Selectivity-directing factors of ammonia oxidation over PGM gauzes in the Temporal Analysis of Products reactor: Secondary interactions of NH3 and NO

Article abstract of DOI:10.1016/j.jcat.2004.09.020

Factors that direct the selectivity of ammonia oxidation for NO were determined previously (J. Catal. 227 (2004) 90) by the investigation of primary NH₃-O₂ interactions over pure Pt and Pt-Rh (95-5) alloy gauzes at 973-1173 K in the temporal analysis of products (TAP) reactor. A solid mechanistic understanding of the processes leading to by-products (N ₂O and N₂) requires an analysis of secondary NH ₃-NO interactions, which we investigated in the TAP reactor with isotopically labeled molecules. Our experiments under transient vacuum conditions indicate that these secondary processes determine the reaction selectivity for N₂O and N₂ in high-temperature ammonia oxidation over noble metal catalysts. Adsorbed oxygen species initiate the reaction of ammonia with nitric oxide. N₂O originates from the coupling of ammonia intermediates (NH_x) and nitric oxide. Different reaction pathways leading to N₂ have been identified, including primary (NH₃ oxidation) and secondary (NH_x and H-assisted NO reduction) processes. The relative contributions of these routes depend on the surface coverage of nitrogen and hydrogen-containing species. A reaction scheme accounting for our experimental observations has been proposed, giving rise to an improved mechanistic description of the complex processes in ammonia burners.

Full text of DOI:10.1016/j.jcat.2004.09.020

Journal of Catalysis 229 (2005) 303–313
www.elsevier.com/locate/jcat
Selectivity-directing factors of ammonia oxidation over PGM gauzes
in the Temporal Analysis of Products reactor:
Secondary interactions of NH3 and NO
J. Pérez-Ramírez a,∗, E.V. Kondratenko b,V.A. Kondratenko b, M. Baerns b
a
Yara Technology Centre Porsgrunn, Catalysis and Nitric Acid Technology, PO Box 2560, N-3908, Porsgrunn, Norway
b
Institute for Applied Chemistry Berlin-Adlershof, Richard-Willstätter-Str.12, D-12489, Berlin, Germany
Received 22 July 2004; revised 14 September 2004; accepted 20 September 2004
Available online 24 December 2004
Abstract
Factors that direct the selectivity of ammonia oxidation for NO were determined previously (J. Catal. 227 (2004) 90) by the investigation
of primary NH –O interactions over pure Pt and Pt–Rh (95–5) alloy gauzes at 973–1173 K in the temporal analysis of products (TAP)
3
2
reactor. A solid mechanistic understanding of the processes leading to by-products (N O and N ) requires an analysis of secondary NH –
2
2
3
NO interactions, which we investigated in the TAP reactor with isotopically labeled molecules. Our experiments under transient vacuum
conditions indicate that these secondary processes determine the reaction selectivity for N O and N in high-temperature ammonia oxidation
2
2
over noble metal catalysts. Adsorbed oxygen species initiate the reaction of ammonia with nitric oxide. N O originates from the coupling
2
of ammonia intermediates (NH ) and nitric oxide. Different reaction pathways leading to N have been identified, including primary (NH
x
2
3
oxidation) and secondary (NH and H-assisted NO reduction) processes. The relative contributions of these routes depend on the surface
x
coverage of nitrogen and hydrogen-containing species. A reaction scheme accounting for our experimental observations has been proposed,
giving rise to an improved mechanistic description of the complex processes in ammonia burners.
2004 Elsevier Inc. All rights reserved.
Keywords: NH oxidation; NO reduction; Platinum; Rhodium; Gauze; Selectivity; N O; N ; Mechanism; Transient experiments; TAP reactor; Isotopes
3
2
2
1. Introduction
ing a high NO selectivity. The formation of N2O and N2
was also discussed in [1], but specific reaction pathways
leading to these by-products could not be fully assessed
merely through the study of the primary interaction of NH3
and O2.
Since both N2O and N2 molecules contain two nitrogen
atoms, one can anticipate that these reaction products should
largely originate from the recombination of N-containing
surface intermediates of NH3 and/or NO, as a result of de-
composition or coupling reactions involving one or both
species. In this sense, studies in the TAP reactor indicated
the high activity of reduced and oxidized PGM gauzes for
the direct decomposition of NO and NH3, respectively, and
the rapid reaction between NO and NH3 [1,2]. Accordingly,
a solid understanding of the complex reaction network in-
volved in ammonia oxidation in industrial burners requires
Our previous study using the Temporal Analysis of
Products (TAP) technique [1] has provided valuable in-
sight into the mechanism of high-temperature NH3 oxi-
dation over platinum group metal (PGM) catalysts. The
analysis of primary interactions between ammonia and oxy-
gen elucidated the process of ammonia activation over Pt
and Pt–Rh gauzes and the factors determining the reac-
tion selectivity for NO. Importantly, these results indicated
that not only the coverage but also the nature of oxygen
species adsorbed to the metal sites are important for obtain-
*
Corresponding authors.
E-mail address: javier.perez.ramirez@yara.com (J. Pérez-Ramírez).
0021-9517/$ – see front matter
doi:10.1016/j.jcat.2004.09.020
2004 Elsevier Inc. All rights reserved.

304
J. Pérez-Ramírez et al. / Journal of Catalysis 229 (2005) 303–313
an in-depth analysis of secondary interactions between the
desired reaction product (NO) and the reactant (NH3).
The mechanism and kinetics of NO reduction by NH3 in
the absence of oxygen (Eqs. (1)–(2)) have been investigated
over platinum catalysts of different natures, including single
crystals (Pt(100) [3,4] or stepped Pt 12(111) × (111) [5]),
polycrystalline Pt wire [6,7] and foils [8–10], and supported
Pt catalysts [11–13]. The process is highly exothermic and
has typically shown an oscillatory kinetics, presenting analo-
gies with the NO–H2 reaction over platinum surfaces [4,14].
The mechanism of these (isothermal) spontaneous oscilla-
tions has been studied at different molar feed NH3/NO ratios
(1–9) in a wide range of pressures (0.1–1000Pa) and temper-
atures (473–873 K) with batch and flow methods and ultra-
high vacuum (UHV) techniques. N2 and N2O were found as
the reaction products, with a typical increase in the N2/N2O
ratio with increasing temperature and/or increasing concen-
tration of NH3 in the feed [6].
and N2. These investigations concluded that N2O formation
depends mostly on the relative coverage of NO and explicitly
excluded the pathway involving the reaction between NO
and NH2 species (Eq. (6)).
Gland and Korchak [5] investigated the NH3–NO inter-
actions over a clean stepped platinum single crystal at pres-
sures of 10−6 Pa and proposed the conversion of HNO to N2
according to Eqs. (10)–(11) instead of Eq. (8), since N2O
was not observed under their UHV conditions. By combin-
ing the mechanisms in [5,11] and excluding Eqs. (6) and (9),
Takoudis and Schmidt [6] described the global kinetics of
Eqs. (1)–(2) with a single Langmuir–Hinshelwood model,
which was said to be valid for UHV and high-pressure con-
ditions.
HNO∗ + H∗ → N∗ + H2O + ∗,
(10)
(11)
2N∗ → N2 + 2∗.
It should be pointed out that the absence of N2O in [5] has
frequently been experienced in other UHV studies dealing
with the NH3–NO reaction over Pt(100) in the temperature
range of 300–700 K [3,16]. As reasoned in [1], the low peak
pressure applied in high and ultrahigh vacuum experiments
may represent an important limitation for N2O analysis and
thus for gaining mechanistic insights into the formation of
this essential by-product in NH3 oxidation or the reduction
of NO by NH3.
Based on the above analysis, it can be concluded that the
reduction of NO by NH3 has not been studied over noble
metal gauzes at relevant temperatures of industrial ammonia
burners. In this work we have investigated the mechanism of
this reaction over Pt and Pt–Rh gauzes in the TAP reactor us-
ing isotopic tracers. The intrinsic features of the TAP reactor
make it possible to study the highly exothermic processes
in Eqs. (1)–(2) under isothermal conditions, as previously
demonstrated for ammonia oxidation [1]. It is shown that
secondary NH3–NO interactions chiefly contribute to the
loss reactions in ammonia burners, yielding N2O and N2.
Reaction pathways leading to these undesired products have
been elucidated.
6NO + 2NH3 → 5N2 + 6H2O
ꢀH o = −452 kJ molN−H1
(1)
3
8NO + 2NH3 → 5N2O + 3H2O
ꢀH o = −473 kJ molN−H1
(2)
3
Otto et al. [11] studied the catalytic reduction of nitric ox-
ide with ammonia by circulating mixtures of 15NH3, 14NO,
and Ar in a gradientless batch reactor over Pt/Al2O3 at
pressures of 0.025–40 kPa and temperatures of 473–523 K.
The proposed reaction mechanism is given in Eqs. (3)–(9),
where ∗ denotes an active site. It was basically concluded
that N2 is predominantly formed via interaction of adsorbed
NO and the ammonia intermediate NH2 (Eq. (5)). N2O
formation involved the interaction of two NO molecules
with chemisorbed hydrogen (obtained from the dissociative
chemisorption of NH3, Eq. (3)), leading to nitroxyl species
(HNO) as a reaction intermediate (Eqs. (7)–(8)). Andrus-
sow claimed that this intermediate plays a central role in
the mechanism of ammonia oxidation [15], although this has
been never observed experimentally. Eqs. (6) and (9) were
considered to be minor paths for N2O and N2 formation, re-
spectively.
NH3 + 2∗ ↔ NH2∗ + H∗,
NO + ∗ ↔ NO∗,
(3)
(4)
(5)
(6)
(7)
(8)
(9)
2. Experimental
The commercial gauzes used in this study, knitted pure
Pt and woven Pt–Rh (95–5 wt%) alloy, and the experimental
procedures during TAP experiments were described in [1].
Briefly, a single piece of gauze catalyst (ca. 30 mm2) was
placed between two layers of quartz particles (sieve fraction
250–350 µm) in the isothermal zone of the TAP microreactor
(6 mm i.d.). The amount of Pt and Pt–Rh gauzes used was
110 and 25 mg, respectively. Prior to the experiments, the
samples were pretreated in pure O2 at 1273 K and ambient
pressure for 2 h. Then the catalysts were exposed to vacuum
conditions (10−5 Pa) and transient studies of the NH3–NO
reaction were performed by means of single and sequential
pulse experiments.
NO∗ + NH2∗ → N2 + H2O + 2∗,
NO∗ + NH2∗ → N2O + 2H∗,
H∗ + NO∗ → HNO∗ + ∗,
2HNO∗ → N2O + H2O + 2∗,
2HNO∗ + 2H∗ → N2 + 2H2O + 4∗.
More recent studies of the NO–NH3 reaction over poly-
crystalline Pt foils up to 623 K by Katona et al. [9,10] have
supported the reaction scheme proposed by Otto et al. [11]
and the relative importance of the steps leading to N2O

J. Pérez-Ramírez et al. / Journal of Catalysis 229 (2005) 303–313
305
Single pulsing was carried out at 1073 K with mixtures
of 14NH3:14NO:Xe = 1:1:1, 15NH3:14NO:Xe = 1:1:1, and
15NO:Ne = 1:1. In sequential pulse (pump-probe) experi-
ments, mixtures of 14NO:Ne = 1:1 and 14NH3:Xe = 1:1 (or
15NH3:Xe = 1:1) were independently introduced into the
reactor via two high-speed pulse valves, with a time delay
(ꢀt) of 0.1–2 s between the two pulses at 1023 or 1073 K.
The pulse size was varied from ca. 1014 to 1016 molecules
to study the reaction in Knudsen and molecular diffusion
regimes, respectively.
In the experiments, Ne (4.5), Xe (4.0), O2 (4.5), 14NO
(2.5), 14NH3 (2.5), 15NH3 (99.9% atoms of 15N), and 15NO
(99.5% atoms of 15N) were used without additional pu-
rification. Isotopically labeled ammonia and nitric oxide
were purchased from ISOTEC. Transient responses were
monitored at atomic mass units (AMUs) related to reac-
tants, reaction products, and inert gases at the reactor out-
let with a quadruple mass spectrometer (Hiden Analyti-
cal). The following AMUs were analyzed: 132 (Xe), 46
(14NO2, 15N15NO), 45 (15N14NO), 44 (14N14NO, CO2), 32
(O2), 31 (15NO, H14NO), 30 (14N14NO, 14NO, 15N15N),
29 (15N14N), 28 ( 14N14NO, 14N14N), 20 (Ne), 18 (H2O,
15NH3), 17 (14NH3, 15NH3, H2O, OH), 16 (O2, H2O,
14NH3, 15NH3), 15 (14NH3, 15NO, 15NH3, 15N15N), and
2 (H2). For each AMU, pulses were repeated 10 times and
averaged to improve the signal-to-noise ratio. The concen-
tration of feed components and reaction products was deter-
mined from the respective AMUs with the use of standard
fragmentation patterns and sensitivity factors. Because of
overlap of the various masses of 15NH3 with other species
in the MS analysis, the degree of 15NH3 conversion was cal-
culated from the N-balance.
14
14
Fig. 1. Transient responses during sequential pulsing of NO and NH
3
14
14
over Pt gauze (ꢀt = 0.2 s) and NH and NO over Pt–Rh gauze
3
14
14
14
(ꢀt = 0.5 s) at 1073 K. Pulse size of NO and NH ∼ 10 molecules.
3
3. Results
3.1. Sequential pulsing of 14NO and 14NH3 in the Knudsen
diffusion regime
cules in the ammonia pulse and (ii) reduction of two NO
molecules in the nitric oxide pulse. H2 was detected in the
ammonia pulse only.
Fig. 2 shows the sharp decrease in the concentration of
gas-phase H2 at the time of the NO pulse during sequential
pulsing of NH3 and NO at ꢀt = 0.2 (Pt gauze) and 0.5 s (Pt–
Rh gauze). This feature indicates that hydrogen-containing
species adsorbed to the gauze surface, from which H2 is
formed, effectively catalyze the reduction of NO to N2 and
H2O. The reduction of NO by H2 at low temperatures (300–
700 K) over Pt-supported catalysts in the presence of O2 was
originally reported by Jones et al. [17]. As clearly shown in
Fig. 3, the H2 yield in the NH3 pulse drops to zero with de-
creasing time delay between the NH3 and NO pulses from
ꢀt = 0.5 s to 0 s. The latter case, when ammonia and ni-
tric oxide were simultaneously pulsed, corresponds with the
highest degree of NO conversion (42%, Fig. 3 and Table 1).
These results indicate the high activity of Pt and Pt–Rh
gauzes for the NO reduction by NH3 to N2 under the vacuum
conditions in the TAP reactor. This is in agreement with pre-
Fig. 1 shows representative normalized transient re-
sponses during sequential pulsing of NH3 and NO over
Pt–Rh gauze and of NO and NH3 over Pt gauze at 1073 K
in the Knudsen diffusion regime (ca. 1014 molecules per
pulse). These experiments make it possible to analyze the
effect of adsorbed species formed during the first pulse on
the primary heterogeneous steps of activation of gas-phase
molecules in the second pulse. In this case, normalization
of the responses is convenient for a proper comparison of
pulse shapes. The results show clear mechanistic analogies
between Pt and Pt–Rh gauzes in NO–NH3 interactions. N2,
H2, and H2O were the only reaction products over the noble
metal gauzes and were not influenced by the order in which
the reactant gases were pulsed. Formation of N2 and H2O
was observed in both the ammonia and nitric oxide pulses.
This suggests two independent reaction pathways leading to
these products, involving (i) oxidation of two NH3 mole-

306
J. Pérez-Ramírez et al. / Journal of Catalysis 229 (2005) 303–313
Table 1
Performance of noble metal gauzes and SiO particles in NO reduction by NH at 1073 K
2
3
14
a
15 14
N
15
15 14
b
14 14
Sample
Diffusion regime
X( NO) (%)
Y(
NO) (%)
Y( NO) (%)
Y(
N
N) (%)
Y(
N
N) (%)
Y(H ) (%)
2
c
c
Pt gauze
Knudsen
Molecular
57
65
n.d.
0.3
4
4
15
20
42
48
n.d.
n.d.
Pt–Rh gauze
Knudsen
Molecular
42
48
n.d.
0.2
6
6
10
13
31
35
n.d.
n.d.
SiO spheres
Molecular
2
n.d.
n.d.
14
1
1.5
n.d.
2
15
14
15
16
Conditions: single pulsing of a mixture NH : NO:Xe = 1:1:1, NH pulse size ∼ 10 molecules (Knudsen diffusion regime) and ∼ 10 molecules
3
3
(molecular diffusion regime).
a
14
15 14
14 14
N
X( NO) = Y (
N
N) + Y (
N).
b
15 14
15 14
o
15 14
N N)/N ( NH ), where N is the number of reactant molecules in the feed and N the number of product
3
o 15
o
Y (
N
N) = N(
N
N)/N (NO) = N(
molecules at the reactor outlet.
c
Not detected.
Fig. 3. NO conversion and H yield vs time delay during sequential pulsing
2
14
14
14
of NH and NO over Pt–Rh gauze at 1073 K. Pulse sizes of NO and
3
14
14
NH ∼ 10 molecules.
3
This is further assessed in Sections 3.2 and 3.3 by means
of single and sequential pulse experiments with 15NH3 and
14NO in the molecular diffusion regime.
3.2. Single pulsing of 14NO and 15NH3: Knudsen vs
molecular diffusion regimes
Fig. 2. Participation of hydrogen-containing species in NO reduction dur-
14
14
The influence of the diffusion regime and peak pressure
in the TAP reactor on the reaction of ammonia with nitric
oxide over the Pt and Pt–Rh gauzes at 1073 K has been eval-
uated. To this end, an equimolar mixture of 15NH3 and 14NO
was pulsed in Knudsen diffusion (pulse size ∼ 1014 mole-
cules, peak pressure ∼ 5 Pa) and molecular diffusion (pulse
size ∼ 1016 molecules, peak pressure ∼ 200 Pa) regimes.
The transient responses of the reactants and N-containing
products are displayed in Fig. 4. Table 1 quantifies the cor-
responding catalytic performance. As shown in the figure,
the mechanism of the high-temperature reduction of NO by
NH3 appears to be similar in general terms over the knitted
Pt and woven Pt–Rh alloy gauzes, as concluded in our pre-
vious study dedicated to primary O2–NH3 interactions [1].
The higher NO conversion over the knitted Pt gauze in Ta-
ble 1 is due to the higher amount of Pt gauze used in the
ing sequential pulsing of NH and NO over Pt gauze (ꢀt = 0.2 s)
3
14
and Pt–Rh gauze (ꢀt = 0.5 s) at 1073 K. Pulse sizes of NO and
14
14
NH ∼ 10 molecules.
3
vious ultrahigh vacuum [3–5], vacuum [6–9], and ambient
pressure [11] studies over different Pt specimens (see Intro-
duction).
Sequential pulse experiments in Fig. 1 also show that
the concentration of gas-phase NO strongly decreases at the
time of the NH3 pulse over the Pt gauze. This indirectly
indicates the reversible dissociation of NO into N and O
atoms on the reduced catalyst surface. The resulting oxygen
species are essential for NH3 conversion into N2, H2O, and
H2. No conclusion can be drawn about N2O formation from
the experiments in Fig. 1, where nonlabeled molecules and a
low peak pressure (Knudsen diffusion regime) were applied.

J. Pérez-Ramírez et al. / Journal of Catalysis 229 (2005) 303–313
307
15
14
Fig. 4. Transient responses during single pulsing of a NH : NO:Xe = 1:1:1 mixture over Pt and Pt–Rh gauzes at 1073 K under (a) Knudsen diffusion
3
15
14
15
16
regime (pulse size of NH ∼ 10 molecules) and (b) molecular diffusion regime (pulse size of NH ∼ 10 molecules). The order of the profiles and the
3
3
associated labels shown in (a) for Pt gauze are the same for all the figures.
experiments (4 times as much) compared with the woven
Pt–Rh gauze (see Section 2).
reaction product of the NO reduction by NH3 below 1073 K
in high-pressure studies [6,7]. As a matter of fact, formation
of nitrous oxide as 15N14NO can be clearly identified when
the pulse experiments were performed in the molecular dif-
fusion regime, that is, with higher peak pressures (Fig. 4b).
The relatively low intensity of the N2O signal in Fig. 4 indi-
cates, once more, that concentrations of this by-product are
very low compared with N2.
The formation of 14N14NO from two 14NO molecules (as
suggested by Otto et al. [11] in Eqs. (7) and (8) at 523 K
cannot be decoupled from CO2 in mass spectrometry (both
have an AMU of 44). It is well known that the surface of
commercial PGM gauzes is covered by substantial amounts
of carbon [18], and part of the carbon remains even after pre-
treatment in O2 flow at 1273 K. As a consequence, a small
but significant amount of CO2 originates from the oxidation
of carbon by O2 or, in this case, NO. This aspect was ex-
tensively discussed in [1] and indeed motivated us to apply
isotopic tracers to elegantly discriminate between N2O and
CO2 in MS analysis.
Regardless of the diffusion regime, labeled nitric oxide
(15NO) and molecular nitrogen were identified as reaction
products. Two nitrogen isotopes can be discerned: 15N14N
and 14N14N. Identification of the other possible nitrogen iso-
tope, 15N15N, in this specific experiment is complicated,
since it overlaps with the reactant 14NO in MS analysis
(AMU 30). To this end, the sequential pulse experiments
in Section 3.3 are required. The diffusion regime has only
a slight influence on the NH3–NO interactions, indicating a
negligible contribution of gas-phase processes (excluded un-
der Knudsen diffusion) to the overall performance. In fact,
the conversion of 14NO over quartz particles during TAP ex-
periments in the molecular diffusion regime was only 2% at
1073 K (see Table 1).
As indicated in Section 3.1, Fig. 4a shows that N2O for-
mation was not observed at the low peak pressures charac-
teristic of the Knudsen diffusion regime. This agrees with
the reported absence of N2O in previous investigations of
the NH3–NO reaction under UHV conditions [3–5,16]. All
of these studies were performed at T < 873 K, and hence a
direct comparison with the present results cannot be estab-
lished. N2O has typically been observed in low amounts as a
As expected, single pulsing of a 15NH3–14NO mixture in
the TAP microreactor filled with quartz particles (i.e., with-
out gauze) did not induce N2O formation in the temperature
range investigated. Accordingly, it can safely be concluded

308
J. Pérez-Ramírez et al. / Journal of Catalysis 229 (2005) 303–313
15 14
15 14
N
15 15
Fig. 6. Transient responses of
N NO formed during sequential pulsing
Fig. 5. Transient responses of
15
NO and
N NO formed during se-
15
14
14
of NH and NO over Pt–Rh gauze at 1023 K at different time delays
(ꢀt, in seconds). Pulse sizes as in caption of Fig. 5.
3
quential pulsing of NH and NO over Pt–Rh gauze at 1023 K. Pulse
sizes of NH and NO ∼ 2.2 × 10 and 3.0 × 10 molecules, respec-
3
15
14
16
16
3
tively.
that the N2O observed in the experiments at higher peak
pressures originates from the heterogeneous reaction of NH3
and NO over the noble metal gauzes. The absence of nitrous
oxide in ammonia-nitric oxide interactions in the Knudsen
regime can be attributed to the detection limit of analysis
in relation to the low amount of N2O formed. This also ex-
plains the unsuccessful identification of this by-product in
UHV studies of NH3 oxidation and NO reduction by NH3
over platinum surfaces, since typical pressures in UHV ex-
periments (10−6–10−7 Pa) are several orders of magnitude
lower by a factor of 106 than the peak pressure in the TAP
reactor under the Knudsen diffusion region.
Fig. 7. Transient responses of N -isotopes formed during sequential pulsing
2
3.3. Sequential pulsing of 14NO and 15NH3 in the
molecular diffusion regime
15
14
of NH and NO (ꢀt = 0.5 s) over Pt–Rh gauze at 1023 K. Pulse sizes
3
as in caption of Fig. 5.
iments on 15NO over PGM gauzes at 1023 K, yielding only
15N15N as the reaction product (not shown). Our previous
paper [1] also reported that single pulsing of 14NO over re-
duced PGM gauzes yielded only 14N14N, and the N2/O2
ratio at the reactor outlet was > 2, indicating oxidation of
the catalyst surface by NO.
To gain additional insights into the mechanism of N2O
and N2 formation during the reduction of nitric oxide by
ammonia, 15NH3 and 14NO were sequentially pulsed over
Pt–Rh gauze at different time delays in the range of ꢀt =
0.1–2 s. As shown in Fig. 5, N2O was observed at the
time of the NO pulse. The main N2O isotope produced
was 15N14NO, which clearly originates from the reaction
between 15NH3 and 14NO. 15N15NO is formed in much
smaller amounts ( 14N15NO/15N15NO ∼ 8). The latter iso-
tope is not a primary reaction product, that is, it is not formed
by direct 15NH3 oxidation. Rather it stems from the reac-
tion between 15NH3 in the first pulse and 15NO formed in
the pulse of 14NO. As shown in Section 3.1 and in previous
single-pulse experiments [1], NO can oxidize the reduced
noble metal gauze, and the deposited oxygen can eventually
react with a surface 15NHx intermediate, leading to 15NO.
Again, N2O formation from 14NO would yield 14N14NO,
but this mass cannot be decoupled from that of CO2 in mass
spectrometry. However, the irrelevance of this route for N2O
formation was indirectly concluded from single-pulse exper-
The heterogeneous nature of the reaction steps leading
to N2O can be conclusively demonstrated by comparing the
formation of the main isotope 15N14NO at different time
delays between the 15NH3 and 14NO pulses. Fig. 6 nicely
shows the decreased formation of nitrous oxide (always at
the time of the NO pulse) with increasing time delay be-
tween ammonia and nitric oxide pulses. Particularly signif-
icant is the decrease observed between ꢀt = 0.1 and 1 s.
In view of this result, formation of nitrous oxide is highly
dependent on the surface coverage by reactive NHx interme-
diates, which couple with NO, yielding N2O.
Sequential pulse experiments with 15NH3 and 14NO have
revealed the formation of three nitrogen isotopes: 15N15N,
15N14N, and 14N14N. As shown in Fig. 7 for the experiment

J. Pérez-Ramírez et al. / Journal of Catalysis 229 (2005) 303–313
309
delay can also be related to the ratio of 15NH3/14NO in
the NO pulse, which is 1 at ꢀt = 0 s and 0 at ꢀt = 2 s.
This can be qualitatively associated with the NH3/NO ra-
tio at different locations of the gauze pack in the ammonia
burner (see Section 4.3). The absence of 15N15N at ꢀt = 0 s
indicates that N2 formation from two 15NH3 molecules is
inhibited in the presence of 14NO. In contrast, the high rel-
ative yield of 14N14N and 15N14N at ꢀt = 0 s is related to
the high activity of the PGM gauzes for 14NO reduction by
H and 15NHx species, respectively. The decreased yield of
these two isotopes with an increased time delay is directly
related to the decreased coverage of intermediate surface
species resulting from ammonia activation, which are essen-
tial for NO reduction to N2. Fig. 9 shows that the ratio of
15N14N/14N14N in the product gas decreases with an in-
creased time delay between the 15NH3 and 14NO pulses.
This indicates that the coupling of NO and NH3 leading to
mixed nitrogen isotope prevails over direct NO reduction
with an increased NH3/NO ratio. The yield of 15N15N in
Fig. 8, which is formed in the 15NH3 pulse (see Fig. 7),
is significant at ꢀt ꢀ 0.5 s and stems from the oxidation
of NH3 by O-species. Such oxygen species may originate
from the dissociation of nitric oxide in the 14NO pulse or
the high-temperature pretreatment of the gauzes in O2 (dis-
cussed later).
Fig. 8. Distribution of isotopically labelled N-atoms in N molecules vs
2
15
14
time delay during sequential pulsing of NH and NO over Pt–Rh gauze
3
at 1023 K. Pulse size as in caption of Fig. 5.
4. Discussion
4.1. Importance of the diffusion regime and the use of
isotopes
Two combined experimental aspects are considered to be
essential for deriving relevant factors that direct selectivity
in the NH3–NO reaction for N2O and N2 in the TAP reac-
tor: (i) the application of high pressures and (ii) isotopically
labeled molecules. Experiments in the Knudsen diffusion
regime (peak pressure of ca. 5 Pa at 1073 K) gave insights
into N2 formation, although N2O was not observed as a re-
action product. This result parallels those of previous surface
science studies carried out at temperatures up to 823 K with
molecular beams in ultrahigh vacuum, where N2O was not
observed in the reaction of NH3 with O2 [19,20] or with
NO [3–5,16]. In contrast, studies of the NH3–NO reaction
over platinum-based catalysts at total pressures in the range
of 0.1–1000 Pa and temperatures up to 823 K observed
N2O as a product along with N2 [6–9,11,21]. In fact, high-
temperature ammonia oxidation over PGM gauzes within ni-
tric acid manufacture is the largest source of nitrous oxide in
the chemical industry (400 kilotons of N2O per year), even
though the N2O selectivity (1–2%) is significantly lower
compared with the N2 selectivity (up to 4%) [22].
15 14
14 14
Fig. 9. Ratio of
N N/ N N vs time delay during sequential pulsing
15
14
of NH and NO over Pt–Rh gauze at 1023 K. Pulse sizes as in caption
3
of Fig. 5.
at ꢀt = 0.5 s and 1023 K, the simple isotopes 15N15N and
14N14N were formed at the time of the 15NH3 and 14NO
pulses, respectively. However, the mixed 15N14N isotope is
formed in both pulses. It should be mentioned here that the
small amount of 14N14N at the time of the ammonia pulse is
due to the presence of impurities in the 15NH3 source.
The distribution of nitrogen isotopes as a function of the
time delay is shown in Fig. 8. The yield of the mixed ni-
trogen isotope was determined by the sum of the 15N14N
yield in each pulse. The 15N14N yield at ꢀt = 0 s (simulta-
neous pulsing of 15NH3 and 14NO) is relatively high (25%)
and strongly decreasing at longer time delays. The decrease
in 14N14N yield with time delay follows a trend similar to
that of 15N14N, whereas an opposite behavior is observed
for 15N15N. The dependences of the distributions of N2 iso-
topes presented in Fig. 8 can be associated with the relative
coverage of N-species on the catalyst surface. This coverage
can be tuned by the application of different time delays be-
tween the 15NH3 and 14NO pulses. Furthermore, the time
The “invisibility” of N2O in (ultra)low-pressure studies
can be related to detection constraints on the analytical meth-
ods as a consequence of the intrinsically low concentrations
of N2O formed. The low pressure is apparently suitable for

310
J. Pérez-Ramírez et al. / Journal of Catalysis 229 (2005) 303–313
a proper analysis of N2, which is formed in a considerably
higher proportion. It should be stressed that the pressure in
UHV studies is up to 9 orders of magnitude lower than that
in TAP experiments in the Knudsen diffusion regime. TAP
experiments with the molecular diffusion regime (peak pres-
sure of ca. 200 Pa at 1073 K, i.e., two orders of magnitude
higher than in the Knudsen diffusion regime, ca. 5 Pa) have
conclusively determined that N2O is formed in the NH3–NO
reaction, and it was observed in the previously investigated
NH3–O2 reaction [1]. Any homogeneous process leading to
N2O can be excluded based on the absence of N2O during
blank experiments in the TAP reactor without the gauzes,
with NH3–NO mixtures and large pulse sizes (Table 1).
The second essential aspect to unraveling the mechanism
of secondary NH3–NO interactions yielding N2O and N2
consists is the use of isotopes in the TAP reactor. This is vital
for two reasons: (i) to discriminate between N2O and CO2 in
MS analysis in mass spectrometry and (ii) to trace the origin
of reaction products by the distinction of the N-label in the
reactants (15NH3 and 14NO). As revealed by the use of iso-
topes, similar amounts of N2O and CO2 were formed in our
TAP experiments during pulsing of NH3–O2 mixtures [1].
investigations concluded that the largest contribution to the
total nitrogen production arose from the mixed 14N15N iso-
tope, suggesting the importance of the interaction between
adsorbed 15NO and 14NH3 species. In a related work, van
Tol et al. [21] investigated the oscillatory behavior of the
NO–NH3 reaction over Pt(100) at low pressures (1 × 10−6
to 4 × 10−5 mbar) at 420-485 K with 15NH3 and 14NO. In
this case, 14N14N and 15N14N were formed in comparable
amounts, whereas the concentration of 15N15N was reported
to be very low. This agrees well with our findings that N2 for-
mation from two NH3 molecules does not occur when NH3
and NO are simultaneously pulsed over the PGM gauzes (see
Fig. 8).
The relevance of the various pathways for N2 formation
depends on the relative coverage of NO, NHx, and H species
on the catalyst surface. It is important to discuss the distribu-
tion of N-labels in N2 as a function of the time delay between
15NH3 and 14NO pulses described in Section 3.3 (see Fig. 8).
These experiments are extremely valuable to understanding
the influence of the NH3/NO ratio on the above reaction
pathways. A short time delay (ꢀt → 0 s) between NH3 and
NO pulses is associated with a relatively high NH3/NO ratio
in the NO pulse and accordingly with a high surface cov-
erage of NHx species on the catalyst surface. In contrast,
a long time delay time indicates a low NH3/NO ratio in the
NO pulse and thus a low coverage of NHx species in the
NO pulse. As shown in Fig. 7, 15N15N was the main nitro-
gen isotope formed (in the 15NH3 pulse) during sequential
15NH3–14NO pulsing at ꢀt = 2 s. This is explained by the
direct oxidation of ammonia to nitrogen by surface oxygen
species (Eq. (12)). A decrease in the time delay between the
15NH3 and 14NO pulses leads to a decreased concentration
of 15N15N at the expense of 14N15N and 14N14N formation
(Fig. 8). The increased formation of 14N14N and 14N15N is
in good agreement with steady-state tests of NO reduction
by NH3 over polycrystalline platinum wire reported by Tak-
oudis and Schmidt [6]. These authors found that the rate
of NO decomposition at 1000 K increased with the molar
feed NH3/NO ratio. However, since no isotopically labeled
molecules were used in that study, no conclusion about the
reaction pathways of nitrogen formation could be drawn.
From a chemical point of view, the pathways for N2 for-
mation with NO participation identified in this study differ
in that Eq. (13) involves the coupling of nitrogen atoms in
adsorbed NO and NHx species, whereas Eqs. (14) and (15)
involve the reduction of two NO molecules by adsorbed NHx
or H species, respectively. The participation of H-containing
species in NO reduction was unequivocally concluded from
the results given in Fig. 2. These routes have different de-
pendences on the NH3/NO ratio, as previously reported in
studies conducted over Pt(100) at 425 K and 450 K [3]. The
decreased molar 14N15N/14N14N ratio with increasing time
delay between the 15NH3 and 14NO pulses in Fig. 9 indicates
that Eqs. (14) and (15) are the major route for N2 at a low
NH3/NO ratio, whereas Eq. (13) becomes more important
with an increasing NH3/NO ratio.
4.2. N2 formation
It has to be emphasized that the TAP technique, despite its
high time resolution, does not provide mechanistic informa-
tion on the level of elementary reaction steps. Our approach
in the previous section was to provide a few relatively ele-
mentary steps from which a basic perception of the reaction
mechanism could be obtained. Such an approach would, of
course, be possible, but it is considered to be merely specu-
lative.
Sequential pulse experiments with isotopes have revealed
that three N2 isotopes, 15N15N, 14N14N, and 15N14N, were
formed in the reaction of 15NH3 with 14NO over the Pt
and Pt–Rh gauzes. Based on our sequential 15NH3–14NO
pulse experiments in Figs. 7 and 8, the reaction pathways in
Eqs. (12)–(15) leading to the various N2 isotopes can be es-
tablished. In these reaction steps, “s” denotes an active site,
and “s–ꢁ” represents an adsorbed species:
2s–15NHx + 2s–O → 15N15N + 2s–HxO + 2s,
(12)
(13)
s–15NHx + s–14NO → 15N14N + s–HxO + s,
s–15NHx + 2s–14NO → 14N14N + s–HxO
+ s–15N + s–O,
(14)
(15)
2s–H + 2s–14NO → 14N14N + 2s–OH + 2s.
N2 formation involves not only the conversion of two NO
or two NH3 molecules, leading to the simple isotopes, but
also the coupling of nitrogen atoms in the reactants, yield-
ing the mixed isotope. These three species were observed in
previous isotopic UHV studies (10−6 mbar) conducted over
Pt(100) at 300–750 K [3,16] with isotopically labeled ni-
tric oxide (15NO) and nonlabeled ammonia ( 14NH3). These

J. Pérez-Ramírez et al. / Journal of Catalysis 229 (2005) 303–313
311
4.3. N2O formation
As pointed out in the Introduction, low-temperature
NH3–NO studies conducted over Pt/Al2O3 found the reac-
tion between adsorbed NH2 and NO species (Eq. (6)) to play
too small a role to explain N2O formation [11]. This step was
discarded because of the absence of H2 in the product gases,
which would have been expected from the spontaneous re-
combination of reactive adsorbed H atoms. Instead, the main
reaction pathway of N2O formation was suggested to in-
volve the reaction of two NO molecules with chemisorbed
hydrogen, via formation of nitroxyl (HNO) as an intermedi-
ate species (Eqs. (7) and (8)). Our results were obtained at
significantly higher temperature and over PGM gauzes and
conclusively show that N2O formation requires the coupling
of N-atoms of adsorbed 15NH3 and 14NO species. Eq. (6),
originally suggested by Otto et al. [11], cannot be considered
the reaction route for N2O formation, since it implies the for-
mation of two H atoms. These atoms can recombine to yield
gas-phase H2, which was not observed in our experiments
when a mixture of NO and NH3 was introduced as a pulse
into the reactor (water was the only hydrogen-containing
product); this was when the highest N2O production was
found. Taking into account this experimental fact and the
importance of oxygen species in the NH3–NO reaction (Sec-
tion 4.4), Eq. (16) can be suggested as the exclusive route for
N2O formation
Fig. 10. Reaction network during the high-temperature NH oxidation over
3
noble metal gauzes, as derived from TAP studies here and in [1]. The dashed
area includes steps at high oxygen coverages, i.e., more representative of
industrial ammonia oxidation. Dashed arrows indicate minor reaction path-
ways and crosses on certain arrows denote that the corresponding process is
inhibited in the presence of oxygen. Two types of adsorbed oxygen species
are distinguished (O and O*, see text).
representing the reaction network and product formation as
a function of oxygen coverage (θO) is considered in Fig. 10.
An accurate value for the coverage in the “low” and “high”
cases cannot be provided at this stage, but extreme values
approaching 0 and 100%, respectively, can be considered
for the purpose of this paper. No reaction products were
obtained upon interaction of NH3 with the reduced PGM
surface (θO = 0). As concluded in previous studies [1,26],
the NH3–O2 reaction is initiated by the activation of the
ammonia molecule by adsorbed oxygen species. These oxy-
gen species induce the consecutive stripping of hydrogen
atoms from ammonia according to Eq. (17), yielding reactive
NH3 − y intermediates (referred to as NHx here). Recom-
bination of two NHx fragments is favored at low oxygen
coverage, leading to N2 (Eq. (18)), with formation of in-
significant amounts of NO [1]. The above-mentioned forma-
tion of N2 via direct NH3 oxidation (see 15N15N in Figs. 7
and 8) likely originates from the combination of Eqs. (17)
and (18). Together with the oxygen coverage, the nature of
the adsorbed oxygen species appears to be of vital impor-
tance in the NO:N2 distribution [1]. To this end, O and O*
are distinguished in Fig. 10, representing strongly bounded
and highly reactive adsorbed oxygen species, respectively.
The pathway in Eq. (18) is virtually suppressed under con-
ditions of industrial relevance because of the high oxygen
coverage. This is due to the excess oxygen in the feed (mo-
lar NH3/O2 ratio ∼ 1/2) and the high affinity of O2 for
dissociative adsorption on a reduced noble metal surface.
Formation of N2 from direct NH3 decomposition would oc-
s–15NHx + s–14NO + s–O → 15N14NO + s–HxO + 2s.
(16)
This reaction resembles the pathway for 15N14N forma-
tion in Eq. (13) and suggests increased N2O formation (rel-
ative to N2 formation) with increased oxygen coverage on
the catalyst surface. Accordingly, N2O would be expected to
be the main by-product of high-temperature ammonia oxi-
dation, which is carried out in the presence of excess oxygen
(typically 10 vol% NH3 in air). However, this is not in line
with results obtained in our study (Figs. 4–6) and reported in
the literature [22–24]. As a matter of fact, N2 is the main N-
containing by-product in industrial NH3 burners (typically
N2/N2O ∼ 2). The relatively high N2/N2O ratio can be ex-
plained we take into account (i) direct N2O decomposition
into N2 and O2, which can occur over the oxidized Pt–Rh
gauze at high temperatures [25], or (ii) reduction of N2O
by NH3 into N2 and H2O. According to [8], the reaction
between NH3 and N2O has to be considered during NO re-
duction with NH3 over polycrystalline platinum foil at high
temperature.
4.4. Reaction network and loss reactions in the ammonia
burner
Based on the results reported here and in our previous pa-
per [1], an improved description of the complex processes
occurring during the high-temperature oxidation of ammo-
nia over PGM gauzes has been attained. A simplified scheme

312
J. Pérez-Ramírez et al. / Journal of Catalysis 229 (2005) 303–313
cur only in the presence of a high local NH3 concentration,
giving rise to a low O/NHx ratio at the surface. N2 produc-
tion by this route is expected to be minimal in view of the
presence of distributor plates before the gauze pack, which
improve the mixing of the ammonia-oxygen mixture and the
excess oxygen in the feed gas. As concluded for ammonia
oxidation over Pt(100) single crystal at temperatures up to
800 K [27] and in our earlier study [1], a high oxygen cov-
erage increases the selectivity for NO. In this situation, the
extraction of H atoms from NH3 and the recombination of
nitrogen and oxygen atoms (Eq. (19)) are accelerated, lead-
ing to the most oxidized N-product:
pretreatment of the as-received gauzes (see Experimental)
and/or (ii) the dissociation of NO over reduced metal sites
(Eq. (20)). These reduced sites are generated via Eqs. (17)–
(19). We have observed a high activity of the H2-pretreated
Pt and Pt–Rh gauzes toward NO conversion to N2. In con-
trast, O2-pretreated Pt and Pt–Rh gauzes are unable to disso-
ciate NO [1]. Therefore, based on this aspect and on the high
O2 content in the burner, direct NO decomposition is not
considered an important loss reaction over the noble metal
gauzes
NO + 2s → s–N + s–O.
(20)
In summary, oxygen species catalyze both the oxidation
of NH3 to NO and the reduction of NO with NH3 to N2O
and N2. It is envisaged that the desired reaction is favored
over the undesired one at increasing oxygen coverages. In
practice, such a situation can be accomplished with a larger
excess of oxygen in the feed. In fact, nearly 100% NO selec-
tivity was reported for ammonia oxidation with an O2/NH3
ratio ca. 3 [29]. However, this operation would dramatically
reduce the lifetime of a gauze pack, since an increase in
partial O2 pressure accelerates platinum loss in the form of
volatile PtO2. Further investigations of the influence of O2
on the kinetics of the NH3–NO reaction are required. To this
end, TAP studies with the ternary NH3–O2–NO system will
be undertaken in the near future.
s–NH3 + ys–O → s–NH3 − y + ys–OH,
s–NHx + s–NHx → N2 + xH2 + 2s,
s–N + s–O → NO + 2s.
(17)
(18)
(19)
The mechanism of ammonia oxidation is complicated by
the occurrence of interactions between NH3 and NO. The re-
sults presented here indicate that by-products (N2 and N2O)
from industrial ammonia burners containing PGM gauze
packs are mainly formed by the reduction of NO by NH3.
These secondary reactions are particularly detrimental to the
economy of the process, since they involve a loss of reac-
tant and the desired product. In industrial burners, the no-
ble metal gauzes are packed snugly against each other in
numbers varying from 4 to 50, depending on the operating
pressure; the low-pressure reactor uses fewer gauzes [24].
Obviously, the origin of secondary NH3–NO interactions
should be situated in a region of the catalytic burner where
NO and NH3 coexist, that is, in the very first gauzes of the
pack. N2O formation principally stems from the coupling of
adsorbed NO and N atoms resulting from NH3 (Eq. (16)).
However, three different reaction pathways have been iden-
tified for N2 formation, involving two NO molecules, two
NH3 molecules, and one NH3 molecule and one NO mole-
cule (Eqs. (12)–(15)). The relative importance of these re-
actions is a function of the NH3/NO ratio, which decreases
along the axial coordinate in the gauze pack. Furthermore,
nitrogen can be formed from nitrous oxide by the reduction
of N2O by NH3 or direct N2O decomposition. The extent of
the NH3–N2O reaction cannot be assessed from the present
study. Under industrial reaction conditions, direct N2O de-
composition over PGM gauzes is not expected, in view of
the large excess of oxygen in the burner (O2/N2O ∼ 70),
which severely inhibits the reaction [25].
5. Conclusions
The results presented in this paper indicate the impor-
tance of secondary NH3–NO interactions for the formation
of N2O and N2 by high-temperatureammonia oxidation over
PGM gauzes as part of the manufacture of nitric acid. Tran-
sient isotopic studies in the temporal analysis of products
(TAP) reactor in the molecular diffusion regime with iso-
topes make it possible to derive valuable information on
reaction pathways leading to these undesired by-products.
Adsorbed oxygen species catalyze ammonia activation into
reactive NHx and H intermediates, which effectively reduce
NO. N2O was found to be purely a product of the secondary
interaction between adsorbed ammonia intermediate species
(NHx) with NO. The absence of N2O during the NH3–O2
and NH3–NO reactions in the Knudsen diffusion regime
(i.e., at a very low peak pressure) is related to the detection
limit of mass spectroscopic analysis. This may also explain
the invisibility of N2O in previous UHV studies, where the
pressure is much lower than in the TAP reactor. Three differ-
ent reaction routes of N2 formation have been established,
involving the secondary interaction between NO and NH3,
but also the direct decomposition of NH3 or NO. The dom-
inant route depends on the surface coverage of N-species,
as demonstrated by sequential pulse studies at different time
delays. Based on the experimental results presented in this
paper and in [1], an improved reaction network for high-
temperature NH3 oxidation has been proposed, taking into
A final aspect to address is the influence of oxygen on
the NH3–NO reaction. This point is relevant since the par-
ticipation of adsorbed oxygen species in ammonia activation
(Eq. (17)) should make them also fundamental to the initia-
tion of the reduction of nitric oxide by ammonia. As stressed
by several authors [28], the rate of the NH3–SCR of NOx is
enhanced by increased partial O2 pressure in the feed. Ka-
tona et al. [9] reported a 10-fold higher reaction rate between
NH3 and NO in the presence of O2. In our study, oxygen
species can be derived from (i) the high-temperature O2

J. Pérez-Ramírez et al. / Journal of Catalysis 229 (2005) 303–313
313
account the influence of oxygen coverage on the formation
of NO, N2O, and N2. This information enables to locate loss
reactions in the gauze pack, providing the basis for rational
strategies for preventing or minimizing by-product forma-
tion in industrial ammonia burners.
[7] C.G. Takoudis, L.D. Schmidt, J. Phys. Chem. 87 (1983) 964.
[8] T. Katona, L. Guczi, G.A. Somorjai, J. Catal. 132 (1991) 440.
[9] T. Katona, L. Guczi, G.A. Somorjai, J. Catal. 135 (1992) 434.
[10] T. Katona, G.A. Somorjai, J. Phys. Chem. 96 (1992) 5465.
[11] K. Otto, M. Shelef, J.T. Kummer, J. Phys. Chem. 74 (1970) 2690.
[12] K. Otto, M. Shelef, J.T. Kummer, J. Phys. Chem. 75 (1971) 875.
[13] R.J. Putasen, J.R. Katzer, W.H. Manogue, AIChE J. 20 (1974) 219.
[14] M. Slinko, T. Fink, T. Loeher, H.H. Madden, S.J. Lombardo, R. Im-
bihl, G. Ertl, Surf. Sci. 264 (1992) 157.
[15] L. Andrussow, Angew. Chem. 39 (1926) 321.
[16] S.J. Lombardo, M. Slinko, T. Fink, T. Löher, H.H. Madden, F. Esch,
R. Imbihl, G. Ertl, Surf. Sci. 269/270 (1992) 481.
[17] J.H. Jones, J.T. Kummer, K. Otto, M. Shelef, E.E. Weaver, Environ.
Sci. Technol. 5 (1971) 790.
[18] P.A. Kozub, G.I. Gryn, I.I. Goncharov, Platinum Metals Rev. 44 (2000)
77.
[19] J.M. Bradley, A. Hopkinson, D.A. King, J. Phys. Chem. 99 (1995)
17032.
[20] J.L. Gland , V.N. Korchak, J. Catal. 53 (1978) 9.
[21] M.F.H. van Tol, J. Siera, P.D. Cobden, B.E. Nieuwenhuys, Surf.
Sci. 274 (1992) 63.
Acknowledgments
Yara International ASA is acknowledged for financial
support and for permission to publish the results for Pt–Rh
alloy gauze. VAK and MB are grateful for financial support
from the Deutsche Forschungsgemeinschaft (DFG) in the
framework of the competence network “Bridging the gap be-
tween real and ideal systems in heterogeneous catalysis SPP
1091.” The authors are indebted to L. Mader for technical
assistance in the TAP experiments.
[22] J. Pérez-Ramírez, F. Kapteijn, K. Schöffel, J.A. Moulijn, Appl. Catal.
B 44 (2003) 117.
[23] R.J. Farrauto, C.H. Bartholomew, in: Fundamentals of Industrial Cat-
alytic Processes, Chapman–Hall, London, 1997, p. 481.
[24] H.F. Rase, in: Handbook of Commercial Catalysts. Heterogeneous
Catalysts, CRC, Boca Raton, 2000, p. 234.
[25] E.V. Kondratenko, J. Pérez-Ramírez, Catal. Lett. 91 (2003) 211.
[26] J.M. Bradley, A. Hopkinson, D.A. King, Surf. Sci. 371 (1997) 255.
[27] M. Kim, S.J. Pratt, D.A. King, J. Am. Chem. Soc. 122 (2000) 2409.
[28] H. Bosch, F. Jansen, Catal. Today 2 (1988) 369.
[29] R.M. Heck, J.C. Bonacci, W.R. Hatfield, T.H. Hsiung, Ind. Eng. Chem.
Process Des. Dev. 21 (1982) 73.
References
[1] J. Pérez-Ramírez, E.V. Kondratenko, V.A. Kondratenko, M. Baerns,
J. Catal. 227 (2004) 90.
[2] J. Pérez-Ramírez, E.V. Kondratenko, Chem. Comm. (2004) 376.
[3] S.J. Lombardo, F. Esch, R. Imbihl, Surf. Sci. 271 (1992) L367.
[4] S.J. Lombardo, T. Fink, R. Imbihl, J. Chem. Phys. 98 (1993) 5526.
[5] J.L. Gland, V.N. Korchak, J. Catal. 55 (1978) 324.
[6] C.G. Takoudis, L.D. Schmidt, J. Phys. Chem. 87 (1983) 958.