Structure sensitivity of ammonia oxidation over platinum

Article abstract of DOI:10.1016/j.jcat.2008.05.032

Ammonia oxidation over different Pt single crystal orientations and a Pt foil has been studied in the 10^-5 to 10^-2 mbar range with rate measurements and LEED. The reaction exhibits only a moderate structure sensitivity with the activ

Full text of DOI:10.1016/j.jcat.2008.05.032

Journal of Catalysis 261 (2009) 129–136
Contents lists available at ScienceDirect
Journal of Catalysis
www.elsevier.com/locate/jcat
Structure sensitivity of ammonia oxidation over platinum
∗
Ying Feng Zeng, Ronald Imbihl
Institut für Physikalische Chemie und Elektrochemie, Leibniz-Universität Hannover, Callinstr. 3-3a, 30167 Hannover, Germany
a r t i c l e i n f o
a b s t r a c t
−5
Article history:
Received 25 February 2008
Revised 16 May 2008
Accepted 17 May 2008
Available online 12 November 2008
Ammonia oxidation over different Pt single crystal orientations and a Pt foil has been studied in the 10
to 10
−2
mbar range with rate measurements and LEED. The reaction exhibits only a moderate structure
sensitivity with the activity varying in the sequence Pt foil:Pt(533):Pt(865):Pt(443):Pt(100) = 6:4:4:2:1.
The activity is apparently mainly determined by the oxygen sticking coefficient which reaches 0.14 on
the Pt foil. Broad hystereses in the reaction rates upon cyclic variation of the temperature are associated
with reversible structural changes. The extent of reaction-induced structural changes depends strongly on
the orientation, the total pressure and also on the ratio of the reactants NH3/O2. On Pt(533) and Pt(443)
Keywords:
Platinum
the extent of restructuring and the hysteretic behavior was found to depend in a non-monotonic way
Stepped platinum surface
Ammonia oxidation
Kinetics
−3
on the total pressure. Exposure to reaction conditions at 10
mbar creates a disordered surface as seen
−2
in LEED with no LEED spots visible; but after exposure to reaction conditions at 10
order beams in LEED are again visible.
mbar the integral
© 2008 Published by Elsevier Inc.
1. Introduction
exhibits an adsorbate-induced surface phase transition between
a catalytically active bulk-like (1 × 1) termination and a quasi-
hexagonal reconstruction of the topmost layer (“hex”) which has
a very low activity [14,25,26]. Catalytic ammonia oxidation on
Pt(100) has been studied quite in detail by the group of King but
The catalytic oxidation of ammonia over platinum is a key step,
both in the industrial manufacturing of nitric acid and in envi-
ronmental chemistry where ammonia is removed in the so-called
selective catalytic reduction (SCR process) [1]. Theoretical studies
as well as experimental data show that ammonia decomposition
on platinum is activated through direct interaction of ammonia
with chemisorbed oxygen or OH species [2–5]. Single crystal stud-
ies have been performed with Pt(100) [2,3], Pt(111) [6,7] and with
stepped Pt(111) orientations [3,8–12]. Already in the very early
studies of Gland et al. it was concluded that the reactivity of Pt
samples in catalytic ammonia oxidation is determined by the den-
sity of steps [8]. Since two elementary steps of the reaction mech-
anism, dissociative oxygen chemisorption and NO decomposition
are highly structure sensitive on Pt, ammonia oxidation should be
structure sensitive too [13–23].
mostly under unstationary reaction conditions and only below a
−5
pressure of 10
mbar [2,4].
The structure sensitivity of ammonia oxidation contains a dy-
namic aspect which is that the substrate structure is modified by
the reaction. With varying extent this is certainly true for all cat-
alytic reactions but the effect is particularly strong for ammonia
oxidation on Pt where a visible roughening of Pt/Rh gauzes used
in the Ostwald process already occurs during the first minutes of
operation. Phenomenological studies focusing on reaction-induced
morphological changes have been conducted with small Pt spheres
at high pressure [27,28]. With LEED and STM restructuring was
−5
−4
mbar range on Pt(533) and Pt(443)
studied in the 10
and 10
[3,11,29]. On Pt(533) a reversible doubling of the step height oc-
curs associated with a change in the selectivity of the reaction.
