Selective oxidation of ammonia over Ru(0 0 0 1)

Article abstract of DOI:10.1016/j.susc.2004.02.022

The decomposition and oxidation of NH₃ have been studied on the Ru(0001) surface in the temperature range from 150 up to 800 K. The results were compared to those found for Ir(110) and Ir(510). TDS results showed that most of the ammonia is dissociatively adsorbed between 150 and 300 K, with formation of H₂ around 300 K and N₂ between 600 and 800 K. N ₂ desorption shifts to lower temperatures with increasing surface oxygen coverage. The products of ammonia oxidation observed were N₂, H₂O and N₂O. Formation of NO was not found. Inhibition of the reaction presumably by N species was observed until 450 and 670 K, depending on the NH₃/O₂ ratios. Above those temperatures the reaction started as manifested by a decrease in the NH₃ and O ₂ pressures and a simultaneous increase in the H₂O, N ₂ and N₂O pressures.

Full text of DOI:10.1016/j.susc.2004.02.022

Surface Science 555 (2004) 83–93
www.elsevier.com/locate/susc
Selective oxidation of ammonia over Ru(0 0 0 1)
a
a,b
S.A.C. Carabineiro , A.V. Matveev a,b, V.V. Gorodetskii
,
a,
*
B.E. Nieuwenhuys
a
Leiden University, Leiden Institute of Chemistry, Surface Science and Heterogeneous Catalysis, Einsteinweg 55, 2333 CC Leiden,
The Netherlands
b
Boreskov Institute of Catalysis, Russian Academy of Sciences, Novosibirsk 630090, Russia
Received 28 November 2003; accepted for publication 6 February 2004
Abstract
The decomposition and oxidation of NH3 have been studied on the Ru(0 0 0 1) surface in the temperature range from
150 up to 800 K. The results were compared to those found for Ir(1 1 0) and Ir(5 1 0). TDS results showed that most of
the ammonia is dissociatively adsorbed between 150 and 300 K, with formation of H2 around 300 K and N2 between
600 and 800 K. N2 desorption shifts to lower temperatures with increasing surface oxygen coverage. The products of
ammonia oxidation observed were N2, H2O and N2O. Formation of NO was not found. Inhibition of the reaction
presumably by N species was observed until 450 and 670 K, depending on the NH3/O2 ratios. Above those temperatures
the reaction started as manifested by a decrease in the NH3 and O2 pressures and a simultaneous increase in the H2O,
N2 and N2O pressures.
ꢀ 2004 Elsevier B.V. All rights reserved.
Keywords: Ammonia; Ruthenium; Oxygen; Nitrogen oxides; Nitrogen molecule; Catalysis; Chemisorption; Surface chemical reaction;
Thermal desorption; Auger electron spectroscopy; Low energy electron diffraction (LEED)
1. Introduction
Reaction (1) is the so-called Ostwald process used
to produce nitric acid. High temperatures (>800
K) are needed for this reaction. It has been studied
intensively in the past because of its industrial
importance.
Relatively few studies were focused on reactions
(2) and (3). However, the interest in reaction (2) is
increasing due to its environmental aspect [5].
Reaction (3) might become a relevant reaction for
N2O production. Recently it has been shown that
this molecule may have a future as oxidant in
selective catalytic oxidation [6–8].
The oxidation of ammonia can proceed via
three main overall reactions [1–4]:
4NH3 þ 5O2 ! 4NO þ 6H2O
4NH3 þ 3O2 ! 2N2 þ 6H2O
4NH3 þ 4O2 ! 2N2O þ 6H2O
ð1Þ
ð2Þ
ð3Þ
* Corresponding author. Tel.: +31-71-527-4545; fax: +31-71-
527-4451.
E-mail address: b.nieuwe@chem.leidenuniv.nl (B.E. Nie-
uwenhuys).
Most of the fundamental studies concerning
ammonia oxidation have been performed in con-
nection with the production of NO on Pt and
0039-6028/$ - see front matter ꢀ 2004 Elsevier B.V. All rights reserved.
doi:10.1016/j.susc.2004.02.022

84
S.A.C. Carabineiro et al. / Surface Science 555 (2004) 83–93
Pt/Rh surfaces [2,3,9–16]. Some studies were also
carried out on Ir in relation to the environmental
aspect [4,17–22].
energy ¼ 1.5 kV) and flashing in UHV to 1400 K.
Surface cleanliness was checked by LEED and
AES. As for the detection of carbon impurities,
since Auger peaks of carbon (C272 eV) and ruthe-
nium (Ru273 eV) overlap, the cleaning procedure
was repeated until no change in the peak at that
energy was detected. In addition, the absence of
CO and CO2 formation upon heating the surface
in an oxygen atmosphere confirmed the absence of
carbon impurities. The Arþ ion sputtering and
flashing treatments were repeated at the beginning
of each series of experiments, and the surface
cleanliness and structure were checked by AES
and LEED. The characteristic LEED pattern of
the (0 0 0 1) surface was clearly observed once the
surface was clean.
