Catalytic Oxidation of Ammonia To Nitrogen
Catalytic Oxidation of Ammonia To Nitrogen
Catalytic Oxidation of Ammonia To Nitrogen
Proefschrift
door
Lu Gang
geboren te Nanjing-China
Dit proefschrift is goedgekeurd door de promotoren:
Gang, Lu
The work described in this thesis has been carried out at the Schuit Institute of
Catalysis, Laboratory of Inorganic Chemistry and Catalysis, Eindhoven University of
Technology, The Netherlands. Finacial support has been supplied by the Netherlands
Technology Foundation (STW) under the auspices of the Netherlands Organization
for Scientific Research (NWO).
Summary…………………………………………………………….……………...125
Samenvatting………………………………………………………………………..129
Publications………………………………………………………….……………...133
Acknowledgments…………………………………………………………………..135
Chapter 1
General Introduction
Abstract
In this chapter the problems related to the emission of NH3, the sources of emission
and the possibilities for their removal are discussed. The emphasis is on the removal
of ammonia from gas phase. The selective catalytic oxidation (SCO) of ammonia is
believed to be a most efficient and promising method for abating ammonia emissions.
The state of the art with respect to catalytic solutions to the problem is discussed in
detail with emphasis on the catalyst components and performance. Finally the
mechanisms of ammonia oxidation on metals and metal oxides are reviewed. The
scope of the investigation is also described briefly.
General Introduction
The emissions of nitrogen oxides (NOx) and sulphur oxides (SOx) give rise to
acidification of the environment. NOx and SOx are converted in the atmosphere to
give nitric and sulphuric acid. However emission of ammonia causes acidification of
the environment in an indirect way. Reaction of ammonia with acidic aerosols in the
atmosphere, such as aerosols of sulphuric acid or nitric acid, gives aerosols containing
ammonium sulphates or ammonium nitrates [1]. Oxidation of ammonia and
ammonium-containing aerosols by microorganisms in the soil then gives acidic HNO3
[2]. Emission of nitrogen oxides and sulphur oxides are the major sources of acidic
deposition in most countries. However emission of ammonia from intensive
agricultural activities, e.g. cattle breeding, makes a significant contribution to the
acidification of the environment in The Netherlands. About 94% of the ammonia
emitted originates from agricultural sources in The Netherlands [3]. The damage
caused by acidification in The Netherlands is serious: about half of the forests and
much of the heather are affected; most of the fens have turned acidic; the nitrate
concentration in the groundwater has increased and is still rising [4]; and the nitrogen
balance in the ecosystem is seriously disturbed [5]. Ammonia can be health damaging
and has an irritating smell. It is known to be a primary pollutant causing severe
irritation and is suspected to have long-term effects such as bronchitis. Furthermore,
ammonia is an undesired impurity for many industrial processes because it can cause
corrosion and plugging of instrumentation.
The major source of ammonia emission has been attributed to the intensive farming
areas and notably to livestock manure. This has led to a recent study of the role of
ammonia in the formation of a rural version of urban smog [6]. In contrast, NH3 is
used beneficially in industry to reduce NOx emissions by the so-called selective
catalytic reduction process (SCR). It is added to the effluent gas as reductant in order
to perform the following reaction: NH3 + NO + 1/4 O2 → N2 + 3/2 H2O. However,
the reaction is complete only with an excess of NH3, giving rise to NH3 “slip”. The
unreacted ammonia is present in the off-gas and must therefore be removed in a
secondary step [7,8]. Ammonia emissions are also caused by various other sources
like: soda production; nitric acid production; metallurgical industry; coal or biomass
gasification.
2
General Introduction
The emission limits for the emission of ammonia to the air are described in the Dutch
Emission Directives (Nederlandse emissie Richtlijn, NeR). For the industrial sector,
ammonia is classified in the class gaseous inorganic 4 (gA.4). This means that,
according to the NeR, the concentration in the outgoing gas flow may not be larger
than 200 mg/Nm3. This limit applies to polluted gas flows of at least 5,0 kg/hr.
However there might be plans to group ammonia in the class gA.3. This would mean
that the concentration must not be larger than 30 mg/Nm3, valid for polluted gas flows
of at least 0,300 kg/hr. The NeR-limits for the agricultural sector are much stricter.
The ammonia concentration in the outgoing gas flow may not be larger than 5
mg/Nm3. This applies to all polluted gas flows. The ammonia emission for the
Netherlands was 188.000 tons in 1996, 188.000 tons in 1997 and 177.000 tons in
1998 [12].
2. Removal of ammonia
Removal of ammonia can be divided in two groups (i) removal of ammonia from
liquid phase and (ii) removal of ammonia from gas phase.
Conventional wastewater treatment plants have only limited capacity for nitrogen
removal. Plant effluents thus contain large concentrations of nitrogen, with ammonia
being the main form. Ammonia concentrations in biologically treated wastewater are
generally in the range 12-35 mg/L [13]. Several technically and economically feasible
methods are available to reduce this concentration. They are characterized in Table 1.
3
General Introduction
The relative costs of the treatment processes listed in Table 1 are controversial, and
some authors claim that ammonia desorption is by far the least expensive ammonia
removal process [16]. It should be noticed that the ammonia desorption method is
actually change the liquid phase problem into a gas phase problem. Other methods
such as catalytic wet oxidation and electrochemical oxidation are still under
development [17,18,76].