The reversible restructuring of Pt(533) shows up in a hysteresis of
the N2 and NO production rates upon cycling the temperature. On
Pt(443) already at 300 K the initially straight step edges start to
meander upon ammonia adsorption [3].
In this report we compare the activity of Pt(533), Pt(443),
Pt(865), Pt(100), and a Pt foil in ammonia oxidation up into the
−2
10
mbar range. The surfaces Pt(533) and Pt(443) are stepped
(111) surfaces which differ in their step orientation as evident
from the microfacet notation which is 4(111) × 1(100) for Pt(533)
and 7(111) × 1(111) for Pt(443). Their kinetics have been studied
−4
The questions are to which extent the different orientations un-
dergo restructuring and how the restructuring depends on the total
pressure. Furthermore, the restructuring should be associated with
an activation or deactivation of the catalyst and with selectivity
changes. In this study we address these questions using rate mea-
surements and LEED to follow reaction-induced restructuring as
quite in detail up to 10
mbar before [11,12]. As model sys-
tem for a kinked surface we take the Pt(865) surface written
11(111) × 2(511) in microfacet notation [24]. The Pt(100) surface
Corresponding author.
E-mail address: gatzen@pci.uni-hannover.de (R. Imbihl).
*
−5
−2
mbar. In this
we increase the pressure from 10
mbar to 10
0021-9517/$ – see front matter © 2008 Published by Elsevier Inc.
doi:10.1016/j.jcat.2008.05.032

130
Y.F. Zeng, R. Imbihl / Journal of Catalysis 261 (2009) 129–136
way we only partially bridge the pressure gap between UHV and
the conditions of the Ostwald process but already in this range we
find quite unexpected results. The ordering we observe after re-
structuring seems to depend in a non-monotonic way on the total
pressure.
foil:Pt(533):Pt(865):Pt(443):Pt(100) = 6:4:4:2:1. We also note that
with exception of Pt(100) the N2 rate maxima shift to lower tem-
perature with increasing catalytic activity of the sample.
A hysteresis in the activity can arise (i) due to reversible struc-
tural changes or an activation/deactivation by oxide formation and
reduction (ii) due to the inhibitory effect of adsorbates on the ad-
sorption of reactants or (iii) through the formation and removal
of surface contaminants. Realistic mathematical models of am-
monia oxidation on Pt did not reveal any kinetic multistability
which makes (ii) rather unlikely [5,9]. In situ XPS measurements of
Pt(533)/NH3 +O2 showed that up to 1 mbar no Pt oxide forms [30].
Reversible restructuring and a potential influence of surface con-
taminants remain therefore as possible causes for the hystereses.
For Pt(100) the hysteresis in the reaction rates is clearly associ-
ated with the (1 × 1) ↔ hex phase transition. The low rate branch
is connected with inactive hex reconstructed surface whereas the
high rate branch belongs to the active (1 × 1) termination of the
substrate [2,31].
As will be shown below LEED data taken after completion of
the measurements reveal a very drastic restructuring of the Pt(865)
surface. This indicates that structural changes take place in the
heating/cooling cycles. For the Pt foil we have no means of detect-
ing structural changes but the texture of the foil contains grains
with (100) and (111) orientations which have been shown to un-
dergo restructuring, i.e. via the adsorbate-induced surface phase
transition in the case of Pt(100) and via reaction-induced faceting
in the case of Pt(111) [26,28]. We can therefore assume that the
large hysteresis we observe with the Pt foil presumably has the
same origin.
2. Experimental
The reaction was studied in a standard UHV system equipped
with LEED, a retarding field analyzer for Auger electron spec-
troscopy and a differentially pumped quadrupole mass spectrom-
eter (QMS) for rate measurement. The base pressure in the UHV
−10
system was 2 × 10
mbar. The sample was heated indirectly by
a filament behind the backside of the crystal via either radiation or
electron bombardment. Samples were prepared by repeated cycles
−6
of Ar-ion sputtering, heating in oxygen (Poxygen = 5 × 10
mbar),
followed by annealing to 1300 K. The sample temperature was
measured with a thermal couple spot-welded to the side of the
sample. A PID controller was used for temperature programmed
reaction (TPR) experiments. The gas purity was 5.0 for oxygen and
2.5 for ammonia. The Pt(533) and Pt(443) samples were the same
as used in previous experiments [3,11,29].