The adsorption of ammonia has also been
studied on several metal surfaces [20,23–33]
including those of Pt [16,31,34–39]. Some studies
were also carried out on Ru surfaces [40–45].
Ammonia can adsorb molecularly [24–31,33,40–
42] or dissociatively [24,25,34,38,43–45]. Molecu-
lar adsorption on a clean surface results in a bond
to the metal via the N atom with H pointing away
from the surface. It has been reported that, in the
presence of preadsorbed oxygen, NH3 interacts
with adsorbed O via a hydrogen bond, leading to
local azimuthal ordering in the absence of long
range order [26].
The present work is part of a collaborative
project in which ammonia oxidation is studied on
several noble metal catalysts, both polycrystalline
[4,5,46,47] and in the form of single crystal sur-
faces. In addition, theoretical calculations are
performed. The ultimate purpose is to understand
the selectivity and the specific differences in
behaviour of the various noble metal surfaces. Our
earlier papers dealt with this reaction on Ir sur-
faces [21,22]. In the present work the reaction was
studied on Ru(0 0 0 1).
High purity gases were used without further
purification. These gases were dosed by means of
three leak valves from the dosing system into the
main chamber. The pressure was read on an ion
gauge and pressure readings were corrected using
relative sensitivities for NH3 and O2 to N2 of 1.2
and 1.0, respectively. In case of dosing NH3 and
O2 together, the pressure of NH3 was stabilized at
1 · 10À6 mbar for 15 min, before O2 was admitted
in the system to the pressure desired.
During the reactions, the crystal was turned in
front of a small opening (diameter of about 1 mm),
which gave access to the quadrupole mass spec-
trometer (QMS) chamber. The distance between
the QMS hole and the crystal was typically 0.2
mm. In this way, the sensitivity was increased and
the contributions from the Ta sample holder,
crystal edges and chamber walls were minimized.
When the crystal was turned away from the QMS
hole, the pressure readings of the QMS returned to
the noise level indicating that the observed phe-
nomena were caused solely by the reaction over
the crystal. Reactions were performed in the flow
mode using a turbomolecular pump. TDS mea-
surements were performed by heating the sample
covered with an adsorbed layer using a constant
heating rate. The desorbing species were moni-
tored by the differentially pumped QMS.
2. Experimental
Experiments were performed in a UHV system
equipped with facilities for LEED, AES, and a
differentially pumped quadrupole mass spectro-
meter. The system was pumped by a turbo
molecular pump (170 l/s) and an ion pump (150 l/
s). The base pressure was always better than
4 · 10À10 mbar.
The Ru sample was cut from a Ru single crystal
by spark erosion to within 0.5ꢁ of the desired
direction and polished down to a grain size of 1
lm. The crystal was spotwelded to a Ta support
and could be heated resistively up to 1400 K. The
temperature was measured using a Pt–Pt/Rh
thermocouple, which was spotwelded to the back
of the crystal. The crystal was cleaned by repetitive
cycles of heating in an oxygen and hydrogen
atmosphere (p(O2 or H2) ¼ 10À6 mbar) and by Arþ
ion sputtering (p(Ar) ¼ 1 · 10À5 mbar, incident
Kinetic experiments were carried out in order to
measure the reaction orders with respect to NH3
and O2. The pressure of either NH3 or O2 was varied
while the pressure of the other gas was kept con-

S.A.C. Carabineiro et al. / Surface Science 555 (2004) 83–93
85
stant and the relative reaction rates determined at
.
.
.
.
.
.
.
.
(a)
steady state conditions. This was done between 550
and 750 K, the temperatures at which reaction took
place (as shown in the heating and cooling cycles).
3. Results
3.1. Adsorption of NH3 on the clean Ru(0 0 0 1)
surface
Fig. 1 shows the results of thermal desorption
spectroscopy (TDS) experiments carried out after
NH3 exposure at 300 K. Spectra obtained at
higher exposures have been shifted to higher val-
ues for better comparison. The only desorption
products observed were H2 (mass 2) and N2
(masses 28 and 14). For N2, the fragment m=e ¼ 14
was monitored in addition to 28, in order to avoid
possible ambiguity with CO.
(b)
Fig. 1a shows that, at low exposures, two dif-
ferent peaks of H2 can be observed (about 360 and
415 K). For pure hydrogen similar peaks have
been reported that have been attributed to
adsorption on hcp and fcc-sites of the Ru(0 0 0 1)
surface [48,49]. At NH3 exposures higher than 4 L,
those two peaks condensate in one peak (around
390 K). A similar peak was obtained when the
surface was exposed to pure H2. Evolution of
the total and individual peak areas is shown in the
insert of Fig. 1, obtained with division of the TD-
curves into two gauss peaks. The peaks area was
calibrated by TPD observed following exposure to
the pure hydrogen at 150 K up to saturation,
assuming that its area corresponds to about 1.0
ML. For exposures higher than 16 L, both the
total peak area and individual areas of the H2
peaks do not change significantly, showing that all
adsorption sites are occupied.