Two main sources of waste gas containing ammonia can be identified: gas from
industrial processes and gas from agriculture. Examples of industrial processes that
produce ammonia are gasification, carbonisation, HDN process and thermal treatment
of coal, liquid petroleum, shale oil, biomass and tar sands. The name of the
contaminated gas flow is Cokes Oven Gas (referred to as COG-gas). The main source
of agricultural pollution in Noord-Brabant is ventilation from sties, stables and
manure processing factories. The name of the contaminated gas flow is Sty-gas.
4
General Introduction
The gas generated from COG-gas usually contains high concentration of ammonia. A
conventional method of removing ammonia from COG is conducted by washing the
COG with dilute sulfuric acid to recover the ammonia as ammonium sulfate. However
the demand of ammonium sulfate for fertilizer has decreased and the market price
greatly lowered. As a result, the profit is remarkably inferior and this process is now
almost worthless from the industry viewpoint. At present, the ammonium sulfate
production process is reduced and changed to orther processes such as the Phosam
process or Chevron WWT process to produce highly pure liquid ammonia, the
Koppers process to separate ammonia followed by direct combustion or the Carl Still
process to burn ammonia in the presence of a catalyst. However in the case of the
direct combustion process, it is difficult to inhibit the production of NOx [19-21]. An
alternative method to remove NH3 from COG is the decomposition of NH3 at high
temperature in the presence of a catalyst [22-24]. A few common techniques are
schematically presented in Fig.1.
Biological and acid scrubbing are two current techniques that are available for the
removal of NH3 from Sty-gas, which usually contains low concentrations of ammonia
(<1000 ppm). In Noord-Brabant only one manure processing plant is operational
(CNC) and another pilot plant is in experimental state (Ferm-O-Feed). Acid scrubbing
and biofilters are techniques that are used in manure processing. The scale of the gas
flow varies strongly per plant (10,000 – 300,000 m3/hr). Biofilters are mainly used for
5
General Introduction
Apart from COG and Sty-gas, many industrial processes, such as textile treatment,
soda production, SCR process and nitric acid production, produce emission streams
containing ammonia. These streams usually contain excess O2, H2O and very low
concentrations of NH3 (<1000 ppm). The conventional method for NH3 removal is
again by acid scrubbing. It should be pointed out that the acid scrubbing actually
change a gas phase problem into a liquid phase problem. At present, the scrubbing
liquid containing either NH4NO3 or (NH4)2SO4 is sold as a fertilizer. As discussed
above, the market for such a fertilizer is reducing due to its low efficiency and the
possibility to cause severe soil acidification. In the long run, it is possible to stop use
such a kind of fertilizer and the scrubbing liquid then becomes a big problem.
Furthermore, acid scrubbing is not suitable to high flow and high temperature flue
gases because of the high operation and investment costs [25].
A new technology for the removal of ammonia is using selective catalytic oxidation
of ammonia to nitrogen and water. This process which provides an efficient, stable,
simple and selective purification of large gas emissions can be applied both in low
and high concentrations of ammonia removal. It is also possible to selectively oxidize
ammonia to nitrogen in the liquid phase. Another potential application of this process
may be in the production of pure nitrogen used as safety gas from air and ammonia.
At present nitrogen is mainly produced from air by physical separation, which needs
large investments and much energy. Ammonia oxidation by air provides a simplified
and flexible process that is especially suitable for small-scale nitrogen production.
The catalytic oxidation of ammonia is one of the most interesting and important
heterogeneous catalytic processes. It can proceed in the following three principal
ways:
6
General Introduction
In the presence of platinum or cobalt oxide at high temperature (750-900 oC), mainly
NOx is produced. This constitutes the basis of the industrial manufacture of nitric acid
(Ostwald process). At low temperature (< 500 oC), all three nitrogen-containing
products (N2, N2O and NO) are formed simultaneously in various proportions in the
presence of many catalysts. The study of low temperature SCO can be further
classified into two categories: high ammonia concentration (1-30%) and low ammonia
concentration (< 1000ppm) oxidation. The latter has only recently become of more
interest because of the environmental problems.
Though much attention has been given to high temperature ammonia oxidation, there
is still a lot of detailed research in the literature about the low temperature ammonia
oxidation over various kinds of metal and metal oxide catalysts. Earlier work on
ammonia oxidation was reviewed by II’chenko [26]. In his review, the activities of
metals and metal oxides for ammonia oxidation at low temperature have been
systematically compared (Table 2 and Table 3).
7
General Introduction
Table 2 shows that at 300 oC the specific catalytic activity of metals for the overall
process decreases in the sequence: Pt>Pd>Cu>Ag>Au>Fe>W>Ti. At low
temperature, the ammonia oxidation products in the presence of metals are N2 and
N2O. The selectivity with respect to nitrous oxide increases with temperature and for
transition metals decreases in the sequence: Pt, Pd>Ni>Fe>W>Ti.
Table 3 shows that at 230 oC the specific catalytic activities of the oxides decrease as
follows:
Co3O4,MnO2>CuO>CaO>NiO>Bi2O3>Fe2O3>V2O5>TiO2>CdO>PbO>ZnO>SnO2
>ZrO2>MoO3>WO3.
In another study [27], a similar sequence was obtained:
Co3O4 >MnO2>Cr2O3> Fe2O3>CuO>NiO>V2O5>MoO3>U3O8>ThO2>WO3>SnO2>
ZrO2>ZnO>Nb2O5>Bi2O3>Sb2O4>Ta2O5.