In order to investigate this reaction in an intermediate pres-
−3
sure range between 10
and 1 mbar, a high pressure reaction cell
was connected to the UHV main chamber via a sample transfer
system. After cleaning, the sample is transferred into the reaction
cell. During rate measurements in the low pressure regime a cone
with 2 mm orifice, which separates the differentially pumped QMS
from the main chamber, was brought 1 mm in front of the sam-
ple surface. In this way only species originating from the sample
surface are detected. For rate measurements in the high pressure
cell a stainless steel tube with a 2 mm orifice was brought approx-
imately 1 mm in front of the sample surface.
For calibration the gases NH3, O2, N2, and NO were introduced
into the main chamber so that the QMS signal could be related to
the real pressure. From the measured pressure increase under reac-
tion conditions the number of particles desorbing from the sample
surface was calculated with ideal gas equation using the known
pumping rate and the surface area of the sample.
For characterizing the surface in situ we can measure the reac-
tive sticking coefficient of oxygen. Due to the geometric arrange-
ment of the QMS which is shielded behind a cone, only molecules
reflected from the surface can enter the cone to be detected. We
obtain the reactive sticking coefficient sreac following the variation
of the partial pressures of the reactants, i.e. of O2 or NH3. Denot-
ing the signal of a gas without reaction by I0 and during reaction
with I we calculate the reactive sticking coefficient
I0 − I
.
sreac
=
I0
3. Results
In this case the reaction rate at 300 K was assumed to be negligi-
ble so that the partial pressures at 300 K should represent I0.
The variation of the reactive sticking coefficient (sreac) of oxygen
during the temperature cycling experiments is reproduced in Fig. 3
for the Pt foil, Pt(100) and Pt(865). The variation of sreac reflects
in general rather well the behavior of the reaction rates during the
T -cycling experiments displayed in Fig. 1. For Pt(100), for example,
the low reactivity of the cooling branch is due to the low oxygen
3.1. Structure sensitivity
In Fig. 1 we compare the activity of two stepped Pt (111) sur-
faces, Pt(533) and Pt(443), with that of a kinked surface, Pt(865),
with a Pt(100) sample, and with a Pt foil. The samples were sub-
jected to heating/cooling cycles in an NH3/O2 atmosphere with
a 1:1 ratio of the partial pressures at a total pressure of 1 ×
−5
10
mbar. The same y-scaling has been used for all samples so
sticking coefficient on the hex phase which according to the lit-
that the rate curves can be compared directly. Structural models of
the single crystal orientations are displayed in Fig. 2. The relatively
small hystereses we observe on Pt(443), Pt(865), and on Pt(533)
can be attributed mainly to transients caused by a heating/cooling
rate of 0.5 K/s which is still too large to ensure true steady state
conditions. The larger hystereses we see on Pt(100), and the Pt foil
are clearly true hystereses caused by reversible structural changes
of the Pt substrate. A broad hysteresis connected to reaction-
−3
erature is as low 10
for a structurally nearly perfect hex phase
[2,14]. During heating up sreac does not reach the value of 0.2 re-
ported for oxygen sticking on the bare Pt(100)-(1 × 1) surface but
only goes up to 0.03 [2]. The discrepancy might indicate that the
hex phase has not been lifted completely during cooling down so
that a significant portion of the surface remains in the inactive hex
state all the time.
In contrast to Pt(100) the activity of the Pt foil increases by
heating up as evidenced by a higher rate maxima for N2 and NO
and a higher sreac in Figs. 1 and 3, respectively. Remarkably, be-
tween roughly 600 and 800 K the selectivity changes drastically
from preferential N2 formation on heating up to NO as main prod-
uct during cooling down. The increase in the overall activity after
heating is reflected by the increase in sreac in Fig. 3. The compar-
ison of the Pt foil with Pt(865) shows that on these two samples
sreac reaches about 0.14 and 0.12, respectively.
induced structural changes has also been observed with Pt(533)
−5
but only at a total pressure substantially beyond 1 × 10
mbar
−5
[11,29]. At the 1 × 10
mbar employed here we therefore do not
see such a broad hysteresis.
