NH3 adsorption and decomposition on Ir(110): A combined temperature programmed desorption and high resolution fast x-ray photoelectron spectroscopy study

Article abstract of DOI:10.1063/1.1893690

The adsorption and decomposition of N H3 on Ir(110) has been studied in the temperature range from 80 K to 700 K. By using high-energy resolution x-ray photoelectron spectroscopy it is possible to distinguish chemically different surface species. At low temperature a N H3 multilayer, which desorbs at ~110 K, was observed. The second layer of N H3 molecules desorbs around 140 K, in a separate desorption peak. Chemisorbed N H3 desorbs in steps from the surface and several desorption peaks are observed between 200 and 400 K. A part of the N H3ad decomposes into N Had between 225 and 300 K. N Had decomposes into Nad between 400 K and 500 K and the hydrogen released in this process immediately desorbs. N2 desorption takes place between 500 and 700 K via Nad combination. The steady state decomposition reaction of N H3 starts at 500 K. The maximum reaction rate is observed between 540 K and 610 K. A model is presented to explain the occurrence of a maximum in the reaction rate. Hydrogenation of Nad below 400 K results in N Had. No N H2ad or N H3ad N H3 were observed. The hydrogenation of N Had only takes place above 400 K. On the basis of the experimental findings an energy scheme is presented to account for the observations.

Full text of DOI:10.1063/1.1893690

N H 3 adsorption and decomposition on Ir(110): A combined temperature programmed
desorption and high resolution fast x-ray photoelectron spectroscopy study
C. J. Weststrate, J. W. Bakker, E. D. L. Rienks, S. Lizzit, L. Petaccia, A. Baraldi, C. P. Vinod, and B. E.
Nieuwenhuys
Citation: The Journal of Chemical Physics 122, 184705 (2005); doi: 10.1063/1.1893690
View online: http://dx.doi.org/10.1063/1.1893690
View Table of Contents: http://scitation.aip.org/content/aip/journal/jcp/122/18?ver=pdfcov
Published by the AIP Publishing
Articles you may be interested in
Mechanism of ammonia decomposition and oxidation on Ir(110): A first-principles study
J. Chem. Phys. 138, 144703 (2013); 10.1063/1.4798970
Adsorption and reaction of NO on the clean and nitrogen modified Rh(111) surfaces
J. Chem. Phys. 131, 084704 (2009); 10.1063/1.3212596
A fast x-ray photoelectron spectroscopy study of the adsorption and temperature-dependent decomposition of
propene on Ni(100)
J. Chem. Phys. 115, 8133 (2001); 10.1063/1.1406502
Evidence of repulsive interactions between NO, O, and N on Ir(110). A fast x-ray photoelectron spectroscopy
study
J. Chem. Phys. 113, 10717 (2000); 10.1063/1.1323727
Nitrogen radical adsorption on InAs (001) surface studied by scanning tunneling microscopy and x-ray
photoelectronic spectroscopy
Appl. Phys. Lett. 73, 3754 (1998); 10.1063/1.122884
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
137.149.200.5 On: Sun, 23 Nov 2014 22:16:38

THE JOURNAL OF CHEMICAL PHYSICS 122, 184705 ͑2005͒
NH adsorption and decomposition on Ir„110…: A combined temperature
3
programmed desorption and high resolution fast x-ray photoelectron
spectroscopy study
C. J. Weststrate, J. W. Bakker, and E. D. L. Rienks
Leids Instituut voor Chemisch Onderzoek, Universiteit Leiden, P.O. Box 9502, Einsteinweg 55, 2333 CC
Leiden, The Netherlands
S. Lizzit and L. Petaccia
Sincrotrone Trieste, Strada Statale 14 km 163.5, I-34012 Basovizza, Italy
A. Baraldi
Physics Department and Center of Excellence for Nanostructured Materials, Trieste University, Via Valerio
2, I-34127 Trieste, Italy and Laboratorio TASC-INFM, Strada Statale 14 Km 163.5, I-34012 Trieste,
Italy
a͒
C. P. Vinod and B. E. Nieuwenhuys
Leids Instituut voor Chemisch Onderzoek, Universiteit Leiden, P.O. Box 9502, Einsteinweg 55, 2333 CC
Leiden, The Netherlands and Schuit Institute of Catalysis, Technische Universiteit Eindhoven, P.O.
Box 5600 MB, Eindhoven, The Netherlands
͑
Received 2 November 2004; accepted 22 February 2005; published online 9 May 2005͒
The adsorption and decomposition of NH on Ir͑110͒ has been studied in the temperature range
3
from 80 K to 700 K. By using high-energy resolution x-ray photoelectron spectroscopy it is
possible to distinguish chemically different surface species. At low temperature a NH multilayer,
3
which desorbs at ϳ110 K, was observed. The second layer of NH molecules desorbs around
3
1
40 K, in a separate desorption peak. Chemisorbed NH desorbs in steps from the surface and
3
several desorption peaks are observed between 200 and 400 K. A part of the NH3ad decomposes into
NHad between 225 and 300 K. NHad decomposes into Nad between 400 K and 500 K and the
hydrogen released in this process immediately desorbs. N desorption takes place between 500 and
2
7
00 K via Nad combination. The steady state decomposition reaction of NH starts at 500 K. The
3
maximum reaction rate is observed between 540 K and 610 K. A model is presented to explain the
occurrence of a maximum in the reaction rate. Hydrogenation of Nad below 400 K results in NHad.
No NH2ad or NH3ad/NH were observed. The hydrogenation of NH only takes place above 400 K.
3
ad
On the basis of the experimental findings an energy scheme is presented to account for the
observations. © 2005 American Institute of Physics. ͓DOI: 10.1063/1.1893690͔
I. INTRODUCTION
sible to generate hydrogen without CO contamination for
application on a small scale, like the PEM fuel cell in auto-
mobiles.