8
General Introduction
The effect of reaction conditions on catalyst performance is the same for both metal
and metal oxide catalysts. Increasing temperature and O2/NH3 ratio will increase the
activity but decrease the selectivity to nitrogen.
Most above mentioned work has been done with single, polycrystaline metals and
simple metal oxides, less with supported metals and metal oxides, metal alloys and
mixed metal oxides. Several papers published some results of ammonia oxidation
over alumina supported Pt catalysts [28-30]. It was discovered that a significant
deactivation took place over these catalysts at low temperature (below 200 oC). The
crystal size of Pt particle had great influence on catalyst performance. However at
high temperature (200-350 oC), no deactivation was observed [30]. For transition
metal exchanged zeolites, zeolite Y was the only zeolite reported [31]. By pulsing
with a dry NH3/O2 mixture, the activity sequence reported was as follows:
CuY>CrY>AgY>CoY>FeY>NiY,MnY, a little different from the sequence for
simple metal oxides. NKK of Japan [32] recently developed a process for
decomposing ammonia recovered from coke oven gas through catalytic oxidation.
Over TiO2-Al2O3 supported Cu-V oxides catalysts, a gas containing about 10%
ammonia at 25500 h-1 space velocity can be completely cleaned from NH3 at a
temperature of about 500 oC with only a few ppm NOx emission. It should be noticed
that there exist some sulphur containing substances (H2S, H2SO4) in their ammonia
containing gas.
In recent years a considerable amount of attention has been focused on the catalytic
purification of ammonia-containing flue gases with ammonia often present in low
concentrations (<1000ppm). Since the surface concentration of nitrogen containing
substances plays an important role in the selectivity of the ammonia oxidation, some
influence of NH3 concentration on the activity and selectivity of the catalyst may be
expected. De Boer’s study [33] shows that both activity and selectivity of ammonia
oxidation over supported molybdenum catalyst are greatly influenced by the ammonia
concentration (see Table 4).
9
General Introduction
The systems investigated in literature for low ammonia concentration oxidation are
V2O5, WO3 and MoO3 on various supports [33-39] at the temperature of 300--400 oC.
The best catalyst was SiO2 supported molybdenum promoted with PbO over which
the ammonia oxidation could be finished completely with 100% nitrogen selectivity at
the temperature of around 400 oC.
Yuejin Li studied the selective NH3 oxidation to N2 in a wet stream over ZSM-5 and
alumina supported Pd, Rh and Pt catalysts at 200-350 oC [42]. They concluded that
the ammonia conversion was not affected at high temperature but decreased at 200-
250 oC with the addition of 5% water vapour. Generally, the ion exchanged ZSM-5
catalysts are more active than the alumina catalysts with an identical metal loading
and less affected by water vapour. The selectivity to N2 is relatively high on Rh and
Pd catalysts and low on Pt catalysts.
10
General Introduction
Wollner reported that high degrees of ammonia conversion of 80-100% was obtained
over mixed copper-manganese oxides supported on titania catalysts at temperatures
greater than about 300 oC [43]. However the selectivity is not reported clearly.
The vast majority of scientific work on ammonia oxidation has been done on platinum
catalysts. Therefor, in this section mechanisms of ammonia oxidation on platinum are
first discussed. Furthermore, mechanisms of ammonia oxidation on other catalysts
such as Cu and Ag, as well as transition metal oxides are also reviewed.
Before 1960 three different general reaction mechanisms on platinum were proposed
on the basis of postulation because of the lack of experimental evidence of
intermediate species. The nitroxyl (HNO) mechanism was suggested by Andrussow
[44]. His hypothesis was based on the formation, during the first stage of the reaction,
of the intermediate compound HNO:
Later Bodenstein [45,46] conjectured that the first stage was the formation of
hydroxylamine (NH2OH), which was converted to nitroxyl in a next step:
The imide (NH) mechanism was proposed by Raschig [47] and Zawadzki [48], in
which the first step yielding imide was postulated. The imide could subsequently react
with oxygen or ammonia to form nitroxyl or hydrazine:
11
General Introduction
According to above mechanisms, the final products NO, N2O, N2 and H2O are formed
via a number of stages involving such intermediate compounds as HNO, HNO2,
HNO3 and N2H4 [44-52].
Fogel et al. used polycrystalline Pt wire, with reactant partial pressure of 10-4 Torr,
and worked over the temperature range 25-1200 oC [53]. Using secondary ion mass
spectrometry (SIMS), they concluded that the intermediate species HNO, NH2ON,
HNO2 and N2O were not formed during the oxidation reaction and therefore proposed
a reaction mechanism which had simple steps, with none of the above intermediates
involved, based on the two reactions:
NH3 + O → NO + H2 + H (10)
NH3 + NO → N2 + H2O + H (11)
It should be pointed out that the above models for the overall behavior of the
ammonia oxidation reaction on Pt appear reasonable. They do, however, address only
the overall reaction mechanism and do not go into the details of the elementary
reaction steps. New insights in the reaction mechanism of ammonia oxidation were
obtained from surface science experiments. Asscher et al. [56] used a laser
multiphoton technique, combined with molecular beams under UHV conditions, to
probe NO formation during reaction on a Pt{111} crystal. By analyzing the decay in
the NO signal after pulsing NH3, they identified two different NO production kinetics
and assigned them to the following mechanisms:
NH + O → NO + H (12)
N + O → NO (13)
12
General Introduction
They found reaction (12) to be the faster mechanism, which was enhanced at high
O2/NH3 ratios and had a higher activation energy than reaction (13), the slower
mechanism.
More recently the ammonia oxidation was studied by Mieher and Ho [57] on the
Pt{111} surface with TPD, temperature programmed reaction spectroscopy (TPRS),
electron energy loss spectroscopy (EELS) and low energy electron diffraction
(LEED). Using EELS they identified OH, NH and NH2 as intermediate species but
were unable to comment on NO due to ambiguities of the NO frequency. They
concluded that the reaction proceeded via the simple stripping of NH3 by oxygen
atoms followed by the combination of nitrogen atoms with oxygen, to form NO, or
with nitrogen atoms, to form N2 (reaction (14)-(18)):
Another recent surface science study on ammonia oxidation was done by Bradley and
King [58] on Pt{100} using surface molecular beams under UHV conditions. The
mechanism proposed is similar to the mechanism of Mieher and Ho. However the
production of nitrogen and NO was found to occur through other paths. They
suggested that NO was directly produced by reacting NH with O. The formation of N2
resulted from the consecutive dissociation of NO:
NH + 2 O → NO + OH (19)
NO → N + O (20)
N + N → N2 (21)
Since only NO and N2 were produced under the reaction conditions of these studies,
the formation of N2O was not discussed. The recent study of Van den Broek and van
Santen [59,60] showed that NH and OH were the main surface species adsorbed on Pt
sponge catalysts after ammonia oxidation in atmospheric pressure. The reaction of
NH and OH was thought to be the rate-determining step under their reaction
conditions. They argued that N2O was produced by NO reacting with adsorbed N. A
density functional study has also been done by Fahmi and van Santen [61] on
ammonia oxidation on a Pt6 cluster. It was found that ammonia only dissociates when
oxygen is present on the surface. The reaction mechanism was found to generally
13
General Introduction
follow the Mieher and Ho mechanism. Furthermore it was established that water
could poison the ammonia oxidation reaction, since water is more acidic than
ammonia.
The XPS, EELS and STM techniques were used by the group of M.W. Roberts [62-
65] to study ammonia oxidation on copper surfaces. They observed that at very low
temperatures (< 300 K) ammonia could be oxidized to adsorbed NH2 and NH species
through oxydehydrogenation steps. At higher temperature (400 K) a fraction of the
imide species was further dehydrogenated into atomic nitrogen:
The STM image clearly showed that the atomic nitrogen produced at 400 K could
block the ends of the –Cu-O- rows, inhibiting further reactions and creating stable
mixed N-O structures on the copper surface. Step defects on the surface had strong
influence on the reactivity of oxygen adatoms. Reactivity was high at the top and
bottom of a [110] step and at the bottom of a [001] step, whereas it was low at the top
of a [001] step.
Ammonia adsorption on Ag(110) has been studied previously by TPD, TPRS and by
EELS with and without coadsorption of molecular and atomic oxygen [66]. It was
concluded that a diffusing nitrogen adatom was the reactive intermediate in NO and
N2 formation. On the basis of a combination of XPS and vibrational electron energy
loss (VEEL) spectra a dioxygen-NH3 complex had been suggested to be a key
intermediate in the oxidation of ammonia on Ag(111) surfaces [67,68]. No report has
been published to-date on the intermediate species and reaction mechanisms on
polycrystalline silver at atmospheric pressure.
Ammonia oxidation mechanisms on Cr2O3, MoO3, Fe2O3 and ZnO have been studied
by the group of Matyshak [69-72] using spectrokinetic method. Surface complexes
NH3ads, NH2ads, and NOads were shown to be intermediate species. They concluded
that ammonia oxidation was first preceded by ammonia adsorption in the form of
coordinated ammonia and then through an oxy-dehydrogen step producing NH2. The
NH2 could be directly oxidized to NO, which will react with NH2 to give N2. The N2O
was believed to be produced by NO2 reacting with NH2:
14
General Introduction
Under SCO catalytic reactions the following reoxidation reaction closes a Mars-Van
Krevelen-type redox mechanism:
In summary on both metal and metal oxide catalysts two major routes have been
proposed in the literature for the selective production of N2 from the oxidation of
ammonia. The first is a direct route based on the oxidation of NHx species directly to
atomic nitrogen and then the recombination of two nitrogen atoms forming N2, or the
recombination of NHx species giving rise to an hydrazinium intermediate N2H4 and
then oxidizing to N2. The other is the in situ or “internal” selective catalytic reduction
(SCR) and is a two-step mechanism, in which the NHx species are oxidized to NOx
species in the first step. The NOx species reacted consequently with NHx species
giving rise to N2 through a surface SCR reaction.
It can be seen from above review that various catalysts of different types have been
tested for the low temperature ammonia oxidation reaction: biological catalysts, metal
oxide catalysts, ion-exchanged zeolites and metallic catalysts. When the various types
of catalysts are compared it appears that the metallic catalysts such as Pt and Ir are the
most active but less selective. Significant amounts of N2O are produced on these
catalysts. The metal oxide catalysts such as CuO, V2O5 and MoO3 show very
15
General Introduction
promising selectivity but the reaction temperature needed is too high to be matched
with some industrial applications.
The research described in this thesis was aimed at the development of new, active and
selective catalysts for low temperature (<300 oC) selective ammonia oxidation to
nitrogen. In Chapter 2 an elementary catalyst screening study was performed in order
to find new catalysts for the SCO reaction. The emphasis was on zeolite-based and
alumina-supported metal or metal oxide catalysts. Copper-based and silver-based
catalysts were found to be the most promising catalysts for low temperature SCO
process. Therefore in Chapter 3 various kinds of copper oxides supported on alumina
and on zeolite Y have been prepared and studied in detail. TPD, TPR, UV-vis
spectroscopy and high-resolution electron microscopy (HREM) were used to
characterize these catalysts in an attempt to shed light on the optimal preparation for
active and selective low-temperature ammonia oxidation catalysts. The results showed
that a CuAl2O4-like phase was more active than a CuO phase for SCO reaction. For
copper zeolite catalysts [Cu-O-Cu]2+-like species or small copper oxygen aggregates
were the likely forms of the catalytically active centers at low temperature. The
activity of CuY was increased by treating the sample with NaOH. This treatment
presumably increases the amount of low temperature active centers. In Chapter 4
ammonia oxidation reaction pathways on high surface area silver powder have been
studied by TPD, TPR, FT-Raman and transient as well as steady-state ammonia
oxidation experiments. It was found that NO was the main reaction intermediate to
give N2O as well as N2. Even at room temperature NO could be formed and, when
oxidized to NOx, it could become adsorbed to the silver surface, which will block the
active sites for oxygen activation. The adsorption of oxygen was thus believed to be
the rate-controlling step for ammonia oxidation. The adsorbed NOx, N2Ox species
were actually inhibitors for ammonia oxidation but these adsorbed species lowered
the surface oxygen coverage. So the selectivity to nitrogen was improved with the
increasing amount of these adsorbed species. In Chapter 5 low temperature gas phase
oxidation of ammonia to nitrogen has been studied over alumina-supported, silica-
supported and unsupported silver catalysts to distinguish the support effect on silver-
based catalysts. TPD, TPR, TEM, XRD and FT-Raman were used to characterize the
different silver catalysts. The results showed alumina-supported silver to be the best
catalyst due to the interaction of silver with alumina. Pretreatment had a great affect
on the catalyst performance. Reduction in hydrogen at 200 oC without any pre-
calcination gave the best activity while reduction at higher temperatures showed little
difference from calcination pretreatment. At least four types of oxygen species were
produced when silver was oxidized at high temperature. These species are adsorbed
molecular oxygen, adsorbed atomic oxygen, strongly adsorbed atomic oxygen and
16
General Introduction
References
17
General Introduction
18
General Introduction
19
General Introduction
[73] J.M.G. Amores, V.S. Escribano, G. Ramis and G. Busca, Appl. Catal. B, 13
(1997) 45.
[74] G. Ramis, L. Yi and G. Busca, Catalysis Today, 28 (1996) 373.
[75] M. Trombetta, G. Ramis, G. Busca, B. Montanri and A. Vaccari, Langmuir, 13
(1997) 4628.
[76] A.C.A. de Vooys, Ph.D. Thesis, Technical University of Eindhoven, Eindhoven,
2001.
20
Chapter 2
NH3 oxidation to nitrogen and water at low temperatures
using supported metal or metal oxide catalysts
Abstract
1. Introduction
Early research involving ammonia oxidation has been reviewed and the activities of
several metals and metal-oxides for nitrogen and NOx formation at low temperature
have been systematically compared [3]. Most of these studies involved either single,
poly-crystalline metals or simple metal oxides. Supported metals, supported metal-
oxide, metal alloys, and mixed- metal oxides have been studied to a lesser extent.
Several papers have been published regarding the ammonia oxidation over alumina-
supported platinum catalysts [4-6]. It was discovered that significant deactivation
occurred on these catalysts at temperatures below 200 oC. To date zeolite Y is to our
knowledge the only zeolite studied in some detail for this reaction [7]. Some initial
screening studies have also been reported for ZSM-5 [8]. Investigations involving
ammonia oxidation on V2O5, WO3 and MoO3 , themselves supported on various
metal-oxides, at temperatures between 300° and 400 oC have been reported [9-15]. Of
these, the best catalyst reported was a silica- supported, PbO-promoted, molybdenum
catalyst on which ammonia could be oxidized with 100 % selectivity to nitrogen at
temperatures of about 400 °C [13]. Li and Armor studied the selective NH3 oxidation
to N2 in wet streams over ion-exchanged ZSM-5 and alumina-supported Pd, Rh and Pt
catalysts at 200° to 350 oC [16]. They concluded that the ammonia conversion was not
affected by co-feeding steam at high temperature but was decreased at lower
temperatures (200o - 250 oC) when 5 vol % water vapor was added. Generally, ion-
exchanged ZSM-5 catalysts were more active than alumina-supported catalysts of
identical metal loadings and were less affected by the addition of water vapor. The
selectivity to N2 was observed to be relatively high on Rh and Pd catalysts and low on
Pt catalysts. Wollner reported that high degrees of ammonia conversion (80 – 100 %)
could be obtained over mixed copper- / manganese-oxides supported on titania at
temperatures greater than about 300 oC [17]. Unfortunately the selectivity of this
process was not clearly reported.
The above studies show that noble metals such as platinum are very active for
ammonia oxidation but form large amounts of nitrogen oxides. Although supported
22
NH3 oxidation to nitrogen and water at low temperatures
using transition metal catalysts
molybdenum and vanadium catalysts show very promising selectivity, the reaction
temperature needed is a too high to be matched with some industrial applications. The
aim of this study is the development of active new catalytic materials that are capable
of selective oxidation of ammonia to nitrogen at low temperatures. For this purpose
we have screened the performance of several transition-metals deposited either onto
alumina (by incipient wetness impregnation) or into sodium zeolite Y by ion-
exchange. The results obtained are reported below.
2. Experimental
23
NH3 oxidation to nitrogen and water at low temperatures
using transition metal catalysts
Various ion-exchanged Na zeolite-Y catalysts were tested and the resulting ammonia
conversion and nitrogen selectivity data are shown in Table 1. Clearly, only CuY is
very active and selective for ammonia oxidation. All of the other ion-exchanged
catalysts had almost no activity, in fact they are worse than NaY itself. This is quite
different from the simple metal-oxide catalysts published in the literature.
24
NH3 oxidation to nitrogen and water at low temperatures
using transition metal catalysts
CuY 250 25 97
(8.4 wt%) 300 88 97
CuY(3.7%) 200 19 97
(after NaOH treatment) 250 56 97
300 100 98
CuY(8.4%) 200 35 95
(after NaOH treatment) 250 68 97
300 100 98
AgY 250 11 75
(after NaOH treatment) 300 81 80
CoY 250 14 71
(after NaOH treatment) 300 42 75
Reaction conditions: NH3 =1.14 vol%; O2 = 8.21%; flow rate = 74.7 Nml/min; cat. weight = 0.2g
Ione et al. [18] and Schoonheydt et al. [19] reported that polynuclear nickel or copper-
ion complexes were formed in the framework of zeolite Y when NaY was ion-
exchanged with aqueous Ni(NO3)2 or Cu(NO3)2 solutions at pH 6-7. These
polynuclear cations gave a higher activity for CO oxidation than did the mononuclear
ions. Suzuki et al. also successfully prepared an excellent CO oxidation catalyst of
highly-dispersed, nickel-oxide catalyst by hydrolysing a Ni2+-exchanged zeolite Y
with either aqueous NaOH or ammonia solutions at different pH values. We were
curious to see the effect that such NaOH treatment would have on the ammonia
oxidation reaction. Table 2 shows resulting conversion and selectivity data obtained
following treatment of zeolite-based catalysts with NaOH solutions (pH = 10)
following ion-exchange. This treatment procedure was the same as proposed by
Suzuki et al. [20,21].
25
NH3 oxidation to nitrogen and water at low temperatures
using transition metal catalysts
6.00E-03
MS intensity, a.u.
4.00E-03 1
2.00E-03
2
0.00E+00
0 100 200 300 400 500 600
Temperature, oC
The activity of all metal zeolite catalysts for ammonia oxidation increased drastically
following this NaOH treatment. This significantly higher activity was apparently
induced by the formation of small metal-oxide particles in the zeolite. As can be seen
in Fig.1 there are two peaks for O2 TPD profiles on an 8.4 wt % copper ion-
exchanged zeolite Y (CuY-8.4) catalyst. Since no O2 desorption is observed on NaY,
these two O2 desorption peaks must involve the Cu on the zeolite. Following NaOH
treatment the amount of O2 desorbed was greatly increased. Many studies on the
complete oxidation of CO and of hydrocarbons have shown that the activity is directly
dependent on the amount of adsorbed oxygen [22,23]. This increased oxygen
adsorption capacity of the catalyst following NaOH treatment may be the reason for
the enhanced performance in ammonia oxidation.
26
NH3 oxidation to nitrogen and water at low temperatures
using transition metal catalysts
Cu/Al2O3 250 15 97
(10 wt%) 300 90 96
350 100 90
Cu/Al2O3 250 9 97
(15 wt%) 300 46 96
350 100 94
Ag/Al2O3 200 11 91
(10 wt%) 250 98 86
Ag/Al2O3 200 19 92
(15 wt%) 250 100 88
Mo/Al2O3 300 8 88
(15 wt%) 400 74 89
V/Al2O3(10%) 400 45 95
Reaction conditions: NH3 = 1.14%; O2 = 8.21%; (vol%)
flow rate = 74.7 Nml/min; cat. weight = 0.2g
27
NH3 oxidation to nitrogen and water at low temperatures
using transition metal catalysts
70
60
NH3 Conversion, %
50
40
30
20
Fig. 2 The stability of Cu-zeolite Y catalysts: NH3 conversion vs. time on stream at different reaction
temperatures.
CuY(1)---ion-exchanged with 8.4 wt% loading;
CuY(2)---after NaOH treatment, 3.7 wt%;
CuY(3)---after NaOH treatment, 8.4 wt%
(Reaction conditions: NH3 = 1.14 vol%; O2 = 8.21 vol%; Flow rate = 74.7 Nml/min; cat. weight =
0.2g)
Fig.2 shows the measured NH3 conversion versus time on stream at various
temperatures for Cu-ion exchanged catalysts under high ammonia concentration
conditions. It can be seen in the figure that the catalysts were fairly stable. Catalysts
subjected to NaOH treatment showed a slight initial deactivation at 250 oC. But at 200
o
C deactivation was absent. The CuY catalyst untreated by NaOH showed a very
stable activity, perhaps even a slight increase in activity with time. Large initial
deactivation was observed for alumina-supported copper and for the reduced noble
metal catalysts (see Fig. 3) during similar ammonia oxidation experiments. At low
ammonia concentration conditions no deactivation was observed during one-day
catalyst testing on all types of catalysts. As all of the alumina-based catalysts were
calcined at high temperature and there were no hydrocarbons in the reaction system,
metal sintering and coking were excluded as possible causes of this deactivation. Our
experiment also showed that the activity could be recovered completely by calcining
the deactivated catalyst at 500 oC again. This indicates that the metal sintering in our
28
NH3 oxidation to nitrogen and water at low temperatures
using transition metal catalysts
case is very unlikely. The most probable reason for deactivation is either
reconstruction of the metal surface or a change in the chemical state of the surface
caused by ammonia-induced species since ammonia concentration has a great
influence on the catalyst deactivation behavior.
120
300 oC
180 oC
NH3 Conversion, %
100
80 Ir/Al2O3
Ag/Al2O3
60 Pt/Al2O3
230 oC Cu/Al2O3
40
165 oC
20
0
0 4.1 8.2 12.3 16.4 20.5
Time on Stream, h
Fig. 3 The stability of various alumina-supported, transition-metal catalysts during ammonia oxidation
at different temperatures (NH3 / O2 = 0.14 v / v, flow rate=74.7 Nml/min, [NH3] = 11400 ppm in He,
cat. weight = 0.2g)
The O2 / NH3 ratio had a large affect on both the activity and the selectivity of all
catalysts. Increasing the O2 / NH3 ratio increased the activity but decreased the
selectivity (see Table 4). It should be noticed that the activity of silver-based catalysts
was greatly increased compared with other catalysts when the O2 / NH3 ratio was
increased. However, the O2 / NH3 ratio had less influence on the selectivity of CuY
catalyst. Co-fed steam in the feed gas lowered the activity of the catalysts at low
temperature but had less effect at high temperature (see Table 5). Since the ammonia
oxidation reaction is irreversible, water should have no effect on this reaction from
thermodynamic point of view. This effect may be caused by the adsorption of water
or by condensation of water on the catalysts resulting in partial blocking of the active
sites.
29
NH3 oxidation to nitrogen and water at low temperatures
using transition metal catalysts
Table 4. The dependence of the activity for ammonia oxidation on O2/NH3 feed ratio.
catalyst temperature O2 / NH3 NH3 N2 selectivity
o
( C) ratio conversion (%) (%)
CuY 300 0.68 21 99.5
(8.4 wt%) 300 2.60 46 99.0
300 7.20 88 97
Table 5. The effect of co-fed water vapor on the activity during ammonia oxidation.
Catalyst Temperature feed NH3 N2 selectivity
o
C conversion % %
Cu/Al2O3 300 dry 90 96
(10 wt%) 300 wet 21 97
350 dry 100 94
350 wet 100 95
In recent years considerable attention has been focused on the catalytic removal of
ammonia from flue gases. The ammonia concentration in such exhaust streams is
often quite low (<1000ppm). Since the surface concentration of nitrogen-containing
30
NH3 oxidation to nitrogen and water at low temperatures
using transition metal catalysts
substances plays an important role in the selectivity of the ammonia oxidation, some
influence of NH3 concentration on the activity and on the selectivity of the catalyst
may be expected. It has been shown previously, over supported-molybdenum
catalysts [15], that both the activity and the selectivity were greatly decreased by
decreasing the ammonia concentration.
Table 6. The activity of various catalysts for ammonia oxidation at low ammonia
feed concentration (1000 ppm) in helium.
catalyst Temperature NH3 conversion N2 selectivity
o
C % %
Ir/Al2O3(1.2 wt%) 180 52 88
(reduced) 190 91 86
200 100 84
Cu/Al2O3 200 16 93
(10 wt%) 230 66 86
250 95 82
Ag/Al2O3 130 56 82
(10 wt%) 140 95 82
160 100 81
Reaction conditions: NH3 = 1000 ppm; O2 = 10%; O2/NH3 = 100
flow rate = 50 Nml/min; cat. weight = 0.1g
Oxidation of ammonia at low concentrations (1000 ppm NH3 in He) was investigated
at low temperatures and the results are shown in Table 6. In all cases, the observed
selectivity of the catalysts decreased relative to those observed previously at higher
concentrations. However, the selectivity observed on CuY catalysts at low
concentrations still exceeded that obtained either on copper-alumina or on noble-
metal catalysts under similar conditions. The measured performance of these catalysts
at higher NH3 concentration (11400 ppm NH3 in He) is shown in Table 7.
31
NH3 oxidation to nitrogen and water at low temperatures
using transition metal catalysts
Ag/Al2O3 200 11 91
(10 wt%) 230 67 89
250 98 86
Rh/Al2O3(reduced) 350 25 90
(1.2 wt%) 380 78 88
400 100 86
250 60 97
Pd/Al2O3(reduced) 280 88 97
(1.2 wt%) 300 100 98
Reaction conditions: NH3 = 1.14 vol %; O2 = 8.21 vol %; O2/NH3 = 7.2
flow rate = 74.7 Nml/min; cat. weight = 0.2g
The results of the steady state kinetic measurements for different catalysts are given in
Table 8. Compared with noble metal catalysts the Ag/Al2O3 and CuY catalysts have a
much lower activation energy. This means that temperature has more influence on
32
NH3 oxidation to nitrogen and water at low temperatures
using transition metal catalysts
noble metal catalysts than on Ag/Al2O3 and CuY catalysts. It is clear that Ag/Al2O3 is
very sensitive to oxygen concentration since it has the highest reaction order in
oxygen. The CuY catalyst shows the highest reaction order in ammonia and for the
same reason the ammonia concentration has a great influence on this catalyst. On the
contrary the reaction orders in ammonia for noble metal catalysts are extremely low,
which indicates much less influence of ammonia concentration on the activities of
these catalysts.
120 100
Pd
CuY
100 Ir Ag Pd Ru
NH3 conversion, %
95 V
N2 selectivity, %
80 CuAl CuAl
Ir Ag
60 90
CuY Mo
40 Rh
85
Pt
20 Pt Mo
V
0 80
100 200 300 400 500 100 200 300 400 500
oC
o Temperature,
Temperature, C
120 100
Pt CuY
100 CuY
NH3 conversion, %
90
N2 selectivity, %
80 Ir
CuAl
60 Ir 80 Ag
Ag
40 Pt
70
20 CuAl
0 60
100 150 200 250 300 100 150 200 250 300
o
Temperature, o
C Temperature, C
For clearly comparison the results of Table 7 and Table 6 are ploted in Fig.4 and
Fig.5. It can be seen evidently that noble metal catalysts are very active especially at
high ammonia concentration, however their selectivity to nitrogen is very poor at low
33
NH3 oxidation to nitrogen and water at low temperatures
using transition metal catalysts
4. Conclusions
The activity of copper ion-exchanged zeolite Y catalysts for ammonia oxidation was
shown to be comparable to that of noble metal catalysts at low temperatures. The
selectivity to nitrogen was much higher for the zeolite catalysts. Treatment of CuY
with NaOH after ion-exchange increased the ammonia oxidation activity. Alumina-
supported silver catalysts were also very active for ammonia oxidation, especially at
high O2/NH3 ratios. With further improving the catalyst selectivity silver-based
catalysts are very promising to be applied to the removal of ammonia in flue gas. Co-
fed steam dramatically decreased catalyst activity, especially at lower temperatures.
Ion-exchanged zeolite Y catalysts were more stable than alumina-supported catalysts
at high ammonia concentration conditions. At low ammonia concentration conditions
both types of catalysts are stable. The deactivation of the catalyst is thus probably
caused by ammonia-induced species. It has been shown that the selectivity to nitrogen
was greatly decreased by decreasing the ammonia concentration. At high O2/NH3
ratio the activity of all catalysts increased but the selectivity for nitrogen production
decreased.
34
NH3 oxidation to nitrogen and water at low temperatures
using transition metal catalysts
References
[1] G.B. Barannik, V.F. Lyakhova and A.V. Simakov: Chemistry for Sustainable
Development,1995.
[2] J.J.P. Biermann, Ph.D. Thesis, University of Twente(1990).
[3] N.I. II’chenko, Russian Chem. Rev., 45(12)(1976)1119.
[4] J.J. Ostermaier, J.R. Katzer and W.H. Manogue, J. Catal., 41(1976)277.
[5] J.J. Ostermaier, J.R. Katzer and W.H. Manogue, J. Catal., 33(1974)457.
[6] J.E. Delaney and W.H. Manogue, Proc. Int. Congr. Catal. 5th, 1(1973)267.
[7] O.V. Al’tshuller and M.Ya. Kushnerev, Problemy Kinetiki I Kataliza,
15(1973)56.
[8] N.N. Sazonova, A.V. Simakov and H. Veringa, React. Kinet. Catal. Lett.,
57(1)(1996)71
[9] F.J.J.G. Janssen and F.M.G. van den Kerkhof, Selective Catalytic Removal of NO
from Stationary Sources, KEMA Sci. & Techn. Reports, 3(6)(1985).
[10] J.P. Chen and R.T. Yang, Appl. Catal., 80(1992)135.
[11] G. Tuenter, W.F. van Leeuwen and L.J.M. Snepvangers, Ind. Eng. Chem. Prod.
Res. Dev., 25(1986)633.
[12] E.T.C. Vogt, A. Boot, J.W. Geus and F.J.J.G. Janssen, J. Catal., 114(1988)313.
[13] E.T.C. Vogt, Ph.D. Thesis, University of Utrecht(1988).
[14] J.J.P. Biermann, F.J.J.G. Janssen and J.W. Geus, J. Mol. Catal., 60(1990)229.
[15] M. de Boer, A.J. van Dillen and J.W. Geus, Catalysis Letters, 11(1991)227.
[16] Y. Li and J.N. Armor, Appl. Catal. B, 13(1997)131.
[17] A. Wollner and F. Lange, Appl. Catal. A, 94(1993)181.
[18] K.G. Ione, P.N. Kuznetsov and V.N. Romannikov, Application of Zeolite in
Catalysis, Akademiai Kiado, Budapest, 87(1979).
[19] R.A. Schoonheydt, K.G. Ione, P.N. Kuznetsov and V.N. Romannikov, J. Catal.,
43(1976)292.
[20] M. Suzuki, K. Tsutsumi, H. Takahashi and Y. Saito, Zeolites, 8(1988)284.
[21] M. Suzuki, K. Tsutsumi, H. Takahashi and Y. Saito, Zeolites, 8(1988)387.
[22] N. Watanabe, H. Yamashita, H. Miyadera and S. Tominaga, Appl. Catal. B:,
8(1996)405.
[23] S.E. Golunski, H.A. Hatcher, R.R. Rajaram and T.J. Truex, Appl. