NO-H2 reaction over Pd(111)

Article abstract of DOI:10.1016/S0039-6028(00)00825-6

The NO-H₂ reaction over Pd(111) was studied at NO pressures in the 1×10^-7 mbar pressure regime and at temperatures between 300 and 750 K. The H₂/NO ratio was varied from 1 to 54. Hysteresis phenomena in the rate of NH₃ and H₂O formation were observed in a heating-cooling cycle and can be attributed to atomic hydrogen absorbed in the bulk.

Full text of DOI:10.1016/S0039-6028(00)00825-6

Surface Science 469 (2000) 196–203
www.elsevier.nl/locate/susc
The NO–H reaction over Pd(111)
2
*
C.A. de Wolf , B.E. Nieuwenhuys
Leiden Institute of Chemistry, Gorlaeus Laboratories, Leiden University, P.O. Box 9502, 2300 RA Leiden, The Netherlands
Received 7 August 2000; accepted for publication 1 September 2000
Abstract
The NO–H reaction over Pd(111) was studied at NO pressures in the 1×10−7 mbar pressure regime and at
2
temperatures between 300 and 750 K. The H /NO ratio was varied from 1 to 54. Hysteresis phenomena in the rate
2
of NH and H O formation were observed in a heating–cooling cycle and can be attributed to atomic hydrogen
3
2
absorbed in the bulk. © 2000 Elsevier Science B.V. All rights reserved.
Keywords: Catalysis; Hydrogen molecule; Low index single crystal surfaces; Nitrogen oxides; Palladium; Surface chemical reaction;
Thermal desorption spectroscopy
1. Introduction
such as the composition of the fuel and the
advanced catalyst design, this disadvantage can
now be satisfactorily overcome [3,4].
To obtain a more fundamental knowledge
regarding the processes that take place in auto-
motive catalysis, a number of surface science
To meet the stringent exhaust emission regula-
tions most cars are equipped with a so-called three-
way catalyst [1,2]. For a long time a catalyst based
on platinum and rhodium was used to simulta-
neously oxidise the CO and unburned hydro-
studies have been dedicated to the CO–O and
2
carbons and reduce the NO . However, in 1992,
85% of the total world supply of rhodium and
CO–NO reactions over palladium (see, for exam-
x
ple, [5–15]). However, only a few papers exist on
43% of platinum were used for automobile cata-
lysts [3]. Therefore, interest in the development of
a palladium-only catalyst, and, due to the unique
capability of rhodium to reduce NO, a Pd/Rh
catalyst has grown. Not only is palladium less
scarce and, hence, more economically attractive, it
also has favourable oxidation activity at low tem-
peratures and a high thermal stability. A major
drawback of palladium in the past has been its
high sensitivity to lead and sulfur poisoning.
However, due to improved boundary conditions,
the NO–H reaction, mainly concentrated on the
Pd(100) surface [15–19].
2
In this paper we describe a study concerning
the NO–H reaction over Pd(111) single-crystal
2
surface. It represents a part of more a extended
research programme, with the main objective of
understanding non-linear processes on various pre-
cious metal surfaces. Non-linear phenomena, such
as oscillations, spatio temporal pattern formation
and hysteresis phenomena, have already been
observed by our group for the NO–H reaction
2
over Pt(100), Rh(111), stepped rhodium surfaces
consisting of (111) terraces, Ru(0001), Ir(100),
Ir(110) and stepped iridium surfaces with (100)
* Corresponding author. Fax: +31-71-5274451.
E-mail address: c.wolf@chem.leidenuniv.nl (C.A. de Wolf )
0039-6028/00/$ - see front matter © 2000 Elsevier Science B.V. All rights reserved.
PII: S0039-6028 ( 00 ) 00825-6

C.A. de Wolf, B.E. Nieuwenhuys / Surface Science 469 (2000) 196–203
197
terraces and (110) steps [20–23]. To our knowl-
edge, only oscillations in the rate of CO oxidation
over Pd(110) surface have been reported so far,
in addition to explosive reactions in NO and CO
coadsorption experiments over the three low-index
planes of palladium [5–8,11].
heat and cool cycle was performed directly after
the AES measurements and the MS intensities
were compared with previous measurements.