NH3 oxidation has been the subject of many fundamen-
tal surface science studies. In particular Pt surfaces have
been studied, because Pt is the main component of the cata-
lyst used in the Ostwald process. The adsorption and disso-
Ammonia is an industrially important molecule, which is
produced in large quantities by the chemical industry. It is
also formed as an unwanted byproduct in several industrial
processes. NH is used for the production of nitric oxide,
3
which is converted to nitrates for use in fertilizers. NH is
3
ciation of NH on Pt surfaces have been investigated using
3
also used in electricity plants, where it selectively reduces
NO to form N and H O. Ammonia itself is a harmful sub-
different experimental techniques such as temperature pro-
grammed desorption ͑TPD͒, x-ray photoelectron spectros-
copy ͑XPS͒, high resolution electron energy loss spectros-
copy ͑HREELS͒, and reflection absorption infrared
spectroscopy ͑RAIRS͒. In these studies different stable sur-
face species were reported, including NH3ad, NH2ad, NHad,
x
2
2
stance, causing eutrophication of lakes and seas, and its re-
lease in the environment is unwanted. An accurate descrip-
tion of the interaction of the NH molecule with catalytically
3
active surfaces is important to understand both ammonia oxi-
dation at low temperatures, forming ideally N and H O, and
1
–5
2
2
and Nad.
for the oxidation at high temperatures forming H O and
Next to Pt some Ru surfaces have been studied as well,
since Ru is a good catalyst for NH3 synthesis. TPD and
2
NO . Another reaction of interest is the NH decomposition
x
3
HREELS indicated the presence of NH3ad, NH2ad, NH , and
into its elements ͑N and H ͒ and via this reaction it is pos-
ad
2
2
Nad ͑Refs. 6 and 7͒ on several crystal planes.
8
Van den Broek et al. have shown that Ir based catalysts
a͒Electronic mail: b.nieuwe@chem.leidenuniv.nl
are more selective towards N during NH oxidation than Pt
2
3
0
021-9606/2005/122͑18͒/184705/8/$22.50
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
37.149.200.5 On: Sun, 23 Nov 2014 22:16:38
122, 184705-1
© 2005 American Institute of Physics
1

184705-2
Weststrate et al.
J. Chem. Phys. 122, 184705 ͑2005͒
based catalysts. Only a few authors have used the surface
since our experiments were performed using a shielded mass
spectrometer. In this setup it is very unlikely that electrons
generated by the mass spectrometer ionizer can reach the
sample.
science approach to study the chemistry of NH on Ir
3
9
–13
surfaces.
In a previous fast-XPS study the reaction be-
tween NO and H was studied on Ir͑110͒. Next to Nad and
NO, another species attributed to NH ͑x=1 or 2͒ was de-
1
4
2
1
8,19
High-energy resolution fast XPS measurements
were
x
tected in the N 1s signal.
performed on the SuperESCA beam line of ELETTRA, the
synchrotron radiation facility in Trieste, Italy. The vacuum
system is equipped with a hemispherical electron energy ana-
lyzer, a sputter gun for sample cleaning, a mass spectrometer,
and LEED optics. The sample can be cooled to ϳ80 K. The
In an earlier paper by our group the interaction of NH3
with Ir͑110͒ was reported, starting at room temperature. In
particular, TPD and temperature programmed reaction ͑TPR͒
were used to study nitrogen desorption, the effect of oxygen
11
−10
and the products of NH oxidation. In the present paper a
base pressure of this system is ϳ1ϫ10 mbar. The Ir͑110͒
3
+
more detailed high-energy resolution fast XPS and TPD/TPR
study is presented. We studied the adsorption and dissocia-
crystal was cleaned with Ar sputtering and annealing cycles
͑ϳ1200 K͒ followed by oxygen treatments at 700–1000 K
and a subsequent hydrogen treatment to remove residual
oxygen. Hydrogen was removed by a final flash to 700 K.
The cleanliness of the sample was checked by XPS, showing
no oxygen and carbon contamination in the O 1s and C 1s
core level regions. The N 1s spectra were measured with a
photon energy of 496 eV. For the temperature programmed
tion of NH on Ir͑110͒ between 80–700 K. Our results also
3
include isotope exchange experiments and the investigation
of steady state NH decomposition over Ir͑110͒.
3
During the interaction of NH with a surface the follow-
3
ing reactions are possible:
NH ͑g͒ ꢀ NH ,NH bilayer/ice,
͑1͒
͑2͒
͑3͒
͑4͒
͑5͒
͑6͒
3
3ad
3
20
XPS ͑TP-XPS͒ measurements a heating rate of 10 K/min
was used.
NH3ad ꢀ NH2ad + Had,
NH2ad ꢀ NH + H ,
The XP spectra were evaluated by fitting the peaks with
Doniach–Šunjić functions convoluted with a Gaussian
ad
ad
21
function. Core level binding energies of the different spe-
cies were measured with respect to the Fermi level.
NH ꢀ N + H ,
ad
ad
ad
Sun et al. and Schwaner et al. have shown that an ad-
5
,17
sorbed NH layer is sensitive to radiation damage. In their
3
2
2
H → H ͑g͒,
ad
2
experiments 50 eV electrons were used to generate NHxad
species on Ag͑111͒ and Pt͑111͒. In our XPS measurements
the surface is exposed to a high x-ray photon flux, so we
expect some radiation damage as well. Furthermore, the pho-
toelectrons have a kinetic energy of ϳ100 eV, i.e., an energy
range where it is likely that they induce dissociation.
N͑H ͒ → N ͑g͒ ͑+ xH ͒.
x ad
2
2
Equation ͑1͒ describes the NH adsorption/desorption
3
equilibrium. Depending on temperature different phases of
NH3ad are observed. Below ϳ120 K NH3 can form a
multilayer, while adsorption between 120 and 150 K results
During the TPD experiments both ND and NH were
3
3
1
5
in the formation of a double layer. Above 150 K only one
used. Both forms of ammonia gave similar results. Some of
layer of NH adsorbs, which is directly coordinated to the
3
the presented material was obtained using ND . This is indi-
3
metal surface. Dissociative reaction steps, described in Eqs.
cated in the text by the use of ND /ND
instead of
3
xad
͑
2͒–͑4͒, do not take place on all metal surfaces. In particular,
NH /NHxad. For the XPS measurements NH3 was used.