Qualitatively the rate curves of all samples are similar. N2
formation is the dominant reaction pathway at low temperature
whereas at high temperature NO production prevails. If we take
the maximum in N2 production as a measure of the catalytic ac-
tivity we obtain the following sequence in catalytic activity: Pt

Y.F. Zeng, R. Imbihl / Journal of Catalysis 261 (2009) 129–136
131
−5
Fig. 1. Comparison of the catalytic activity of Pt(533), Pt(443), Pt(865), Pt(100), and of a Pt foil in the 10
mbar range with a 1:1 mixing ratio of the gases. The ramping
speed in the T -cycling experiments was 30 K/min. Due to the relatively high ramping rate the hystereses with exception of Pt(100) are mainly caused by transients.
3.2. Influence of total pressure
speed. In all experiments we started with a freshly prepared sur-
face after Ar-ion sputtering, oxygen treatment and annealing to
−3
1100 K. Experiments with a pressure at or above 10
mbar were
We studied the influence of the total pressure on the kinetics
−2
of N2 and NO production in the 10−5–10
mbar range. Figs. 4
and 5 display the results for Pt(533) and Pt(443) obtained for
a 1:1 ratio NH3/O2 of the reactants and with identical ramping
carried out in the high pressure chamber, and with a pressure be-
mbar in the main chamber. Fig. 6 displays for Pt(533),
Pt(443), and Pt(865) a plot of the maximum N2 production rate
−3
low 10

132
Y.F. Zeng, R. Imbihl / Journal of Catalysis 261 (2009) 129–136
Fig. 3. Variation of reactive oxygen sticking coefficient (sreac) during temperature
cycling experiments with a Pt foil, Pt(865) and Pt(100). The total pressure is in the
−5
10
mbar range, and a 1:1 mixing ratio of the gases NH3 and O2 is used. The
ramping speed in the T -cycling experiments was 30 K/min.
Fig. 2. Structural models of different Pt single crystal surfaces: (a) Pt(533), (b) Pt
(443), and (c) Pt(865).
−3
versus the total pressure. Up to 10
mbar the N2 production
−3
scales with the total pressure but beyond 10 mbar the rate slows
−2
down. At 10
mbar the measured rate is lower by roughly a fac-
tor of three compared to the rate that scaling with total pressure
would predict. According to experiments conducted with He mix-
ing mass transport limitations start to play a role only beyond
10 mbar at catalyst temperatures above 660 K [5]. More likely is
−2
that at 10
mbar the adsorbate coverages become large enough
to slow down considerably the adsorption kinetics so that scaling
of the rate with the total pressure can longer be expected.
−5
As shown in Fig. 4, at a total pressure of 6 × 10
mbar Pt(533)
displays a broad hysteresis in the N2 and NO production which was
shown to be due to a reversible doubling of the step height upon
−5
heating [11,29]. At 6 × 10
mbar the double-atomic steps were
demonstrated to be present roughly between 600 K and 800 K on
−3
the heating branch. Increasing the total pressure to 1 × 10
mbar
considerably enhances the N2 hysteresis while the NO hystere-
sis practically disappears. The latter is probably a consequence of
shifting the NO rate maximum to a higher temperature outside the
Fig. 4. Effect of the total pressure on the kinetics of N2 and NO production on
−2
Pt(533), the experiments are conducted in the 10−5–10
mbar range under a feed
composition 1:1 and with a ramping speed of 0.5 K/s.

Y.F. Zeng, R. Imbihl / Journal of Catalysis 261 (2009) 129–136
133
again similar to the behavior of Pt(533). The hystereses of Pt(533)
and Pt(433) are all counterclockwise, i.e. heating up leads to an
activation of the catalyst.
The preceding measurements show an unexpected result: a
non-monotonic variation of the hysteretic behavior with pres-
sure for both orientations, Pt(533) and Pt(443). At low pressure
−5
(p < 10
mbar) no hysteresis occurs, at intermediate pressure
a very pronounced hysteresis is observed and at high pressure
−2
(10
mbar) the hysteresis vanishes again. The two orientations
Pt(533) and Pt(443) differ in so far as Pt(533) displays a pro-
−5
−4
mbar whereas no hysteresis
nounced hystereses at 10
and 10
exist for Pt(443) in this pressure range. The explanation as veri-
fied by LEED and STM is that on Pt(533) a reversible doubling of
the step height occurs while the average step separation remains
constant on Pt(443).
With increasing total pressure in general the risk of a sur-
face contamination rises for two reasons. The partial pressure of
contaminants in the gas phase increases and the higher chemical
potential of the reactants enhances the segregation of bulk con-
taminants to the surface. One might suspect that contamination
effects are responsible for the pressure dependent vanishing of the
hystereses but so far Auger electron spectroscopy provided no in-
dication of surface contaminants. Moreover, the segregation of Si
which leads to SiO2 formation on the Pt surface would cause an ir-
reversible reduction of catalytic activity which is not what we see.
On the other hand, since we cannot completely rule out a poten-
tial influence of contaminants this possibility should still be taken
into account. In the following we try to establish a connection be-
tween the rate hystereses and changes in the catalytic properties
of Pt(533) and Pt(443).
For Pt(533) it was demonstrated that the doubling of the step
height under reaction conditions is associated with a change in
the selectivity from N2 towards NO formation. Since the availabil-
ity of oxygen controls the selectivity towards NO an increase in
Fig. 5. Effect of the total pressure on the kinetics of N2 and NO production on
−2
Pt(443), the experiments are conducted in the 10−5–10
mbar range under a feed
composition 1:1 and with a ramping speed of 0.5 K/s.
the oxygen sticking coefficient could explain the observed change
−4
in selectivity. The data in Fig. 7a recorded in the 10
mbar range
show that upon heating sreac increases sharply at 700 K which is
roughly the temperature where we expect the doubling of the step
height to occur. sreac, however, also remains high during cooling
down but during cooling down the steps remains single-atomic as
shown in earlier measurements [11,29]. Evidently sreac cannot be
determined by the step structure alone but apparently also the ad-
sorbate coverages determine sreac by blocking sites for adsorption.
In addition, also local structural changes which do not show up in
LEED might play a role. Measurements of sreac with different ra-
tions NH3:O2 show that the hysteresis becomes smaller the more
oxygen is in excess. For a 10 fold excess of O2 the hysteresis prac-
tically vanishes.
With LEED the surface structure of Pt(533), Pt(443), and Pt(865)
was examined after exposure to reaction conditions. The beam
profiles of the split (1, 0)-beam after the hysteresis measurements
−2
in the 10
mbar range are displayed in Fig. 8. On all three sur-
faces the beam intensity strongly decreased in comparison to their
initial values. On Pt(865) the ordering after the reaction is so poor
that the LEED spots are hardly discernible from the background.
The splitting of the spot is no longer visible and instead a very
small peak appears at the middle position of the original spot
splitting. On Pt(533) we observe a widening of the spot splitting
by roughly 20%. On Pt(433) the spot splitting remained unchanged
but the FWHM of the spots increased and the maximum intensity
dropped by 50%. This indicates some disordering caused by the re-
action. Apparently the average terrace width on Pt(443) has not
been changed by the reaction in contrast to Pt(533) where the av-
erage terrace width decreased by around 20%.
Fig. 6. Dependence of the maximum rate of N2 production on the total pressure.
The Pt orientations Pt(533), Pt(443), and Pt(865) are compared. The same reaction
conditions are applied with a mixing ratio of ammonia to oxygen of 1:1.
T -cycling window. Quite surprisingly, a further increase of the to-
−2
tal pressure to 1 × 10
mbar leads to a vanishing hysteresis.
−5
On Pt(443) practically no hysteresis is seen at 10
mbar and
−4
−3
at 10
mbar but increasing the total pressure to 1 × 10
mbar
causes the appearance of a substantial hysteresis as evidenced by
An overview summarizing the observations with LEED and the
hysteresis measurements of the different orientations in the differ-
−2
Fig. 5. After a further rise to 1 × 10
mbar the hysteresis vanishes

134
Y.F. Zeng, R. Imbihl / Journal of Catalysis 261 (2009) 129–136
(a)
(b)
(c)
(a)
Fig. 7. Reactive sticking coefficient of oxygen during ammonia oxidation on Pt(533)
under varying feed composition NH3:O2, from 1:1 to 1:10. The ramping speed is
(b)
−4
10 K/min. The total pressure is at around 10
mbar.
Fig. 8. (a) A LEED pattern of Pt(533) showing the profile window along which the
intensity profiles reproduced in (b) were taken. (b) LEED spot profiles showing
ent pressure ranges is given in Table 1. We first focus on Pt(533)
and Pt(443). The LEED data as well as the hysteresis measurements
do not exhibit a monotonic variation with pressure. Both orienta-
−2
reaction-induced restructuring in the 10
mbar range. Beam intensity profiles of
the (1,0)-beam are taken between the (1, −1/2) and the (1, 1/2) beam which is the
−2
direction of the spot-splitting. The sample was exposed to NH3/O2 at 10
mbar
−3
tions, Pt(533) and Pt(443) become disordered at 10
at higher pressure, at 10
mbar but
during temperature cycling from 300 to 800 K for 150 min. The mixing ratio is 1:1,
LEED beam energy E = 110 eV.
−2
mbar, some ordering is present again.
Similarly, both orientations display broad hystereses at 10
but at 10 mbar the hystereses vanish for both orientations. In or-
der to exclude a contamination effect, we changed the sequence in
−3
mbar
−2
On Pt(865) the behavior of Pt(865) is insofar different from that
of Pt(533) and Pt(443) as a hysteresis is only visible at low pres-
−4
mbar, where the surface is also still well ordered.
which the experiments were performed. First ammonia oxidation
sure, at 10
−2
was carried out at 10
mbar, subsequently the total pressure was
At higher pressure the surface undergoes complete restructuring
and the hysteresis disappears. The most stable surface of all orien-
tations we investigated is Pt(100) as evidenced by Table 1. Up to
−3
reduced to 10
mbar. Also in this sequence we find a hysteresis
mbar but not at 10
−3
−2
at 10
mbar.

Y.F. Zeng, R. Imbihl / Journal of Catalysis 261 (2009) 129–136
135
Table 1
Overview of the LEED and T -cycling experiments over different pressure ranges for Pt(533), Pt(443), Pt(865), and Pt(100).
Pressure range
Pt(533)
Pt(443)
Pt(865)
Pt(100)
Hysteresis in
reaction rate
Surface
ordering/LEED
Hysteresis in
reaction rate
Surface
ordering/LEED
Hysteresis in
reaction rate
Surface
ordering/LEED
Hysteresis in
reaction rate
Surface
ordering/LEED
−4
10
10
10
mbar
mbar
mbar
Yes
Yes
No
Ordered (1 × 1)
No
Yes
No
Ordered (1 × 1)
Yes
No
No
Ordered (1 × 1)
Yes
Yes
Yes
Ordered (1 × 1)
Ordered (1 × 1)
Ordered (1 × 1)
−3
−2
Disordered
Disordered
Disorder
a
Ordered (1 × 1)
Ordered (1 × 1)
a
Diffuse broad spot with low intensity.
−2
10
mbar this surface remains well-ordered independent of the
itor structural changes during the rate hystereses. The fact that at
−3
pressure range.
10
mbar large hystereses on Pt(533) and Pt(443) are connected
with a disordered surface (see Table 1) seems to be in contradic-
tion to the previous statement relating rate hystereses with re-
versible structural changes. However, the LEED data in Table 1 only
show the state of the surface after completion of the experiment.
This leaves the possibility that during the T -cycles an ordered sur-
face may exist in a certain parameter range. A second possibility is
that LEED only shows the absence of long range order while some
short range order may still exist. Local restructuring without long
range order might therefore be responsible for the rate hystereses
found on disordered surfaces.
One rather unexpected phenomenon was the non-monotonic
variation of the hysteretic behavior and of the surface ordering
with total pressure observed with Pt(533) and Pt(443) (Table 1).
In principle, one could envision a surface which is roughened by
the reaction at some intermediate pressure but which orders again
as the dynamics of the reaction beyond a certain threshold become
large enough to overcome kinetic barriers preventing ordering. At
4. Discussion
Using the maximum in N2 production as a measure for catalytic
activity the activity ratio Pt foil:Pt(533):Pt(865):Pt(443):Pt(100) =
6:4:4:2:1 was obtained. The choice of the maximum in N2 produc-
tion as a measure for catalytic activity contains some arbitrariness
because the height of the maximum will also depend on the tem-
perature where the branching to preferential NO formation takes
place. If we use instead the maximum in the reactive oxygen stick-
ing coefficient, sreac, we come to a corresponding activity ratio of
4.5:2:4:2:1. With exception of Pt(533) the picture is very similar to
the ratio obtained with the maximum in N2 production.
In this study of different orientations the Pt(111) surface has
not been included but since the oxygen sticking coefficient of a
structurally perfect Pt(111) surface is very low [32], it is probably
safe to assume that their catalytic activity is very low too. The
higher activity of Pt(533) in the N2 production rate as compared to
Pt(443) which amounts to a factor of two (Fig. 1) can be attributed
to the higher step density of Pt(533) which is higher by a factor
of 7/4. What also might play a role could be dependence of the
activity on the step orientation. With regard to Pt(100)-(1 × 1) one
might suspect a higher activity of (100) steps as compared to steps
with (111) orientation but this remains to be shown. Comparing
Pt(865) with the stepped orientations in Fig. 1, one can conclude
from their similar activity that the additional presence of kinks
does not have a dramatic effect on the activity.
the present stage, this is pure speculation and experiments extend-
−2
ing beyond 10
mbar are required in order to substantiate this
interesting possibility.
5. Conclusions
The results obtained with different Pt orientations studied in
−5
−2
mbar show that ammonia oxidation on plat-
the 10
to 10
inum is only a moderately structure sensitive reaction. The dif-
ferent reactivity is essentially determined by the difference in the
oxygen sticking coefficient. The catalytic activity is highest on the
Pt foil and then decreases in the order Pt(865), Pt(533), Pt(443)
and Pt(100). The extent of reaction-induced restructuring depends
strongly on the total pressure, the orientation and the mixing ratio
of reactants. On Pt(533) and Pt(443) we observe a non-monotonic
dependence of the amount of restructuring on the total pressure:
Quantum chemical studies of ammonia activation by Oad and
OHad on flat and stepped Pt(111) surfaces indicated a small in-
fluence of the surface structure on the Pt activity as far as the
pure decomposition steps of NHx (x = 1–3) are concerned [3]. The
main factors causing a structure sensitivity of ammonia oxidation
on Pt are the adsorption of oxygen and ammonia and NO forma-
tion/decomposition. Ammonia adsorption on Pt surfaces exhibits
only a moderate structure sensitivity [33], but oxygen sticking on
increasing restructuring with pressure leads to a completely disor-
−3
dered surface in LEED at 10
at 10
mbar but at even higher pressure,
Pt is known to vary over three orders of magnitude depending on
the structure from about 10
−2
mbar, the ordering of the surfaces is partially restored.
−3
on Pt(100)-hex to a value close to
one on Pt(210) [2,15,22]. Detailed studies of oxygen adsorption on
Pt(533) have been performed in the group of Hayden [17]. The dif-
ferent studies all agree that oxygen adsorption on Pt is largely de-
termined by structural defects, i.e. by atomic steps and kinks [34].
The data shown here indicate that kinked surface have a low
structural stability under reaction conditions followed by stepped
surfaces. The Pt(100) surface exhibits the highest stability of all
studied surfaces under reaction conditions. This is consistent with
the results of Schmidt et al. who studied small single crystal
spheres of Pt in a NH3/O2 atmosphere and found that after ex-
posure to reaction conditions in the mbar range (100) orientations
survive [27,28].
Acknowledgments
Assistance from Dr. Sebastian Günter in the experiments is
gratefully acknowledged. The authors are indebted to the group
of M. Baerns for supplying a Pt foil as sample. This work was sup-
ported by the DFG under the priority program 1091 ‘Bridging the
gap between ideal and real systems in heterogeneous catalysis’.
References
[1] T.H. Chilton, in: The Manufacture of Nitric Acid by the Oxidation of Ammonia,
in: Chemical Engineering Progress Monograph Series, vol. 3, Am. Inst. Chem.
Eng., New York, 1960.
[2] J.M. Bradley, A. Hopkinson, D.A. King, J. Phys. Chem. 99 (1995) 17032.
[3] R. Imbihl, A. Scheibe, Y.F. Zeng, S. Gunther, R. Kraehnert, V.A. Kondratenko,
M. Baerns, W.K. Offermans, A.P.J. Jansen, R.A. Van Santen, Phys. Chem. Chem.
Phys. 9 (2007) 3522.
Previous LEED and STM investigations of Pt(533)/NH3 + O2 con-
−5
−4
mbar range have shown that the
ducted in the 10
and 10
rate hysteresis is connected with reversible structural changes, i.e.
a reversible doubling of the step height [29]. In the pressure ar-
range investigated here no in situ methods were available to mon-

136
Y.F. Zeng, R. Imbihl / Journal of Catalysis 261 (2009) 129–136
[4] M. Kim, D.A. King, S.J. Pratt, J. Am. Chem. Soc. 122 (2000) 2409.
[20] J.M. Gohndrone, Y.O. Park, R.I. Masel, J. Catal. 95 (1985) 244.
[21] M. Sander, R. Imbihl, G. Ertl, J. Chem. Phys. 97 (1992) 5193.
[22] R.J. Gorte, L.D. Schmidt, J.L. Gland, Surf. Sci. 109 (1981) 367.
[23] T. Sugisawa, J. Shiraishi, D. Machihara, K. Irokawa, H. Miki, C. Kodama, T. Kuri-
yama, T. Kubo, H. Nozoye, Appl. Surf. Sci. 169–170 (2001) 292.
[5] W.K. Offermans, A.P.J. Jansen, R.A.v. Santen, Surf. Sci. 600 (2006) 1714.
[6] M. Asscher, W.L. Guthrie, T.-H. Lin, G.A. Somorjai, J. Phys. Chem. 88 (1984)
3233.
[7] W.D. Mieher, W. Ho, Surf. Sci. 322 (1995) 151.
[8] J.L. Gland, V.N. Korchak, J. Catal. 53 (1978) 9.
[24] M.A.V. Hove, G.A. Somorjai, Surf. Sci. 92 (1980) 489.
[9] W.L. Guthrie, J.D. Sokol, G.A. Somorjai, Surf. Sci. 109 (1981) 390.
[10] D.G. Löffler, L.D. Schmidt, Surf. Sci. 59 (1976) 195.
[11] A. Scheibe, U. Lins, R. Imbihl, Surf. Sci. 577 (2005) 1.
[12] A. Scheibe, M. Hinz, R. Imbihl, Surf. Sci. 576 (2005) 131.
[13] H. Wang, R.G. Tobin, D.K. Lambert, A.L. DiMaggio, G.B. Fisher, Surf. Sci. 372
(1997) 267.
[14] K. Griffiths, T.E. Jackman, J.A. Davies, P.R. Norton, Surf. Sci. 138 (1984) 113.
[15] N. Freyer, M. Kiskinova, G. Pirug, H.P. Bonzel, Surf. Sci. 166 (1986) 206.
[16] A. Rar, T. Matsushima, Surf. Sci. (1994) 89.
[25] P.A. Thiel, R.J. Behm, P.R. Norton, G. Ertl, J. Chem. Phys. 78 (1983) 7448.
[26] X.-C. Guo, J.M. Bradley, A. Hopkinson, D.A. King, Surf. Sci. 310 (1994) 163.
[27] R.W. McCabe, T. Pignet, L.D. Schmidt, J. Catal. 32 (1974) 114.
[28] M. Flytzani-Stephanopoulos, L.D. Schmidt, Prog. Surf. Sci. 9 (1979) 83.
[29] A. Scheibe, S. Günther, R. Imbihl, Catal. Lett. 86 (2003) 33.
[30] S. Günther, A. Scheibe, H. Bluhm, M. Haevecker, E. Kleimenov, R. Schlögl,
A. Knop-Gericke, R. Imbihl, J. Phys. Chem. C 112 (2008) 15382.
[31] M. Rafti, F. Lovis, Y. Zeng, R. Imbihl, Chem. Phys. Lett. 446 (2007) 323.
[32] C.T. Campbell, G. Ertl, H. Kuipers, J. Segner, Surf. Sci. 107 (1981) 220.
[33] 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.
[17] A.T. Gee, B.E. Hayden, J. Chem. Phys. 113 (2000) 10333.
[18] V.I. Savchenko, N.I. Efremova, React. Kinet. Catal. Lett. 56 (1995) 97.
[19] J.M. Gohndrone, R.I. Masel, Surf. Sci. 209 (1989) 44.
[34] L. Vattuonea, L. Savioa, M. Roccaa, Surf. Sci. Rep. 63 (2008) 101.