High-purity NO (99.5%) and H (99.999%)
2
gases were used without further purification. The
pressure readings were corrected for the ionisation
efficiencies of the ion gauge using relative sensitivi-
ties for NO and H to N of 1.3 and 0.46,
2
2
2. Experimental
respectively.
The measurements were performed in a stan-
dard ultrahigh vacuum (UHV) system equipped
with facilities for low-energy electron diffraction
(LEED), Auger electron spectroscopy (AES) and
a differentially pumped (60 l s−1) quadrupole mass
spectrometer. The base pressure of the system was
always less than 5×10−10 mbar.
3. Results
To create a basis for the study of the NO–H
2
reaction, the interaction of NO with Pd(111) was
studied first. Fig. 1 shows the thermal desorption
(TD) spectra of NO adsorbed at 300 K. NO
desorbs in a single peak, which shifts from 508 K
for low NO coverages to 488 K for the saturation
coverage. Small amounts of N and N O, esti-
mated as less than 5% at saturation, desorb around
495 K. Remarkably, the quantity of N and N O
that was formed did not change with increasing
NO coverage. The minor amounts of oxygen atoms
were removed by flash annealing to 1600 K as was
confirmed by AES [30].
The palladium sample was cut from a single-
crystal palladium rod by spark erosion and pol-
ished within 1° of the desired direction. The crystal
was spotwelded to a tantalum support and could
be heated resistively up to 1600 K. The temper-
ature was measured with a Pt–Pt/Rh thermocou-
ple, which was spotwelded to the back of the
crystal. The crystal was cleaned by repetitive cycles
of oxygen treatment, Ar+-ion sputtering and
flashing in UHV to 1600 K. The surface cleanliness
and structure were checked by AES and LEED.
2
2
2
2
During the reactions the NO and H pressures
2
were kept constant while the system was pumped
by a turbomolecular pump. Prior to the measure-
ments the NO pressure was stabilised for more
than 1 h and only the second and subsequent heat
and cool cycles for each H /NO ratio were
2
recorded. In this way the hysteresis phenomena
proved to be reproducible. The crystal was turned
in front of a small opening which gave access to a
small chamber containing the mass spectrometer
(MS) to increase the sensitivity and to minimise
the contributions of the edges of the crystal. For
NH , 16 amu was followed instead of 17 amu to
3
decrease the contribution of the cracking product
of H O (OH). The cracking pattern of N O was
2
2
estimated to be 100% at 44 amu 30% at 30 amu
and 10% at 28 amu.
The AES measurements were performed using
a single-pass cylindrical mirror analyser with an
incident electron energy of 2.0 kV. To check the
influence of the electron beam on the reaction, a
Fig. 1. TD spectra of NO (m/e=30), N (m/e=28) and N O
2
2
(m/e=44) after dosing various amounts of NO at 300 K. The
heating rate is 5 K s−1.

198
C.A. de Wolf, B.E. Nieuwenhuys / Surface Science 469 (2000) 196–203
The NO–H reaction over Pd(111) was studied
in the temperature regime between 300 and 750 K
around 620 K, while the NH and H O partial
2
3
2
pressures reach their maximum when the temper-
ature is further lowered to 530 K. Around 400 K
the NO can no longer dissociate on the surface
and the reaction is stopped.
at NO pressures ranging from 7.7×10−8 to
7.7×10−7 mbar and H /NO ratios varying from
2
1 to 54. The primary reaction products were N ,
2
NH , H O and small amounts of N O.
A typical example of the product formation
during a heating–cooling cycle in the reaction
As can be seen from Fig. 2, between roughly
470 K and 600 K small hysteresis phenomena in
the rate of H O and NH formation occur during
3
2
2
2
3
mixture is shown in Fig. 2 for a H /NO ratio of
the heating–cooling cycle for the NO–H reaction
2
2
2.8. The consumption of NO starts around 470 K
over Pd(111). Note that the gradual increase of
and leads to the formation of N , NH , H O and
the background pressures of H O and NO is
caused by the slow pumping rate of these gases.
2
3
2
2
minor amounts of N O. Around 550 K the NO
2
conversion and the formation of H O and NH
To examine the influence of the H /NO ratio
2
3
2
start to decrease, whereas the rate of N formation
is still increasing and reaches its maximum around
on the product formation, a series of heating–
2
cooling cycles were performed with increasing H
pressure. Fig. 3 shows the results for the heating
2
620 K. Around 750 K the surface has become
completely inactive with respect to the NO
conversion.