Fig. 1. TDS experiments carried out after NH3 exposure at 300
K. Desorption as H2 (a) and N2 (b) are shown (heating rate of
25 K/s). The insert in (a) shows the evolution of the individual
areas of the H2 peaks (detected at 360 and 415 K, respectively)
and total H2 peak area, with NH3 exposure.
The TD spectra of N2 (Fig. 1b) show one peak
around 680 K. That has been attributed to N-
adsorbed on hcp-sites [50]. The peak maximum
shifts to higher temperatures with increasing
exposure. This shift to higher temperatures is in
contrast with the results obtained by other authors
who reported a shift to lower temperatures [43].
The ratio of the peak areas of N2 and H2 is 3:1,
taking in account the different ionization factors.
If we accept the assignment of the H and the N-
desorption peaks to adsorption on hcp and fcc-
sites (from Refs. [48–50]) then we can conclude
that the nitrogen coverage, under saturation con-
ditions, is hhNcp ꢀ 0:25 ML, and hydrogen coverage
hhHcp ꢀ 0:25 ML and hHfcc ꢀ 0:5 ML. The hhHcp=hHfcc
ratio is 0.5, which is the value of the ratio of the
areas of low and higher-temperature hydrogen
desorption peaks, seen in the insert of Fig. 1a.

86
S.A.C. Carabineiro et al. / Surface Science 555 (2004) 83–93
Desorption as NH3 was not observed following
(a)
(b)
exposure at 300 K. Molecular adsorption of NH3
was observed on Ru(0 0 0 1) by other authors at
lower temperatures (around 100 K) [40,42]. In fact,
some desorption as NH3 was found following
exposure at 150 K. Fig. 2 shows the resulting TD
spectra. The small and broad NH3 desorption
peak at 195 K increases with increasing exposure
and saturates at 5 L. The TD spectra of H2 and N2
are consistent with those observed following
exposure at 300 K. H2 desorbs at 395 K (Fig. 2b),
and N2 at 620 K after 0.5 L exposure (Fig. 2c),
shifting to higher temperature (760 K, for 10 L
exposure). The intensity of the N2 desorption peak
was larger by one order of magnitude than that
of NH3. These results show that desorption of
molecular ammonia is a minor process under our
conditions and only occurs at lower temperatures
(around 150 K). The major processes are decom-
position of ammonia and subsequent formation of
N2 and H2. Similar results were also reported by
other authors [24,25,34,38,43–45].
3.2. Effect of oxygen precoverage on NH3 adsorp-
tion
(c)
Fig. 3 shows the TD spectra obtained after
exposing the sample to 5 L NH3 at 300 K, on a
surface precovered with different amounts of
oxygen, up to about 0.5 ML, as was determined by
our AES measurements and calibrated using lit-
erature data [51–55]. In the presence of oxygen, the
N2 desorption peak (Fig. 3a) is shifted from
approximately 710 to 540 K. The spectra practi-
cally do not change for oxygen exposures higher
than 10 L (0.25–0.3 ML). Also the size of the peak
does not vary with increasing oxygen precoverage,
that is, the amount of N2 desorbing from an O-
saturated surface is almost similar to that from
clean Ru(0 0 0 1). However, the large shift to lower
temperatures suggests that the N-atoms are
strongly affected by the presence of Oads. It points
to a large reduction of the activation energy of
nitrogen desorption, which is in agreement with
the results obtained on other Pt-group metal sur-
faces including iridium [21,22]. Huang et al. [43]
found an additional nitrogen desorption peak at
620 K on a Ru(0 0 0 1) surface covered with a 0.5
Fig. 2. TDS experiments carried out following various expo-
sures to NH3 at 150 K. Desorption as NH3 (a), H2 (b) and N2
(c) are shown (heating rate of 25 K/s).
ML of oxygen, in addition to the nitrogen
desorption peak at 700 K. The reason for the ob-
served differences, in comparison with our results,
is related to the preheating of the oxygen layer up
to 600 K, used by these authors [43].
Fig. 3b shows that hydrogen desorption is
strongly suppressed with increasing O-precover-

S.A.C. Carabineiro et al. / Surface Science 555 (2004) 83–93
87
(a)
(a)
(b)
(b)
(c)
Fig. 3. Effect of preadsorbed oxygen. TDS experiments carried
out after 5 L NH3 exposure at 300 K, on a surface exposed to
various amounts of O2. Desorption as N2 (a) and H2 (b) are
shown (heating rate of 25 K/s).
age, and is not observed anymore when the surface
is pre-exposed to 2 L O2 or higher. This decrease
is, most probably, due to an almost immediate
desorption of formed H2 (the same was observed
at 150 K as will be shown later on).
Fig. 4. Effect of preadsorbed oxygen. TDS experiments carried
out after 5 L NH3 exposure at 150 K, on a surface exposed to
various amounts of O2. Desorption as NH3 (a), H2 (b) and N2
(c) are shown (heating rate of 25 K/s).
At 150 K, the intensity of the NH3 peak de-
creases with O-precoverage (Fig. 4a), whereas that
of the nitrogen desorption peak is almost not af-
fected (Fig. 4c). It means that reactivity of the
oxygen precovered surface (to what concerns
the decomposition of NH3) is higher than that of
the clean surface, that is, less molecular ammonia
remains on the O-covered surface. Therefore, some
beneficial effect of Oads on NH3 decomposition was
found on Ru(0 0 0 1), as was reported by other
authors on Pt(1 0 0) [31,56], Ir(1 0 0) [20], Cu(1 1 0)
[33] and for Ru(0 0 0 1) at 480 K [43]. That is
however in contrast with results obtained for

88
S.A.C. Carabineiro et al. / Surface Science 555 (2004) 83–93
Ir(1 1 0) [21] and Ir(5 1 0) [22], where no beneficial
effect was observed.
The hydrogen peak also decreases with O-pre-
coverage at 150 K (Fig. 4b), and it also ‘‘splits’’
into two peaks that become less intense with
increasing pre-exposure to oxygen. H2O desorp-
tion was not observed. The N2 peak (Fig. 4c) shifts
to lower temperatures with increasing O-preco-
verage, as observed following exposure at room
temperature. That again suggests that there is no
reaction between oxygen and hydrogen, otherwise,
most likely, one would observe
a constant
desorption temperature for N2, until a critical
coverage of oxygen was achieved and all the H2
would react away.
No desorption of NO was observed, in contrast
with the results obtained by other authors on a
Pt(1 1 1) surface [16]. It should be noted that for-
mation of NO is more likely to occur at higher
temperatures [3,4,15,57].
3.3. Heating and cooling cycles in the presence of
NH3 and O2 in the gas phase
Fig. 5 shows the results of heating and cooling
cycles from 300 to 800 K performed in a flow of
NH3 at a constant pressure of 1 · 10À6 mbar, in the
absence and presence of O2. These results dem-
onstrate that, upon increasing the O2/NH3 ratio,
the temperature at which the reaction starts de-
creases. The NH3 conversion obtained varies, un-
der our experimental conditions, from 4% (for a
ratio of 1.2) to 38% (for a ratio of 15). The con-
version values were determined by the drop in
NH3 pressure observed in the heating branch (Fig.
5). These results are similar to those obtained for
Ir(1 1 0) [21]. In contrast, for Ir(1 1 0), where no
NH3 conversion was observed in the absence of
oxygen [21], a conversion of around 3.5% was
observed for Ru(0 0 0 1), without oxygen, as shown
in Fig. 5. This surface is thus capable to decom-
pose NH3 in the absence of an oxidizing gas.
AES experiments were carried out during
heating and cooling cycles to examine what was
present on the surface. However, due to equipment
limitations, that had to be done in the 10À8 mbar
range. Under those conditions no significant build-
up of N or O on the surface was found. However,
Fig. 5. Heating and cooling cycles performed in 1 · 10À6 mbar
NH3, for several O2/NH3 ratios (R) at a heating rate of 20 K/
min. Change in pressure of NH3 is shown.
the strong dependence of the temperature at which
the reaction starts on the O2/NH3 ratio, shown in
Fig. 5, suggests that there is inhibition of the
reaction until a certain temperature (between 450
and 670 K, depending on the O2/NH3 ratio). This
can be caused by adsorbed N- or O-species.
However, it is more likely to be N-caused, since the
inhibition ends at the desorption temperature of
nitrogen, and oxygen desorbs at much higher
temperatures [58]. It is interesting to note that

S.A.C. Carabineiro et al. / Surface Science 555 (2004) 83–93
89
there is a significant hysteresis, with a higher NH3
conversion in the heating branch then in the
cooling branch.
Fig. 6 shows changes in pressure of O2 during
heating and cooling cycles. Increasing the O2/NH3
ratio results in a decrease in the temperature at
which the reaction starts, which is consistent with
what was shown for NH3 in Fig. 5. These results
again suggest that the inhibitive species are N-
species and not O-species.
The reaction products are N2, H2O and N2O, as
observed for Ir(1 1 0) [21]. Fig. 7 shows the heating
and cooling cycles for all the reactants and products
for a O2/NH3 ratio of 2.4. It can be seen that the
sharp increase in the H2O, N2 and N2O pressures at
the temperature of 615 K coincides with a sharp
decrease in the NH3 and O2 pressures. The con-
version was approximately 10%. The N2O:H2O:N2
ratio was 1:1.3:5.4. With increasing O2/NH3 ratio,
the conversion increased, as stated before. How-
ever, the ratio between the products did not vary
much. A small increase in the percentage of N2,
compared to N2O and H2O was obtained (thus for
a O2/NH3 ratio of 15, the relationship between
N2O:H2O:N2 was 1:1.2:6.6). Obviously, the reac-
tion starts at that temperature, for that ratio, and
overcomes the inhibition present until then, possi-
ble by N-species, as described earlier. A similar
behaviour was observed for other ratios. No for-
mation of NO was detected, in contrast to results
obtained by other authors on different surfaces
[3,5,12–14,16,59,60]. Again, as can be seen in the
figure, it is the heating branch of the hysteresis at
Fig. 7. Heating and cooling cycles performed in 1 · 10À6 mbar
NH3, for a O2/NH3 ratio of 2.4 at a heating rate of 20 K/min.
For a better view the concentrations are not scaled. The
product relationship between N2O:H2O:N2 was 1:1.2:6.6.
Fig. 6. Heating and cooling cycles performed in 1 · 10À6 mbar
NH3, for several O2/NH3 ratios (R) at a heating rate of 20 K/
min. Change in pressure of O2 is shown.

90
S.A.C. Carabineiro et al. / Surface Science 555 (2004) 83–93
which the surface is most active, in contrast to re-
sults obtained on Ir [21,22]. A possible explanation
for the observed low activity in the cooling branch
of the hysteresis is that for temperatures higher
than 800 K all the nitrogen is desorbed from the
surface. In the cooling branch, the rate of O
adsorption may exceed that of its removal by NH3,
resulting in O-build-up and inhibition of the reac-
tion.
It was not possible to examine the surface
structure during the reactions by LEED, since the
pressure used (10À6 mbar range) is too high. The
original pattern characteristic of the (0 0 0 1) sur-
face structure was, however, observed before and
after the reaction (in the latter case, after evacua-
tion of the gases).
surface following exposure at 150 and 300 K dis-
sociates upon heating to nitrogen and hydrogen.
A possible interpretation for the disappearance
of hydrogen desorption peaks as a result of pre-
exposure to oxygen (Figs. 3 and 4) is that adsorbed
oxygen reacts with hydrogen and forms water.
However, desorption of water was not found. Also
Hrbek could not find any reaction between oxygen
and hydrogen [67,68]. Thus, this decrease is, most
probably, due to almost immediate desorption of
formed H2 (since the same was observed at 150 K).
The disappearance of the hydrogen peak can be
caused by the same effect explained above, that is,
the reduction of the Edes in the presence of oxygen.
It is known that the presence of oxygen atoms on
the Ru surface results in a shift in the desorption
temperature from 300 K to less than 150 K [53,69].
Assuming that an adsorbed atom of oxygen re-
duces the heat of adsorption of only six adjacent
hydrogen atoms, then all the hydrogen should
desorb at an oxygen coverage of 0.25 ML. This
value is in good agreement with the value of 0.25–
0.3 ML, that corresponds to the adsorption of 10
L oxygen.
As shown in Section 3.3, in the heating branch,
the reaction starts at temperatures at which
nitrogen desorption begins. This suggests that the
rate limiting process in the lower temperature re-
gion is nitrogen desorption. Figs. 5 and 6 show
that an increase of the O2/NH3 ratio results in a
shift of the active branch in the heating curve to
lower temperatures. Obviously, this effect is con-
nected with reduction of heat of nitrogen adsorp-
tion on metal surfaces as the amount of adsorbed
oxygen increases.
3.4. Kinetic experiments
The reaction orders with respect to NH3 and O2
were determined between 550 and 750 K. For that
purpose, the oxygen pressure was kept constant at
5 · 10À8 mbar, while the NH3 pressure was varied
between 1 · 10À8 and 1 · 10À5 mbar, and the NH3
pressure was kept constant at 5 · 10À8 mbar, while
the oxygen pressure was varied from 1 · 10À8 to
1 · 10À5 mbar, respectively. In this temperature
range, the reaction order with respect to NH3 is
around 0.4 and with respect to O2, around 0.5.
Similar values were obtained on Ir surfaces [21,22]
and by other authors on supported Pt [46]. At
much higher pressures, for supported catalysts,
reaction orders close to zero have been reported
[46]. The positive order of reaction suggests there
is no significant inhibition of the reaction by N- or
O-species in that temperature range. However, as
discussed above, that is only true above 450 and
670 K, depending on the O2/NH3 ratio.
At higher temperatures, there is a competition
between ammonia conversion and ammonia
desorption [62]. Increasing O2/NH3 ratio results in
an increase in product formation, i.e. ammonia
conversion. This apparently suggests that for
reaction in flow condition, as for our coadsorption
experiments, oxygen covered Ru(0 0 0 1) surface is
more reactive for ammonia conversion, compared
with a clean Ru(0 0 0 1) surface.
4. General discussion
In several papers [40,42,61–63,65,66], it has
been reported that ammonia dissociation on
Ru(0 0 0 1) surface is negligible at temperatures
lower than 300–400 K. Our results, as well as re-
sults from other authors [43,64] clearly proved that
most of the ammonia adsorbed on the Ru(0 0 0 1)
The following processes describe the ammonia
and oxygen adsorption:
NH3ðgÞ $ NH3ads
ð4Þ

S.A.C. Carabineiro et al. / Surface Science 555 (2004) 83–93
91
O2ðgÞ ! 2Oads
ð5Þ
NH3ads þ OHads ! NH2ads þ H2Oads
ð16Þ
ð17Þ
ð18Þ
King and co-workers [15] suggested that there are
two routes for N2 production on Pt(1 0 0): above
350 K N2 is formed mainly from NO dissociation,
and below 350 K when NO dissociation is not
possible, N2 comes from NH3 molecules that have
been stripped of hydrogen by oxygen adatoms.
Hence as NH3 molecules are stripped, they are
NH2ads þ OHads ! NHads þ H2Oads
NHads þ OHads ! Nads þ H2Oads
Then N2 can be formed as in reaction (11). An-
other alternative for N2 formation according to
the same authors [4], would be from NHXads
(X ¼ 1 or 2) via a N2HY (Y ¼ 2–4):
converted to either N adatoms or NOads
:
NHXads þ NHY ads ! N2HXþY ads
ð19Þ
NH3ads þ Oads ! NHads þ OHads þ Hads
ð6Þ
ð7Þ
ð8Þ
N2HXþY ads þ ðX þ Y ÞOHads
! N2ðgÞ þ ðX þ Y ÞH2O
These authors found that most of the nitrogen on
the surface was NH and concluded that the last
dehydrogenation step 18 (NH with OH to N and
water) is the rate determining step [4].
Alternatively, it might also be that under our
experimental conditions, N2 and H2O are formed
via:
NHads þ Oads ! Nads þ OHads
ð20Þ
NHads þ 2Oads ! NOads þ OHads
Higher O coverages favor steps 7 and 8.
OHads þ Hads ! H2O
Above 350 K NOads can dissociate:
ð9Þ
ðgÞ
NOads ! Nads þ Oads
ð10Þ
ð11Þ
2Nads ! N2ðgÞ
This step competes with NO desorption, so at
higher temperatures (above 0.2 ML), NO
formed [15,70]:
NH3ads ! Nads þ 3Hads
(via NHX with X ¼ 1 and 2), followed by:
2Nads ! N2ðgÞ
and:
ð21Þ
ð22Þ
ð23Þ
is
ðgÞ
NOads ! NO
ð12Þ
ðgÞ
Oads þ 2Hads ! H2O
ðgÞ
in preference to N2ðgÞ (steps 10 and 11).
The present results cannot discriminate between
the various mechanisms. Recent XPS results [71]
showed three component peaks (possibly due to
adsorbed NH3, NH2, NH and/or N species). Since
dissociation of NH3 takes place at room temper-
ature, shown by formation of H2, it is possible that
the main species on the surface is NH, since other
authors reported that NH is the main species
present on the catalyst [4].
According to the same authors, N2 production
is suppressed at high oxygen coverages that is, step
10 will not occur. Other authors reported that
ammonia oxidation on silver is a two step reaction,
where ammonia is first oxidized to NO and NOads
further reacts with NHX to produce N2 [5]. In our
case no formation of gaseous NO was observed,
only N2. It is however possible that there may be
formation of NOads that readily dissociates into
N2.
Some authors suggested that the reaction mech-
anism must proceed through a stepwise dehydro-
genation of the ammonia molecule [4]:
5. Conclusions
TDS results show that following exposure at
both 150 and 300 K, decomposition of ammonia is
the dominant process, with formation of N2 and
H2. At 150 K some molecular adsorption was
found. A decrease in the nitrogen and (apparently)
hydrogen desorption temperatures with increasing
NH3ads þ Oads ! NH2ads þ OHads
NH2ads þ Oads ! NHads þ OHads
ð13Þ
ð14Þ
NHads þ Oads ! Nads þ OHads
ð15Þ

92
S.A.C. Carabineiro et al. / Surface Science 555 (2004) 83–93
[18] G. Papapolymerou, V. Bontozoglou, J. Mol. Catal. 120
(1997) 165.
surface oxygen coverage was found. The products
obtained by ammonia oxidation are N2, H2O and
N2O. No formation of NO was observed. Inhibi-
tion of the reaction, presumably by N species,
was observed until 450 and 670 K, depending on
the NH3/O2 ratios. At higher temperatures,
the decomposition of ammonia seems to limit the
reaction.
[19] A.K. Santra, B.K. Min, C.W. Yi, K. Luo, T.V. Choudh-
ary, D.W. Goodman, J. Catal. Lett. 77 (2001) 1.
[20] A.K. Santra, B.K. Min, C.W. Yi, K. Luo, T.V. Choudh-
ary, D.W. Goodman, J. Phys. Chem. B106 (2002) 340.
[21] S.A. Carabineiro, B.E. Nieuwenhuys, Surf. Sci. 505 (2002)
163.
[22] S.A.C. Carabineiro, B.E. Nieuwenhuys, Surf. Sci. 532–535
(2003) 87.
[23] K. Yoshida, G.A. Somorjai, Surf. Sci. 75 (1978) 46.
[24] M. Grunze, M. Golze, R.K. Driscoll, P.A. Dowben, J. Vac.
Sci. Technol. 18 (1981) 611.
Acknowledgements
[25] A.R. Cholach, V.A. Sobyanin, V.V. Gorodetskii, React.
Kinet. Catal. Lett. 18 (1981) 391.
This work has been performed under the aus-
pices of NIOK, The Netherlands Institute for
Catalysis Research, laboratory report no. UL 01-
2-03 and of NRSC-Catalysis. The project was
supported by the NRSC-Catalysis and partly by
NWO (47-015.002).
[26] F.P. Netzer, T.E. Madey, Surf. Sci. 119 (1982) 422.
[27] J.L. Gland, B.A. Sexton, G.E. Mitchell, Surf. Sci. 115
(1982) 623.
[28] C. Egawa, S. Naito, K. Tamaru, Surf. Sci. 131 (1983) 49.
[29] V.V. Gorodetskii, M.Yu. Smirnov, A.R. Cholach, Proc.
10th Int. Cong. Catal., Budapest, (1992) 192.
[30] R.M. van Hardeveld, R.A. van Santen, J.W. Niemants-
verdriet, Surf. Sci. 369 (1996) 23.
[31] J.M. Bradley, A. Hopkinson, D.A. King, Surf. Sci. 371
(1997) 225.
References
[32] A. Galtayries, E. Laksono, J.-M. Siffre, C. Argile, P.
Marcus, Surf. Interface Anal. 30 (2000) 140.
[1] J.K. Dixon, J.E. Longfield, in: P.W. Emmel (Ed.), Catal-
ysis, VII, 1960, p. 281.
ꢀ
[33] C.M. Pradier, A. Adamski, C. Methivier, I. Louis-Rose, J.
Mol. Catal. A: Gen. 186 (2002) 193.
[34] J.L. Grand, Surf. Sci. 71 (1978) 327.
[2] R.W. McCabe, T. Pignet, L.D. Schmidt, J. Catal. 32 (1974)
114.
[35] G.B. Fisher, Chem. Phys. Lett. 79 (1981) 452.
[36] J.L. Gland, E.B. Kollin, Surf. Sci. 104 (1981) 478.
[37] N.R. Palmer, J.C. Campuzano, N.V. Richardson, D.A.
King, Vacuum 31 (1981) 679.
[3] J.L. Gland, V.N. Korchak, J. Catal. 53 (1978) 9.
[4] A.C.M. van den Broek, J. van Grondelle, R.A. van Santen,
J. Catal. 185 (1999) 297.
[5] L. Gang, B.G. Anderson, J. van Grondelle, R.A. van
Santen, W.J.H. van Gennip, J.W. Niemantsverdriet, P.J.
Kooyman, A. Knoester, H.H. Brongersma, J. Catal. 206
(2002) 60.
[38] A.R. Cholach, V.A. Sobyanin, V.V. Gorodetskii, Proc.
Soviet-Japanese Seminar on Catalysis, Novosibrisk, (1983)
177.
[39] J.M. Gohndrone, C.W. Olsen, A.L. Backman, T.R. Gow,
E. Yagasaki, R.I. Masel, J. Vac. Sci. Technol., A 7 (1989)
1986.
[6] V.I. Sobolev, G.I. Panov, A.S. Kharitonov, V.N. Roman-
nikov, A.M. Volodin, K.G. Ione, J. Catal. 139 (1993) 435.
[7] H. Orita, H. Kondoh, H. Nozoye, J. Catal. 177 (1998) 217.
[8] S.A.C. Carabineiro, Ph.D. Thesis, New University of
Lisbon, (2000).
[40] L.R. Danielson, M.J. Dresser, E.E. Donaldson, J.T.
Dickinson, Surf. Sci. 71 (1978) 599.
[41] C. Egawa, S. Naito, K. Tamaru, Surf. Sci. 138 (1984) 279.
[42] Y. Zhou, S. Akhter, J.M. White, Surf. Sci. 202 (1988) 357.
[43] H.H. Huang, C.S. Seet, G.Q. Xu, Surf. Sci. 317 (1994) 353.
[44] K. Jacobi, Y. Wang, C.Y. Fan, H. Dietrich, J. Chem. Phys.
115 (2001) 4306.
[9] T. Pignet, L.D. Schmidt, Chem. Eng. Sci. 129 (1974) 1123.
[10] R.T.K. Baker, R.B. Thomas, J.H.F. Norton, Plat. Met.
Rev. 18 (1974) 130.
[11] T. Pignet, L.D. Schmidt, J. Catal. 40 (1975) 212.
[12] J.L. Gland, G.C. Woodard, V.N. Korchak, J. Catal. 61
(1980) 543.
[45] Y. Wang, A. Lafosse, K. Jacobi, Surf. Sci. 507–510 (2002)
773.
[13] M. Asscher, W.L. Guthriet, T.-H. Lin, G.A. Somorjai, J.
Phys. Chem. 88 (1984) 3233.
[46] A.C.M. van den Broek, Ph.D. Thesis, Technische Univer-
siteit Eindhoven, (1998).
[14] W.D. Mieher, W. Ho, Surf. Sci. 322 (1995) 151.
[15] J.M. Bradley, A. Hopkinson, D.A. King, J. Phys. Chem. 99
(1995) 17032.
[47] L. Gang, B.G. Anderson, J. van Grondelle, R.A. van
Santen, Appl. Catal. B: Env. 40 (2003) 101.
[48] J.Q. Barteau, Broughton, D. Menzel, Surf. Sci. 133 (1983) 443.
[49] A. Schiffer, P. Jacob, D. Menzel, Surf. Sci. 465 (2000) 210.
[50] H. Shi, K. Jacobi, G. Ertl, J. Chem. Phys. 102 (1995) 1432.
[51] S. Talat, S. Rahman, Phys. Rev. Lett. 51 (1983) 1979.
[16] D.S.Y. Hsu, D.W. Squire, M.C. Lin, J. Chem. Phys. 89
(1998) 2861.
[17] V.V. Gorodetskii, V.A. Sobyanin, Proc. 7th Int. Cong.
Catal., Tokio, (1980) 566.

S.A.C. Carabineiro et al. / Surface Science 555 (2004) 83–93
93
[52] S.K. Shi, J.A. Schretels, J.M. White, Surf. Sci. 105 (1981) 1.
[53] M.H. Koch, P. Jacob, D. Menzel, Surf. Sci. 367 (1996) 293.
[54] T.E. Madey, H.A. Engelhardt, D. Menzel, Surf. Sci. 48
(1975) 304.
[63] J.E. Parmeter, Y. Wang, C.B. Mullins, W.H. Weinberg, J.
Chem. Phys. 88 (1988) 5225.
[64] T. Sasaki, T. Aruga, H. Kuroda, Y. Iwasawa, Surf. Sci. 240
(1990) 223.
[55] S.L. Parrot, G. Praline, B.E. Koel, J.M. White, T.N.
Taylor, J. Chem. Phys. 71 (1979) 2788.
[65] K. Jacobi, H. Dietrich, G. Ertl, Appl. Surf. Sci. 121/122
(1997) 558.
[56] M.F.H. van Tol, J. Siera, P.D. Cobden, B.E. Nieuwenhuys,
Surf. Sci. 274 (1992) 63.
[66] S. Dahl et al., Phys. Rev. Lett. 83 (1999) 1814.
[67] J. Hrbek, J. Phys. Chem. 90 (1986) 6217.
[68] J. Hrbek, J. Catal. 100 (1986) 523.
[57] Y. Li, J.N. Armor, Appl. Catal. B: Env. 13 (1997) 131.
[58] B.E. Nieuwenhuys, G.A. Somorjai, Surf. Sci. 72 (1978) 8.
[59] E. Molinari, F. Cramarossa, A. Pullo, L. Triolo, J. Catal. 4
(1965) 341.
[69] A. Schiffer, P. Jacob, D. Menzel, Surf. Sci. 465 (2000) 198.
[70] M. Kim, S.J. Pratt, D.A. King, J. Am. Chem. Soc. 122
(2000) 2409.
[60] J. Zawadzki, Discuss. Farad Soc. 8 (1950) 140.
[61] C. Benndorf, T.E. Madey, Surf. Sci. 135 (1983) 164.
[62] W. Tsai, W.H. Weinberg, J. Chem. Phys. 91 (1987) 5305.
[71] E.D.L. Rienks, C.J. Weststrate, J.W. Bakker, S.A.C.
Carabineiro, B.E. Nieuwenhuys, S. Lizzit, A. Baraldi,
unpublished results.