions in the
ammonia oxidation are observed, depending on the surface
coverage. At low coverage (catalyst is active) N2 and H2O
are selectively produced via stripping of hydrogen from am-
monia by oxygen atoms. Moreover as shown in Section 3.6
the recombination of OH(a) toward water and active O(a)
takes place at all studied temperatures. With PEP, it was
observed that already when the catalyst is still active, ad-
sorbed nitrogen species are formed which stay irreversibly
adsorbed at the catalyst surface and do not form any gaseous
products. These nitrogen species mainly cause the poison-
ing of the platinum catalyst, thus retarding the adsorption
of reactants. XPS and temperature-programmedexperiments
identified those as NH2(a) and NH(a). The oxygen species do
not poison the catalyst, unless oxygen is preadsorbed on the
platinum. The main reason for the deactivation of the plat-
inum catalyst is the fact that the NHx(a) + OH(a) reaction
is much slower than formation of water via hydroxyls. In
this way the NH and NH2 remain unreactive at the surface
and they block the active sites. The nitrogen formation pro-
ceeds via the recombination of atomic nitrogen and not via
an NO(a) intermediate. Low surface coverage does not favor
NO(a) formation and below 400 K the dissociation of NO
is prohibited. Moreover, the addition of NO to the reaction
flow causes faster deactivation of the platinum sponge. The
preoxidized platinum sponge deactivates faster than the plat-
inum sponge due to surface poisoning caused partly by the
nitrogen species and partly by unreactive oxygen.
[9] V.A. Matyshak, E. Lefler, K.Kh. Schnabel, Kinet. Katal. 28 (1987)
1199.
[
[
[
10] J.J. Ostermaier, J.R. Katzer, W.H. Manogue, J. Catal. 33 (1973) 457.
11] J.J. Ostermaier, J.R. Katzer, W.H. Manogue, J. Catal. 41 (1976) 277.
12] D.P. Sobczyk, A.M. de Jong, E.J.M. Hensen, R.A. van Santen,
J. Catal. 219 (2003) 156.
13] P.R. Norton, J. Catal. 36 (1975) 211.
14] N. Wagner, O. Brümmer, Krist. Tech. 13 (1978) K71.
15] J.L. Gland, Surf. Sci. 93 (1980) 487.
16] H. He, N. Swami, B.E. Koel, J. Chem. Phys. 110 (1999) 1173.
17] B. Laurßen, S. Günther, H. Marbach, M. Kiskinova, J. Janek, R. Im-
bihl, Chem. Phys. Lett. 316 (2000) 331.
18] W. Vaalburg, W.A.A. Kamphuis, H.D. Beerling van der Molen,
S. Reiffers, A. Rijskamp, M.G. Woldring, Int. J. Appl. Radiat. Isot. 26
[
[
[
[
[
[
(1975) 316.
[
19] K. Suzuki, Y. Yoshida, Appl. Radiat. Isot. 50 (1999) 497.
[20] D.P. Sobczyk, J. van Grondelle, A.M. de Jong, M.J.A. de Voigt, R.A.
van Santen, Appl. Radiat. Isot. 57 (2002) 201.
21] M.J. Welch, M.R. Kilbourn, J. Labelled Compd. Radiopharm. 22
1985) 1193.
22] B.M. Palmer, M. Sajjad, D.A. Rottenberg, Nucl. Med. Biol. 22 (1995)
41.
23] A.V.G. Mangnus, L.J. van IJzendoorn, J.J.M. Goeij, R.H. Cunning-
ham, R.A. van Santen, M.J. de Voigt, Nucl. Instrum. Methods Phys.
Res. B 99 (1995) 649.
[
[
[
(
2
[
24] B.G. Anderson, R.A. van Santen, A.M. de Jong, Top. Catal. 8 (1999)
125.
[
25] A. Bogicevic, J. Strömquist, B.I. Lundqvist, Phys. Rev. B 57 (1998)
R4289.
[26] P.J. Feibelman, Phys. Rev. B 56 (1997) 10532.

478
D.P. Sobczyk et al. / Journal of Catalysis 225 (2004) 466–478
[
[
[
[
27] M. Lynch, P. Hu, Surf. Sci. 458 (2000) 1.
[38] R. Zehr, A. Solodukhin, B.C. Haynie, C. French, I. Harrison, J. Phys.
28] Ž. Šlijvan cˇ anin, B. Hammer, Surf. Sci. 515 (2002) 235.
29] A. Fahmi, R.A. van Santen, Z. Phys. Chem. 197 (1996) 203.
30] F. Illas, N. López, M. García-Hernández, P.R. de Moreira, J. Mol.
Struct. 458 (1999) 93.
Chem. B 104 (2000) 3094.
[39] J. Kiss, D. Lennon, S.K. Jo, J.M. White, J. Phys. Chem. 95 (1991)
8054.
[40] T.S. Amorelli, A.F. Carley, M.K. Rajumon, M.W. Roberts, P.B. Wells,
Surf. Sci. 315 (1994) L990.
[41] N.R. Avery, Surf. Sci. 131 (1983) 501.
[42] J.H. Campbell, C. Bater, W.G. Durrer, J.H. Craig Jr., Surf. Sci. 380
(1997) 17.
[43] R.J. Gorte, J.L. Gland, Surf. Sci. 102 (1981) 348.
[44] D. Burgess Jr., D.S. King, R.R. Cavangh, J. Vac. Sci. Technol. A 5
(1987) 2959.
[45] C.G. Takoudis, L.D. Schmidt, J. Catal. 80 (1983) 274.
[46] H.G. Lintz, L. Riekert, J. Catal. 88 (1984) 244.
[47] S.J. Lombardo, F. Esch, R. Imbihl, Surf. Sci. Lett. 271 (1992)
L367.
[
[
[
31] M. Matsumo, N. Tatsumi, K. Fukutani, T. Okano, Surf. Sci. 513 (2002)
485.
32] H. Aizawa, Y. Morikawa, S. Tsuneyuki, K. Fukutani, T. Ohni, Surf.
Sci. 514 (2002) 394.
33] A.C.M. van den Broek, Low-Temperature Oxidation of Ammonia over
Platinum and Iridium Catalysts, Universiteitsdrukkerij, Eindhoven,
1
998, p. 60.
34] Y.M. Sun, D. Sloan, H. Ihm, J.M. White, J. Vac. Sci. Technol. A 14
1996) 1516.
35] S. Sugai, K. Takeuchi, T. Ban, H. Miki, K. Kawasaki, Surf. Sci. 282
1993) 67.
[
[
[
(
(
[48] S.J. Lombardo, M. Slinko, T. Fink, T. Löher, H.H. Madden, F. Esch,
36] T. Sugisawa, J. Shiraishi, D. Machihara, K. Irokawa, H. Miki, C. Ko-
dama, T. Kuriyama, T. Kubo, H. Nozoye, Appl. Surf. Sci. 169/170
R. Imbihl, G. Ertl, Surf. Sci. 269/270 (1992) 481.
[
[
49] G. Veser, F. Esch, R. Imbihl, Catal. Lett. 13 (1992) 371.
50] J.J. Vajo, W. Tsai, W.H. Weinberg, J. Phys. Chem. 89 (1985) 3243.
(
2001) 292.
[
37] G. Szulczewski, R.J. Lewis, J. Chem. Phys. 103 (1995) 10238.