Upon cooling the NO conversion increases
immediately and reaches it maximum around
branch. No hysteresis phenomena in the N and
2
N O formation rate were observed for the ratios
2
studied. The rates of NH and H O formation
3
2
were always more pronounced on the cooling
branch and shifted to slightly lower temperatures
with respect to the heating branch.
540 K, and, simultaneously, the rates of N and
2
H O formation begin to increase. A minor amount
2
of N O starts to be formed around 670 K.
Accompanied by the increase in the rate of NH
formation, the N formation rate starts to decrease
From Fig. 3 it is clear that the onset of the
2
main N formation shifts from around 450 K at
3
2
H /NO=1 to ca. 600 K at H /NO=54. In addi-
2
2
2
tion to the aforementioned N maximum, a low-
temperature maximum is observed which becomes
2
more pronounced at higher H pressures. Precisely
2
in between the maxima in the rate of N formation,
2
a peak in the NH formation is observed which
3
shifts to higher temperatures upon increasing the
H /NO ratio.
2
The results in Fig. 3 were used to quantitatively
compare the reaction orders in H for the forma-
2
tion of N , NH and N O. For that purpose, the
2
3
2
slopes of the ln [product] versus ln[ p(H )] plots
were determined at a temperature at which all
2
products were formed simultaneously, i.e., at
567 K. The pressures of the product gases were
corrected for the background pressures of these
gases at zero catalytic activity.
The dependence of the three nitrogen-contain-
ing reaction products on the H pressure is shown
2
in Fig. 4. The order in H of the N formation
2
2
becomes strongly negative at H /NO ratios over
Fig. 2. The rates of NO conversion and N O, NH , H O and
2
2
3
2
14, whereas the order for NH formation is posi-
N formation during a heating–cooling cycle at an NO pressure
3
tive, but becomes less so at H /NO ratios over
2
2
of 7.7×10−7 mbar and an H /NO ratio of 2.8. The heating and
2
cooling rate is 1.6 K s−1.
2.8. The order in H of the N O formation was
2
2

C.A. de Wolf, B.E. Nieuwenhuys / Surface Science 469 (2000) 196–203
199
Fig. 3. The rates of N , NH and N O formation during the heating branch of a heating–cooling cycle at an NO pressure of
2
3
2
7.7×10−7 mbar and H /NO ratios of 1, 2.8, 5.6, 14, 25 and 54. The H /NO ratio increases in the direction of the arrows. The heating
2
2
rate is 1.6 K s−1.
Fig. 4. A schematic representation of the reaction orders in H at 567 K for the formation of N , NH and N O at various H /NO
2
2
3
2
2
ratios, based on the results of Fig. 3. The NO pressure was 7.7×10−7 mbar.
slightly negative over the whole range of measured
during the heating–cooling cycle, indicating that
O is removed by hydrogen as soon as it is
H /NO ratios.
2
ads
For a better understanding of the observed
formed. NO is not visible at this energy due to
ads
phenomena, AES measurements were performed
during the heating–cooling cycle. In this way some
knowledge on the nitrogen species present on the
surface during the reaction may be obtained. The
shielding.
At room temperature a N
sured whose intensity gradually diminishes upon
signal is mea-
381 eV
heating the surface. Around 520 K, N starts to
2
N
peak-to-peak height was measured and
leave the surface and at 560 K the N
peak
381 eV
381 eV
normalised against the Pd
peak to exclude
has completely vanished. On the cooling branch
279 eV
the influence of a fluctuating electron beam.
The results are shown in Fig. 5 for the heating–
cooling cycle at p(NO)=1.5×10−7 mbar and
the decrease in the rate of N formation, which
2
starts around 600 K, was accompanied by an
increase in the rate of NH formation. In addition,
3
H /NO=28. The O
signal remained very low
the N
signal starts to increase around 560 K,
2
510 eV
381 eV

200
C.A. de Wolf, B.E. Nieuwenhuys / Surface Science 469 (2000) 196–203
like steps or kinks. The N
recombines with
ads
N
or NO to form N or N O, respectively
ads
[25,30,31]. Since the amount of N and N O
ads
2
2
2
2
desorbing from the surface in this study proved to
be independent of the NO coverage, we conclude
that the dissociation can be mainly attributed to
the presence of some more active sites.
In the temperature regime above 370 K, NO
desorbs in one single peak [26,28,29]. Although
the peak temperature shifts to lower temperatures
upon increasing the NO coverage, the desorption
of NO can be considered as a first-order process,
the temperature shift being caused by a coverage-
dependent activation energy for desorption
[24,25,28,30]. Using this assumption, the activa-
tion energy for desorption and the pre-exponential
factor in the low coverage limit were calculated by
the Chan-Aris-Weinberg method [33]. A value of
100 kJ mol−1 was found with a pre-exponential
factor of 1×109.6 s−1, which is within the range
of the values calculated by Ramsier et al. [30] who
found pre-exponential factors ranging from 1017
to 108 s−1 and activation energies of ca 180 to
90 kJ mol−1 depending on the NO coverage.
According to these authors, strong repulsive inter-
actions between the adsorbed NO molecules lower
the kinetic parameters at higher NO coverages.
Therefore, it is likely that the highest value of the
activation energy reported by Ramsier et al. was
determined at a lower NO coverage in comparison
with the lowest exposure of Figure 1. The differ-
ence may also be partly attributed to a higher
concentration of defect sites or to the presence of
Fig. 5. The N
/Pd
279 eV
formation rates at an NO pressure of
AES signal ratios in addition to the
381 eV
2
NH and
N
3
1.5×10−7 mbar and an H /NO ratio of 28. The solid symbols
2
represent the heating branch and the open symbols represent
the cooling branch.
thus deviating from the heating branch. The rate
of NH formation reaches its maximum around
3
500 K. At the same temperature the growth of
residual O on the crystal used in this study. This
ads
the N
N
heating branch.
signal starts to level. Around 410 K the
is supported by the results of Schmick et al. [25]
381 eV
signal collides again with the one of the
who found values for the activation energy of
desorption ranging from ca 146 to 71 kJ mol−1 on
a stepped Pd(111) surface exhibiting average ter-
race widths of 16 atoms that was pre-exposed to
381 eV
No indication of any oscillatory behaviour was
found under the conditions studied.
0.75 L O .
2
Clearly, the high desorption temperature and
low dissociation probability of NO influence the
4. Discussion
rate of the NO–H reaction at low temperatures.
As can be seen in Fig. 2, no reaction takes place
2
The interaction of NO with Pd(111) has pre-
viously been investigated by several authors [24–
32]. From these studies it was concluded that NO
is molecularly adsorbed in the temperature range
between 100 and 373 K. Upon heating only minor
dissociation takes place, presumably at active sites
until some NO molecules desorb from the surface
around 500 K. The creation of vacant sites allows
NO to dissociate into N and O . Subsequently,
ads
ads
N
is formed in addition to the formation of
2
N O, via N +NO N O(g). The formation
ads ads
2
2

C.A. de Wolf, B.E. Nieuwenhuys / Surface Science 469 (2000) 196–203
201
of intermediate NO dimers, which was suggested
for Pd(100) [34] and Pd(111) [25], was excluded
for Pd(111) on the basis of more recent
high-energy electron energy-loss spectroscopy
on the surface is simply to short to dissociate at
temperatures above its desorption temperature.
On lowering the temperature, the NO consump-
tion rate increases again, resulting in the formation
of N and H O. Around 660 K the temperature is
(HREELS) results [28]. The formed O immedi-
ads
2
2
ately reacts with two hydrogen atoms to form
low enough to allow NO to react with N to
ads
ads
H O(g) [15]. The origin of the atomic hydrogen
form N O. At a slightly lower temperature NH
also starts to be formed.
2
2
3
will be discussed later.
Below the desorption temperature of NO, the
To understand the hysteresis in the hydrogen-
major pathways to N formation over precious
metal surfaces are either the recombination of two
adsorbed nitrogen atoms, or a disproportionation
reaction between adsorbed NO and N [35]. Both
containing species, knowledge on the interaction
2
between H and Pd(111) is indispensable. Over
2
the years a lot of research has been performed on
this subject [47–54]. From these studies it was
concluded that hydrogen adsorbs dissociatively on
Pd(111) and can penetrate into the subsurface
region already at 90 K. At temperatures above
350 K diffusion from the subsurface to the bulk
can take place. Furthermore, it was found that in
the H O formation the reaction between O and
reactions were used to describe the N formation
2
over platinum, rhodium and platinum–rhodium
alloy surfaces [36–46], although in recent studies
on Rh(111) and Rh(100) the latter reaction path
was excluded [45,46]. The presence of a low-
temperature N peak in the TD spectrum shown
2
2
ads
in Fig. 1, which coincides with the NO desorption
H from the bulk is preferred over the reaction of
peak, might suggest that the disproportionation
reaction may play a minor role over Pd(111).
Further heating of the Pd(111) surface in the
O
with H adsorbed on the surface [19,55].
ads
Therefore, the rate of H O formation is dominated
by the transport of atomic hydrogen between the
2
reaction mixture leads to the formation of NH . As
bulk and the surface. The presence of major
3
can be seen in Figs. 3 and 4, the formation of N
amounts of O can slow down the diffusion rate
2
and NH are competitive processes in the temper-
3
ads
and, hence, the rate of H O formation [55].
2
ature regime between 500 and 700 K for the highest
Using this information the increased rate of
H /NO ratios and between 500 and 600 K for lower
NH and H O formation on the cooling branch
2
3
2
ratios. The rate of NH formation shows a positive
with respect to the heating branch can be under-
stood as follows. On the heating branch the
amount of bulk H is limited due to the low
temperature, which does not support bulk diffu-
sion, in combination with a high surface coverage
3
order in the H pressure. The higher positive order
2
at low H /NO ratios is caused by the temperature
shift of the maximum in the rate of NH formation
towards T=567 K. The order in H for the N
formation is slightly negative and becomes very
2
3
2
2
of NO and possibly some O
that hinders the
ads
negative when increasing amounts of NH start to
dissociative hydrogen adsorption. On the begin-
ning of the cooling branch, however, both the
temperature and the abundance of vacant sites
promote the diffusion of atomic hydrogen into the
3
be formed. Due to the removal of N by H, the
ads
formation of N O is also negatively influenced by
2
an increase in the H pressure.
2
For H /NO=2.8, the formation of N becomes
bulk. Since H O is mainly formed by the reaction
2
2
2
more favourable above 560 K. According to the
TDS results of Fig. 1, only minor amounts of
molecular NO are present on the surface at this
temperature. As a consequence, the formation of
of O with bulk H, an increased supply of bulk
ads
H will lead to a higher reaction rate. Most likely,
the same reaction path holds for the formation of
NH . Moreover, the fast removal of N
ads
creates vacant sites that promote the dissocia-
and
3
N O is becoming negligible and the only way to
O
2
2
ads
N
formation is by the recombination of N
tive adsorption of NO and, hence, increase the
total reaction rate.
ads
atoms. In addition, the rate of NO consumption
starts to decrease. This can be understood if one
considers the low tendency of NO to dissociate on
Pd(111). In other words, the residence time of NO
The presence of bulk H also explains the forma-
tion of NH already at H /NO ratios as low as 1
3
2
(Fig. 3).

202
C.A. de Wolf, B.E. Nieuwenhuys / Surface Science 469 (2000) 196–203
As can be seen in Fig. 5, a nitrogen-containing
species is built up on the surface during the NH
formation on the cooling branch. Unfortunately,
using conventional AES, no distinction can be
was performed under the auspices of NIOK, the
Netherlands Institute for Catalysis, lab. report no.
UL00-2-07.
3
made between N, NH and NO. However, since
x
NH is the only reaction product under these
References
3
conditions and the formation of NH , is consid-
2 ads
ered to be the rate-determining step in the NH
3
[1] K.C. Taylor, Catal. Rev. Sci. Eng. 35 (1993) 457.
[2] M. Bowker, R.W. Joyner, Insights Spec. Inorg. Chem
(1995) 145.
formation, this species can probably be assigned
to NH in combination with molecular NO [15].
This indicates that the dissociation of NO is faster
than the rate of NH formation under these
ads
[3] B.H. Engler, D. Lindner, E.S. Lox, A. Schafer-Sindlinger,
¨
K. Ostgathe, Stud. Surf. Sci. Catal. 96 (1995) 441 and
references therein.
3
conditions.
[4] J. Dettling, Z. Hu, K. Lui, R. Smaling, C.Z. Wan, A.
Punke, Stud. Surf. Sci. Catal. 96 (1995) 461.
[5] M. Ehsasi, M. Berdau, A. Karpowicz, K. Christmann, J.H.
Block, in: L. Guczi (Ed.), Proc. 10th ICC, 1992, Budapest,
New Frontiers in Catalysis A (1993) 321, and references
therein.
The nitrogen-containing species on the cooling
branch at temperatures lower than 420 K and the
nitrogen-containing species on the heating branch
of Fig. 5 are primarily connected to NO , because
now the dissociation rate of NO is rather slow.
ads
[6] N. Hartmann, K. Krischer, R. Imbihl, J. Chem. Phys. 101
(1994) 6717.
[7] S. Ladas, R. Imbihl, G. Ertl, Surf. Sci. 219 (1989) 88.
[8] M. Berdau, M. Ehsasi, A. Karpowicz, W. Engel, K.
Christmann, J.H. Block, Vacuum 45 (1994) 271.
[9] M. Bowker, I.Z. Jones, R.A. Bennett, S. Poulston, Stud.
Surf. Sci. Catal. 116 (1998) 431.
[10] B.N.J. Persson, Surf. Sci. Rep. 15 (1992) 1.
[11] M. Date´, H. Okuyama, N. Tagagi, M. Nishijima, T. Aruga,
Surf. Sci. 350 (1996) 79.
[12] S.M. Vesecky, D.R. Rainer, D.W. Goodman, J. Vac. Sci.
Technol. A 14 (1996) 1457.
[13] D.R. Rainer, S.M. Vesecky, M. Koranne, W.S. Oh, D.W.
Goodman, J. Catal. 167 (1997) 234.
[14] I. Kobal, K. Kimura, Y. Ohno, T. Matsushima, Surf. Sci.
445 (2000) 472.
[15] A. Obuchi, S. Naito, T. Onishi, K. Tamaru, Surf. Sci. 122
(1982) 235.
[16] T. Yamada, K. Tanaka, J. Am. Chem. Soc. 111 (1989)
6880.
5. Conclusions
After adsorption at 300 K, NO desorbs in one
single peak from the Pd(111) surface, with an
activation energy of desorption of 100 kJ mol−1
and a pre-exponential factor of 1×109.6 s−1. Only
minor amounts of N and N O were formed during
temperature-programmed desorption.
During the NO–H reaction over Pd(111), three
nitrogen-containing products were observed: N ,
2
2
2
2
NH and N O.
3
2
Hysteresis phenomena in the rate of NH and
3
H O formation were observed during a heating–
2
cooling cycle in the reaction mixture. The cooling
[17] C. Nyberg, P. Udval, Surf. Sci. 211/212 (1989) 923.
[18] I. Matsuo, J. Nakamura, H. Hirano, T. Yamada, K.
Tanaka, K. Tamura, J. Phys. Chem. 93 (1989) 7747.
[19] H. Neergaard-Waltenburg, Q. Gao, R.D. Ramsier, J.T.
Yates Jr., J. Phys. Chem. 98 (1994) 12075.
[20] P.D. Cobden, N.M.H. Janssen, Y. van Breugel, B.E. Nieu-
wenhuys, Faraday Disc. 105 (1996) 57.
[21] N.M.H. Janssen, P.D. Cobden, B.E. Nieuwenhuys,
J. Phys.: Condens. Matter 9 (1997) 1889.
[22] P.D. Cobden, C.A. de Wolf, M.Yu. Smirnov, A. Makeev,
B.E. Nieuwenhuys, J. Mol. Catal. 158 (2000) 115.
[23] C.A. de Wolf, M.O. Hattink, B.E. Nieuwenhuys, in
preparation.
branch showed a higher activity for the formation
of hydrogen-containing species than the heating
branch, probably due to the bigger supply of
atomic hydrogen dissolved in the bulk. Both
H O and NH are considered to be formed by the
2
3
reaction between bulk H and O , respectively,
ads
with N . No hysteresis phenomena were observed
ads
in the rate of N formation.
2
Acknowledgements
[24] H. Conrad, G. Ertl, J. Kuppers, E.E. Latta, Surf. Sci. 65
¨
(1977) 235.
NWO is gratefully acknowledged for financial
support. Part of the work was supported by
INTAS, reference number 99-01882. This work
[25] H.-D. Schmick, H.-W. Wassmuth, Surf. Sci. 123 (1982)
471.
[26] M. Bertolo, K. Jacobi, Surf. Sci. 226 (1990) 207.

C.A. de Wolf, B.E. Nieuwenhuys / Surface Science 469 (2000) 196–203
203
[27] M. Bertolo, K. Jacobi, Surf. Sci. 236 (1990) 143.
[28] D.T. Wickham, B.A. Banse, B.E. Koel, Surf. Sci. 243
(1991) 83.
[29] P.J. Chen, D.W. Goodman, Surf. Sci. 297 (1993) L93.
[30] R.D. Ramsier, Q. Gao, H. Neergaard-Waltenburg, K.-W.
Lee, O.W. Nooij, L. Lefferts, J.T. Yates Jr., Surf. Sci. 320
(1994) 209.
[41] T.W. Root, L.D. Schmidt, G.B. Fisher, Surf. Sci. 150
(1985) 173.
[42] P. Ho, J.M. White, Surf. Sci. 137 (1984) 103.
[43] S.B. Schwarz, G.B. Fisher, L.D. Schmidt, J. Phys. Chem.
92 (1988) 389.
[44] A. Siokou, R.M. van Hardeveld, J.W. Niemantsverdriet,
Surf. Sci. 402–404 (1998) 110.
[31] R.D. Ramsier, Q. Gao, H. Neergaard-Waltenburg, J.T.
Yates Jr., J. Chem. Phys. 100 (1994) 6837.
[32] D. Loffreda, D. Simon, P. Sautet, J. Chem. Phys. 108
(1998) 6447.
[45] D.N. Belton, C.L. DiMaggio, S.J. Schmieg, K.Y.S. Ng,
J. Catal. 157 (1995) 559.
[46] M.J.P. Hopstaken, J.W. Niemantsverdriet, J. Phys. Chem.
B 104 (2000) 3058.
[33] C.-M. Chan, R. Aris, W.H. Weinberg, Appl. Surf. Sci. 1
(1978) 360.
[47] H. Conrad, G. Ertl, E.E. Latta, Surf. Sci. 41 (1974) 435.
[48] G.E. Gdowski, T.E. Felter, R.H. Stulen, Surf. Sci. 181
(1987) L147.
[49] T. Engel, H. Kuipers, Surf. Sci. 90 (1979) 162.
[50] A. Noordermeer, G.A. Kok, B.E. Nieuwenhuys, Surf. Sci.
165 (1986) 375.
[51] O.M. Løvvik, R.A. Olsen, J. Chem. Phys. 104 (1996) 4330.
[52] W. Dong, G. Kresse, J. Hafner, J. Mol. Catal. A 119
(1997) 69.
[53] R.A. Olsen, P.H.T. Philipsen, E.J. Baerends, G.J. Kroes,
O.M. Løvvik, J. Chem. Phys. 106 (1997) 9286.
[34] C. Nyberg, P. Udval, Surf. Sci. 204 (1988) 517.
[35] B.E. Nieuwenhuys, Adv. Catal. 44 (1999) 259.
[36] H. Hirano, T. Yamada, K.I. Tanaka, J. Siera, P.D.
Cobden, B.E. Nieuwenhuys, Surf. Sci. 262 (1992) 97.
[37] J. Siera, B.E. Nieuwenhuys, H. Hirano, T. Yamada, K.I.
Tanaka, Catal. Lett. 3 (1989) 179.
[38] J. Siera, P. Cobden, K. Tanaka, B.E. Nieuwenhuys, Catal.
Lett. 10 (1991) 335.
[39] C.T. Campbell, J.M. White, Appl. Surf. Sci. 1 (1978) 347.
[40] T.W. Root, L.D. Schmidt, G.B. Fisher, Surf. Sci. 134
(1983) 30.
[54] R. Lober, D. Hennig, Phys. Rev. B 55 (1997) 4761.
¨
[55] T. Engel, H. Kuipers, Surf. Sci. 90 (1979) 181.