3
4,16
on Pt dissociation does not take place
below the NH3
Both NH and ND were dosed via background dosing.
3
3
desorption temperature. For Ir, on the contrary, NH disso-
3
We studied the Ir͑110͒ surface, which presents a “miss-
ing row” reconstruction in the clean state. As observed by
LEED, the substrate surface geometry is not affected by the
presence of NHx.
ciation has been observed on the ͑110͒, ͑510͒, and the ͑100͒
9
,11,12
crystallographic planes.
Equations ͑5͒ and ͑6͒ summa-
rize the possible mechanisms for N and H desorption.
2
2
II. EXPERIMENT
III. RESULTS AND DISCUSSION
The temperature programmed desorption/reaction mea-
surements were performed at the Leiden Institute of Chem-
istry. The vacuum system is equipped with a differentially
pumped, shielded mass spectrometer, to reduce the contribu-
tion of the heating wires and the edges of the crystal. The
sample is placed in front of a 2 mm wide hole at a distance
of 2 mm. The system is equipped with a sputter gun for
sample cleaning, low-energy electron diffraction ͑LEED͒ op-
tics and an Auger/XPS system. The base pressure of the sys-
tem is ϳ5ϫ10− mbar. The sample holder can be cooled
A. NH3 „/ND … desorption and decomposition:
3
An adsorbed layer
Figures 1͑a͒–1͑c͒ show the thermal behavior of an ad-
sorbed ammonia layer after adsorption at ϳ100 K. Part ͑a͒
shows the molecular desorption of ND , while parts ͑b͒ and
3
͑c͒ show the desorption of the dissociation products, D /HD
2
and N . Figure 1͑d͒ shows the XPS results obtained during
2
the heating of an adsorbed NH layer.
3
10
The molecular ND desorption shows several peaks. At
3
with liquid N to ϳ100 K. Different heating rates were used
118 K a large desorption peak is observed, which grows con-
tinuously with increasing exposure. This indicates that it cor-
2
in the range of 0.4–5 K/s.
5
Radiation damage as described by Sun et al. and
responds to desorption of a ND multilayer ͑ice͒. Next to this
3
1
7
Schwaner et al. can be excluded in this vacuum system,
multilayer peak a shoulder is observed at 140 K, which ini-
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
37.149.200.5 On: Sun, 23 Nov 2014 22:16:38
1

184705-3
NH3 adsorption and decomposition on Ir͑110͒
J. Chem. Phys. 122, 184705 ͑2005͒
present at 280, 375, and 440 K. Both the 187 K peak and the
three broad ND desorption peaks are assigned to desorption
3
of ND from the first ͑chemisorbed͒ layer, but we cannot
3
exclude that the small component at 440 K is due to recom-
bination of NDxad and D.
The ND desorption from the Ir͑110͒ surface is similar
3
to the NH desorption from Pt͑110͒ ͑Ref. 4͒ ͑heating rate
3
1
4 K/s͒. Gohndrone et al. show that NH desorption from
3
Pt͑110͒ shows desorption peaks at very similar temperatures.
The NH desorption at 180 K, labeled ␣Ј by Gohndrone et
3
2
al. ͓only present on Pt͑110͒ and Pt͑211͔͒, is also observed on
Ir͑110͒. Several of the ND desorption peaks correspond
3
quite well to those reported for NH desorption from Ir͑100͒
3
9
by Santra et al. Especially the high temperature NH de-
3
sorption peaks, at 280 K and 375 K, are similar for both Ir
surfaces.
The formation of D /HD and N ͓Figs. 1͑b͒ and 1͑c͔͒
2
2
11
shows that ND dissociates on Ir͑110͒. NH decomposition
3
3
was also observed on Ir͑100͒ ͑Refs. 9 and 10͒ but not on
1
3
Ir͑111͒.
Due to the presence of H in the residual gas it is nec-
2
essary to monitor both HD and D desorption to make sure
2
that all D that desorbs is detected. Figure 1͑b͒ shows that the
HD desorption maximum is at 370 K. The HD desorption
also shows a small desorption peak at 470 K. The D desorp-
2
tion peak has a maximum at 470 K, next to a smaller desorp-
tion peak at 370 K.
The desorption peak at 370 K is similar to normal hy-
2
4
drogen desorption from Ir͑110͒. This indicates that hydro-
gen combination is the rate determining step for this desorp-
tion peak. In this temperature regime Had and Dad coexist,
which results in HD as the major product. The desorption
peak at 470 K only occurs after ammonia adsorption. This
indicates that hydrogen formed at this temperature is due to
NDxad decomposition, and that NDxad decomposition reac-
tion is the rate determining step. Above 400 K there is no H
on the surface, so D is the main product. An evaluation of
2
the amount of D desorbing as HD compared to D desorption
2
shows that 60% of the D desorbs as HD, while 40% desorbs
as D . If we assume that HD is formed via D from the first
2
ND decomposition steps ͓Eqs. ͑2͒ and ͑3͔͒, and that D is
3
2
formed in the last decomposition step ͓Eq. ͑4͔͒, we can as-
sign the D desorption peak at 470 K specifically to ND
2
ad
decomposition. This assignment is confirmed by the XPS
measurements ͓Figs. 1͑d͒ and 2͔, which show that NHad de-
composition takes place around 420 K.
A comparison with hydrogen desorption due to NH de-
3
composition on Ir͑100͒ ͑Ref. 9͒ shows that on Ir͑110͒ the
hydrogen desorption due to NH decomposition occurs at a
3
similar temperature.
FIG. 1. TPD after ND adsorption at 100 K. ͑a͒ Molecular ND desorption
The N ͓Fig. 1͑c͔͒ desorption spectra from the Ir͑110͒
3
3
2
11
͑
͑
3 K/s͒, ͑b͒ HD/D2 desorption ͑5 K/s͒ and ͑c͒ N2 desorption ͑5 K/s͒.
have been reported previously. N desorbs between 600
2
d͒ TP-XPS during heating of an adsorbed NH layer ͑0.4 K/s͒.
3
and 700 K. The observed desorption temperature is higher
9
than the 500 K reported by Santra et al. for Ir͑100͒. This
tially appears after an exposure of ϳ1 L, and saturates
indicates that Nad is more strongly bound to the Ir͑110͒ sur-
face than to the Ir͑100͒ surface.
at ϳ2 L. This desorption feature is assigned to a ND double
3
layer, which is hydrogen bonded to the ND molecules in the
The results presented in the Figs. 1͑b͒ and 1͑c͒ show that
the mechanism of N formation is via N combination rather
3
1
5,22,23
first layer.
Another sharp ND desorption peak is ob-
3
2
ad
served at 187 K. Three broad ND desorption peaks are
than via NDxad combination. In fact N desorption is ob-
3
2
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
37.149.200.5 On: Sun, 23 Nov 2014 22:16:38
1

184705-4
Weststrate et al.
J. Chem. Phys. 122, 184705 ͑2005͒
spond to the surface species are also shown in this table.
Using these six components the series of spectra can be fit-
ted. The result of this fit is shown in Fig. 1͑d͒.
At 80 K the XP spectrum shows only one broad peak at
4
00.2 eV. This peak is assigned to a multilayer of NH . Due
3
to the high exposure ͑50 L͒ the multilayer is the only species
that is observed. Signal from the first and second layer is
effectively shielded. When the surface is heated the NH3
multilayer rapidly desorbs, in agreement with the ND de-
3
sorption at 118 K ͓Fig. 1͑a͔͒. The XP spectrum changes after
desorption of the NH multilayer, and a new peak appears at
3
4
00.8 eV. Based on literature reports about the special sta-
1
5,22,23
bility of an NH double layer,
we assign this new XPS
3
peak to the molecules in the second layer. At 140 K the
double layer, which partly screens signal coming from the
first ͑chemisorbed͒ layer, is removed. The desorption of the
second layer is observed as a shoulder in the ND desorption
3
trace. The XP spectrum after desorption of the double layer
can either be described as a convolution of two peaks at
4
00.1 eV and 399.5 eV ͓the result of this approach is shown
in Fig. 1͑d͔͒, or as a single peak which shifts from
00.1 eV to 399.5 eV. In the following part the spectra will
4
be interpreted as consisting of two peaks. Both components
at 400.1 eV and 399.5 eV are assigned to molecular forms of
chemisorbed NH . The 400.1 eV species desorbs around
3
2
00 K and is also observed in the TPD experiment as a
rather sharp ND desorption peak around 187 K ͓Fig. 1͑a͔͒.
3
The 399.5 eV species is very stable and is observed up to
FIG. 2. Six different surface species observed during TP-XPS of a surface
ϳ450 K. This is in line with the TPD experiment, in which
exposed to 50 L of NH at 80 K ͑0.4 K/s͒. The markers show the measured
3
molecular ND desorption was observed up to 440 K.
data points, the line through the data points represents the sum of the fitting
components.
3
By using a different assignment we could alternatively
attribute the 400.1 eV peak to NH3ad. The 399.5 eV peak
could be assigned to NH2ad. It would mean that the thermal
served only when D is completely desorbed. This means
2
desorption of ND above 200 K is due to recombination of
3
that the reaction described in Eq. ͑6͒ can be simplified to Eq.
ND2ad and D , rather than molecular desorption. This would
ad
͑7͒,
imply that all NH3ad readily dissociates on Ir͑110͒, around
200 K. This is very unlikely, since on Ru ͑a more reactive
2
N → N ͑g͒.
͑7͒
ad
2
metal͒ NH dissociation starts at 280 K, as was reported by
3
This is in contrast to experimental findings on an Ir field
6
Jacobi and co-workers. On Ir͑100͒ only a small part of the
2
5
emitter by Gorodetskii et al. They presented evidence for
9
adsorbed NH decomposes, and on Ir͑111͒ NH does not
3
3
NHad combination as the main mechanism for N formation.
13
2
dissociate at all. Furthermore, even on Pt͑110͒ NH desorp-
3
Figure 2 shows the N 1s XP spectra during a TP-XPS
experiment of an adsorbed multilayer of NH3.
The upper panel shows selected spectra in which the
fitting components are indicated. Table I shows the assign-
ment of the different binding energy ͑BE͒ components that
are observed in the spectra. The desorption peaks that corre-
tion was observed up to 450 K, although NH does not dis-
3
4
sociate on this surface. Therefore, we conclude that two
different NH3ad species are present on Ir͑110͒ at low tem-
perature, with different N 1s BEs as observed by means of
XPS.
Above 200 K only three surface components are ob-
served in the N 1s spectrum. Two new peaks appear at
397.6 eV and 396.5 eV.
TABLE I. Binding energies and TP-desorption maxima found for NH ad-
3
sorption
on Ir͑110͒.
The peak at 396.5 eV is assigned to N . N can be
ad
ad
1
4
formed on Ir͑110͒ via NO dissociation, which indeed re-
sults in a N 1s component at 396.5 eV. The decrease in the
Species
N 1s BE ͑eV͒
TPD maxima ͑K͒
Nad signal also corresponds to N desorption shown in the
2
NH ice
400.2
400.8
400.1
399.5
397.6
396.5
118
140
3
TPD experiment ͓Fig. 1͑c͔͒.
NH second layer
3
The 397.6 eV species is already observed at low tem-
perature, even during uptake experiments at ϳ80 K. We as-
sume that intense x-rays, rather than the surface chemical
properties, are responsible for the dehydrogenation of NH3ad
below 200 K. Uptake experiments done at different tempera-
NH3ad ͑1͒
NH3ad ͑2͒
NHad
187
280, 375, 440
͑470͒
650
Nad
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
37.149.200.5 On: Sun, 23 Nov 2014 22:16:38
1

184705-5
NH3 adsorption and decomposition on Ir͑110͒
J. Chem. Phys. 122, 184705 ͑2005͒
ture show that thermal decomposition takes place between
25 and 300 K. In these experiments the growth of NHxad is
2
faster than what we expect from radiation induced dissocia-
tion.
We assign the 397.6 eV species to NH . This assign-
ad
ment is based on the BEs of the different NHxad species. The
⌬
E between Nad and NHad is ϳ1 eV, the ⌬E between NHad
and NH is ϳ2 eV. According to the assignment by Sun et
al., the ⌬E between different NH species is 1 eV. We
3
ad
5
xad
expect that NH2ad has a BE of ϳ398.6 eV, between those of
NH3ad and NH . From the XPS results we conclude that
ad
NHad starts to decompose at 380 K. This corresponds nicely
to the onset of the NDad decomposition peak observed in the
TPD experiment. Some difference might occur in the exact
temperatures, since we used different heating rates in both
experiments ͑5 K/s and 0.4 K/s͒.
Our assignments lead to the conclusion that NH2ad is not
a stable surface species. Upon thermal formation it immedi-
ately decomposes further into NH . Semiempirical calcula-
ad
2
6,27
tions by Cholach et al.
have shown that NH2ad species
are not very stable on noble metals such as Pt, Ir, and Rh. A
2
8
comparable observation was reported by Fuhrmann et al.
for CHxad on Pt͑111͒. Only CH3ad and CHad were observed
as stable CHxad surface species.
We cannot exclude the presence of an unresolved NH2ad
component in the N 1s spectrum. It would be located be-
tween the NH3ad and the NHad peak, which coexist on the
surface between 200 and 400 K. In this case the intensity of
the NH2ad species is shared between NH3ad and NHad.
With XPS we found that NH3ad dissociation is inhibited
by NH . After the NH concentration has reached a maxi-
ad
ad
mum value, the remaining NH3ad does not dissociate any-
more ͑not even due to the x-ray beam͒. During further heat-
ing NH3ad just desorbs. To circumvent inhibition by NHad,
NH exposure was performed at different surface tempera-
3
FIG. 3. Steady state ND3 decomposition at different surface temperatures
tures. TPD shows that the subsequent N desorption does
2
and different ND pressures. ͑a͒ D production at different ND pressures.
3
2
3
depend on the surface temperature during uptake. Exposure
͑
b͒ Surface species during NH3 decomposition. ͑c͒ Results of a kinetic
at 100, 200, and 300 K result in similar N desorption, but
model: the surface coverage and the reaction rate are shown.
2
exposure at 400 and 500 K results in a larger amount of N2
formation. The NH exposure needed to saturate N desorp-
3
2
2
NH ͑g͒ → N ͑g͒ + 3 H ͑g͒.
͑8͒
tion is also much higher. At 100 K saturation was reached
after ϳ1 L, while at 500 K saturation was obtained after an
exposure of ϳ2 L. Exposure at 500 K leads to about three
times as much N desorption compared to N desorption fol-
3
2
2
Santra et al. showed that a steady state NH decomposi-
3
9,10
tion reaction can take place on Ir͑100͒. We also found that
2
2
it is possible to have a steady state ND ͑or NH ͒ decompo-
3
3
lowing exposure at 100 K.
sition reaction on Ir͑110͒, at different ND pressures ͑5
3
Hydrogen desorption after an uptake at 500 K is very
low. All NDad that forms can decompose further during ex-
posure into N . Thus inhibition of NH decomposition by
−8
−6
ϫ10 –4ϫ10 mbar͒. Figure 3͑a͒ shows the D formation
2
for different ND pressures. The experiment was performed
3
ad
3ad
in two different ways, both with the same result. The isother-
mal reaction was studied by keeping the temperature of the
crystal constant for a few minutes. This results in a steady
state N and D production. The temperature was increased
in several steps. The data obtained in this way is shown as
markers in Fig. 3. The solid lines in Fig. 3 are the result of a
subsequent measurement. In this experiment the temperature
NHad is circumvented, and the maximum Nad coverage, in-
stead of the maximum NHad coverage, is the limiting factor
for the amount of N formation.
2
2
2
The N desorption temperature decreases when the N
2
ad
concentration is higher. This indicates that N formation is a
2
second-order desorption process, described by Eq. ͑7͒.
was continuously decreased, while the D pressure was mea-
sured. Both types of experiments give the same result. This
means that a steady state reaction takes place in both cases.
2
B. NH „/ND … decomposition: A steady state reaction
3
3
2
9
Papapolymerou et al. reported that Ir catalysts catalyze
the steady state decomposition reaction of NH ͓Eq. ͑8͔͒
The results for different ND pressures show similar
thermal behavior. The reaction starts around 500 K. This co-
3
3
with higher efficiency than Pd, Pt, and Rh,
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
37.149.200.5 On: Sun, 23 Nov 2014 22:16:38
1

184705-6
Weststrate et al.
J. Chem. Phys. 122, 184705 ͑2005͒
TABLE II. Kinetic parameters used in the modeling of the NH decompo-
sition reaction.
incides with the onset of N desorption due to Nad combina-
3
2
tion ͑at a high Nad concentration͒. The reaction rate reaches a
maximum between 540 K and 610 K ͑depending on the
pressure͒ and goes down at higher temperature.
pNH ͑mbar͒
1ϫ10−7
3
S NH ͑is kept constant͒
1
0
3
Figure 3͑b͒ shows the TP-XPS during heating ͑0.4 K/s͒
−1
6ϫ1014
Ades NH3ad ͑s
Edes NH3ad ͑kJ/mol͒
Adec NH3ad ͑s−1
Edec NH3ad ͑kJ/mol͒
͒
−
8
in the presence of 5ϫ10 mbar NH . During the reaction
3
100
1ϫ1012
80
͑
above 500 K͒ the surface is covered with Nad and NH . At
͒
ad
the maximum reaction the NHad concentration is low. When
the temperature is further increased the NHad concentration
drops to zero, while the Nad concentration decreases as well.
Adec NHad ͑s−1
͒
1ϫ1011
Edec NHad ͑kJ/mol͒
110
Arecomb Nad ͑s−1
͒
1ϫ1011
122
From this measurement we conclude that the N formation
2
Erecomb Nad ͑kJ/mol͒
Heating rate ͑K/s͒
takes place via Eq. ͑7͒ rather than via Eq. ͑6͒.
0.85
On the basis of these results we propose the following
model to describe the NH decomposition reaction. Below
3
5
00 K, NHad and Nad inhibit steady state NH decomposi-
3
C. Hydrogenation of Nad and NHxad
tion. At 500 K Nad starts to desorb thus removing the inhi-
bition. The reaction rate initially increases with increasing
temperature because the rate of N desorption increases. At
higher temperature the NH3ad availability limits the reaction
rate. At these temperatures there is a competition between
NH3ad desorption and NH3ad decomposition. An increase of
the temperature favors the desorption pathway and as a con-
sequence the reaction rate drops. The model also explains the
We studied the formation of NHxad starting from Nad.
First a layer of Nad was prepared. The surface was exposed to
2
1
4,31
NO, which dissociates on Ir͑110͒.
The Oad was removed
by exposure to H . The N 1s and O 1s XPS signals were
2
monitored in situ during the hydrogen treatment, which made
it possible to remove all the Oad without hydrogenation of
N . Figure 4 shows the result of a subsequent exposure to
ad
−
7
3
observed effect of NH pressure. An increase of the NH3
H ͑5ϫ10 mbar͒ between 250–100 K. Hydrogenation of
2
pressure means and increase of the NH3ad equilibrium cov-
erage at a certain temperature. In that case NH3ad availability
starts to play a role at a higher temperature. A similar model
was proposed by Mortensen et al. They use their model to
explain the thermal behavior of the observed dissociative
Nad takes place between 200 K and 250 K. The only product
of Nad hydrogenation is NH . Further hydrogenation to-
ad
wards NH2ad or NH3ad does not take place.
7
The TP-XPS results of an adsorbed NH3 layer also
showed that NHad is the product of NH3ad decomposition.
This leads to the conclusion that NHad is the most stable
NHxad surface species. Several other authors such as Cholach
et al. ͓Ir͑110͔͒ ͑Refs. 26 and 27͒, Rod et al. ͓Ru͑0001͔͒,
and Tanaka et al. ͓Pd͑100͒, Rh͑100͒, and Pt-Rh͑100͔͒ ͑Ref.
sticking coefficient of NH on Ru͑0001͒.
3
To test the validity of our model we have written a set of
differential equations to describe the kinetics of the reaction.
The equations are based on the following steps:
3
2
3
3͒ confirm that NHad is the most stable NHxad species.
NH ͑g͒ ꢀ NH3ad,
͑9͒
͑10͒
͑11͒
͑12͒
3
Figure 5 shows the results of an isotope exchange ex-
periment between NH and D . A layer of NH ͑adsorbed
3
2
3ad
NH → → NH ,
at 150 K͒ was heated in the presence of D . This allows us to
3
ad
2
monitor the hydrogenation of the different NHxad species.
The data presented in Fig. 5 is corrected for the overlapping
cracking patterns of the different products. We assume that in
the ionizer the NH bond has an equal probability of being
NH → N ,
ad
ad
2
N → N ͑g͒.
ad
2
The real NH3ad desorption has several desorption peaks
͓
Fig. 1͑a͔͒, but for the model only the NH desorption peak
3
at the highest temperature is taken into account. The system
is further simplified by the assumption that adsorption is re-
versible, but the decomposition steps are irreversible. The
NH coverage is independent of N and NHad coverage. Nad
3
ad
and NHad occupy the same site.
The parameters shown in Table II are used in the model.
The temperature dependence is described by Arrhenius be-
havior. The activation energies were derived empirically by
finding the closest match to experimental observations. The
estimated prefactors were taken from Ref. 30.
The results of the model are shown in Fig. 3͑c͒. They
nicely reproduce experimental findings. The most important
result of our simulation is that it reproduces the maximum in
FIG. 4. Nad hydrogenation at low temperature. The temperature was de-
the reaction rate. An increase in pNH also results in a shift
3
creased in steps, while the N 1s signal was measured ͑H pressure
2
of the maximum rate to higher temperature.
5ϫ10−7 mbar͒.
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
37.149.200.5 On: Sun, 23 Nov 2014 22:16:38
1

184705-7
NH3 adsorption and decomposition on Ir͑110͒
J. Chem. Phys. 122, 184705 ͑2005͒
FIG. 6. A qualitative energy scheme proposed on the basis of our experi-
mental results.
FIG. 5. The result of an isotope exchange experiment, in which a NH
3
ad
−
6
layer is heated up in D ͑heating rate 5 K/s, D pressure 1ϫ10 mbar͒.
2
2
D. An energy scheme
broken as the ND bond. This assumption is supported by the
similarity of the cracking patterns of NH and ND .
The result presented in Fig. 5 can be divided into two
temperature regimes, i.e., below and above 400 K. Below
Based on our experimental findings we present the en-
ergy scheme depicted in Fig. 6. It qualitatively resembles the
energy scheme that was found by means of theoretical cal-
3
3
3
2,34
4
00 K only NH and NDH ͑and some ND H͒ are observed.
culations for Ru͑0001͒ by Rod et al. and Zhang et al.
3
2
2
Above 400 K all possible ND H ͑g͒ species are observed.
One of the main features of the scheme is that NHad is the
most stable species. The Nad species is less stable, which
accounts for the experimental finding that Nad is easily hy-
drogenated towards NH . NH decomposes when hydrogen
x
y
To understand this result we need to look at the forma-
3
4
tion of NH starting with N . Calculations by Zhang et al.
3
ad
for Ru͑0001͒ show that the NHad hydrogenation has a high
ad
ad
activation barrier, comparable to that of N dissociation. This
desorbs from the surface, thus shifting the equilibrium shown
2
is in line with our findings that NHad is the only species
formed upon Nad hydrogenation.
in Eq. ͑13͒ towards Nad and Had,
NH ꢀ N + H .
͑13͒
ad
ad
ad
For the NO+H reaction on Ir͑110͒ NH formation takes
2
3
3
1
place above 400 K. The onset of NH formation is not
NH2ad on the other hand is an unstable surface species
and it is even less stable than NH3ad. This explains why
NH2ad is not observed on the surface. The equilibrium con-
centration of NH2ad on the surface will be low at all times,
and NH2ad that is formed either decomposes into NHad or
hydrogenates to NH3ad. Another feature of the energy
scheme is that the energy difference between NH2ad ͑and
NH3ad͒ and NHad is large, and a high activation barrier is
associated with NHad hydrogenation. This explains why NH3
formation starting with Nad only takes place above 400 K,
when the energy barrier for NHad hydrogenation can be over-
come.
3
limited by NO dissociation, because XPS shows that NO
dissociates already at 300 K. We have shown that the Nad
hydrogenation towards NHad takes place below 250 K. This
means that the next hydrogenation steps have a significant
activation barrier and that NHad hydrogenation is the rate
determining step in the Nad hydrogenation towards NH . The
3
hydrogenation of NHad only takes place above 400 K, when
there is enough energy available to overcome the barrier.
When we apply these considerations to the isotope ex-
change experiment it means that below 400 K NH cannot
3
be formed due to hydrogenation of Nad or NH . Below
ad
4
00 K a large amount of NDH desorbs. This means that at
2
this low temperature another process, presumably involving
NH2ad, takes place. We propose the following model to ex-
IV. CONCLUSIONS
plain the NDH desorption below 400 K. At low temperature
The interaction of NH ͑/ND ͒ with Ir͑110͒ was studied
2
3
3
NH3ad and D coexist on the surface. In the preceding section
it was concluded that NH2ad is not a stable surface species,
but that does not mean that it cannot be formed. When NH2ad
between 80 K and 700 K. We used TPD/R and high-energy
resolution fast XPS, which enabled us to get a detailed un-
derstanding of the adsorption, desorption, and decomposition
processes taking place on the surface.
forms via NH decomposition, it will initially decompose
3
further to form NHad ͑this occurs between 225 and 300 K͒.
Below 110 K NH forms a multilayer ͑for exposures
3
At a certain NHad concentration further NH decomposition
Ͼ2 L͒. This multilayer is characterized by a single broad
peak at 400.2 eV in the N 1s region of the XP spectrum. The
multilayer desorbs at 118 K. A specific structure is stable
between 110 K and 150 K. This structure is described as a
3
is inhibited. Any NH2ad formed at this stage will immediately
hydrogenate ͑since NH3ad is more stable than NH2ad͒, but in
the presence of D there is a probability that NH D is
2
ad
formed. This process can explain the desorption of NDH2
observed in the experiment. When the surface is heated
above 400 K all possible deuterated products are observed,
which indicates that at these temperatures NHad and Nad hy-
drogenation takes place as well.
double layer, in which a second layer of NH is hydrogen
3
1
5,22,23
bonded to the first, chemisorbed layer of NH3.
Above
150 K two different molecular forms of NH3ad seem to be
present. One of them desorbs around 190 K, the other des-
orbs in three steps, at 280, 375, and 440 K.
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
37.149.200.5 On: Sun, 23 Nov 2014 22:16:38
1

184705-8
Weststrate et al.
J. Chem. Phys. 122, 184705 ͑2005͒
Both radiation damage and thermal dissociation result in
the formation of the NH dissociation products NH and
knowledged for technical support. A.B. acknowledges finan-
cial support from the MUIR and INFM under the programs
PRIN2003, FIRB, and PURS2001. C.J.W. acknowledges
STW for financial support.
3
ad
N . The NH species is not stable on the surface and is,
ad
2ad
therefore, not observed. Thermal NH decomposition into
3
NHad takes place between 225 K and 300 K. The hydrogen͑/
deuterium͒ from the NH ͑/ND ͒ decomposition desorbs at
1
D. G. Löffler and L. D. Schmidt, Surf. Sci. 59, 195 ͑1976͒.
3
3
2
J. L. Gland and E. B. Kollin, Surf. Sci. 104, 478 ͑1981͒.
3
70 K, the normal hydrogen desorption temperature from
Ir͑110͒. NHad decomposition into Nad takes place between
00 K and 500 K. Hydrogen͑/deuterium͒ from NHad ͑/NDad͒
3
W. L. Guthrie, J. D. Sokol, and G. A. Somorjai, Surf. Sci. 109, 390
͑1981͒.
4
4
J. M. Gohndrone, C. W. Olsen, A. L. Backman, T. R. Gow, E. Yagasaki,
and R. I. Masel, J. Vac. Sci. Technol. A 7, 1986 ͑1988͒.
Y.-M. Sun, D. Sloan, H. Ihm, and J. M. White, J. Vac. Sci. Technol. A 14,
decomposition desorbs at 470 K, in a separate desorption
peak.
5
1516 ͑1995͒.
6
7
Nitrogen desorbs at 650 K. This temperature is higher
Y. Wang, A. Laflosse, and K. Jacobi, Surf. Sci. 507-510, 773 ͑2002͒.
H. Mortensen, L. Diekhöner, A. Baurichter, E. Jensen, and A. C. Luntz, J.
Chem. Phys. 113, 6882 ͑2000͒.
9
than 500 K, the N desorption observed for Ir͑100͒. This
2
indicates that Nad binds stronger to the Ir͑110͒ surface. Nad
combination rather than NHxad combination is responsible
8
9
0
A. C. M. van den Broek, J. van Grondelle, and R. A. van Santen, J. Catal.
185, 297 ͑1999͒.
for N formation.
A. K. Santra, B. K. Min, C. W. Yi, K. Luo, T. V. Choudhary, and D. W.
Goodman, J. Phys. Chem. B 106, 340 ͑2001͒.
T. V. Choudhary, A. K. Santra, C. Sivadinaranaya, B. K. Min, C. W. Yi, K.
Davis, and D. W. Goodman, Catal. Lett. 77, 1 ͑2001͒.
S. A. C. Carabineiro and B. E. Nieuwenhuys, Surf. Sci. 505, 163 ͑2002͒.
S. A. C. Carabineiro and B. E. Nieuwenhuys, Surf. Sci. 532, 87 ͑2003͒.
R. J. Purtell, R. P. Merrill, C. W. Seabury, and T. N. Rhodin, Phys. Rev.
Lett. 44, 1279 ͑1980͒.
2
We found that NHad inhibits the dissociation of NH3ad
.
1
When the NHad concentration has reached its maximum the
NH3ad dissociation stops, and the remaining NH3ad desorbs
as molecular NH . The maximum concentration of NH de-
11
1
1
2
3
3
ad
termines the maximum amount of N that desorbs. Above
2
400 K inhibition effect of NHad can be circumvented, be-
14
C. A. de Wolf, J. W. Bakker, P. T. Wouda, B. E. Nieuwenhuys, A. Baraldi,
S. Lizzit, and M. Kiskinova, J. Phys. Chem. B 105, 4254 ͑2001͒.
G. B. Fisher, Chem. Phys. Lett. 79, 452 ͑1981͒.
cause NHad decomposes during the adsorption process. Thus
it is possible to obtain a three times higher concentration of
Nad on the surface.
1
1
5
6
A. Fahmi and R. A. van Santen, Z. Phys. Chem. ͑Munich͒ 197, 203
17͑
1996͒.
We observed NH3 ͑/ND ͒ decomposition over the
3
A. L. Schwaner, E. D. Pylant, and J. M. White, J. Vac. Sci. Technol. A 14,
Ir͑110͒ in a steady state reaction at different pressures. The
steady state reaction ͑forming N and H ͒ starts at 500 K.
1
453 ͑1995͒.
1
1
8
9
A. Baraldi, G. Comelli, S. Lizzit, M. Kiskinova, and G. Paolucci, Surf.
Sci. Rep. 49, 169 ͑2003͒.
2
2
The reaction rate reaches a maximum between 540–610 K.
At higher temperature the reaction rate decreases. We pro-
pose a kinetic model to explain the maximum in the reaction
rate. The key step in the model is that at high temperature
NH3ad availability limits the reaction. At high temperature
there is a competition between NH desorption and NH dis-
A. Baraldi, M. Barnaba, B. Brena, D. Cocco, G. Comelli, S. Lizzit, G.
Paolucci, and R. Rosei, J. Electron Spectrosc. Relat. Phenom. 76, 145
0͑
1995͒.
2
A. Baraldi, G. Comelli, S. Lizzit, D. Cocco, G. Paolucci, and R. Rosei,
Surf. Sci. 367, L67 ͑1995͒.
21
J. J. Joyce, M. del Guidice, and J. H. Weaver, J. Electron Spectrosc. Relat.
Phenom. 49, 31 ͑1989͒.
3
3
2
2
2
2
3
4
sociation. An increase in temperature favors the desorption
pathway, and the reaction rate decreases. A simple kinetic
model shows that this mechanism indeed results in a maxi-
mum in the reaction rate.
C. Benndorf and T. E. Madey, Surf. Sci. 135, 164 ͑1983͒.
D. R. Jennison, P. A. Schultz, and M. P. Sears, Surf. Sci. 368, 253 ͑1996͒.
D. E. Ibbotson, T. S. Wittrig, and W. H. Weinberg, J. Chem. Phys. 72,
254
885 ͑1980͒.
V. V. Gorodetskii and V. A. Sobyanin, Interaction of hydrogen, nitrogen
and ammonia with Iridium, Proceedings of the VII International Congr.
Catalysis, edited by T. Seiyama and K. Tanabe ͑Elsevier and Kodansha
LTD, Tokyo, 1980͒.556–577.
Nad hydrogenation takes place below 400 K. NHad is the
only product, the formation of NH2ad or NH3ad/NH ͑g͒ was
3
not observed. This indicates that NHad is the most stable
species on the surface.
26
A. R. Cholach, N. N. Bulgakov, and B. E. Nieuwenhuys, Catal. Lett. 86,
9
͑2003͒.
2
2
2
3
7
8
9
0
We present a qualitative energy scheme which summa-
rizes all the experimental findings. The main features of this
model are that NHad is the most stable NHxad species and that
NH2ad is very unstable. The model is similar to what was
A. R. Chloach, N. N. Bulganov, and B. E. Nieuwenhuys, Surf. Sci. 573,
64 ͑2004͒.
T. Fuhrmann, M. Kinne, C. M. Whelan, J. F. Zhu, R. Denecke, and H.-P.
Steinrück, Chem. Phys. Lett. 390, 208 ͑2004͒.
G. Papapolymerou and V. Bontozoglou, J. Mol. Catal. A: Chem. 120, 165
͑1997͒.
J. A. Dumesic, D. F. Rudd, L. M. Aparicio, J. E. Rekoske, and A. A.
Treviño, The Microkinetics of Heterogeneous Catalysis ͑American Chemi-
cal Society, Washington, DC, 1993͒.
2
found for NH on Ru͑0001͒ ͑Refs. 32 and 34͒ using density-
3
functional theory calculations.
3
3
1
C. A. de Wolf, J. W. Bakker, P. T. Wouda, B. E. Nieuwenhuys, A. Baraldi,
S. Lizzit, and M. Kiskinova, J. Chem. Phys. 113, 10717 ͑2000͒.
T. H. Rod, A. Logadottir, and J. K. Nørskov, J. Chem. Phys. 112, 5343
ACKNOWLEDGMENTS
2
The authors would like to thank ELETTRA and the E.U.
for financial support to perform measurements at ELETTRA.
C.J.W. acknowledges STW for financial support from the
project under Grant No. UPC.5037 R.C.V. van Schie is ac-
33͑
2000͒.
K. Tanaka, T. Yamada, and B. E. Nieuwenhuys, Surf. Sci. 242, 503
4͑
1991͒.
3
C. Zhang, Z.-P. Liu, and P. Hu, J. Chem. Phys. 115, 609 ͑2001͒.
This article is copyrighted as indicated in the article. Reuse of AIP content is subject to the terms at: http://scitation.aip.org/termsconditions. Downloaded to IP:
37.149.200.5 On: Sun, 23 Nov 2014 22:16:38
1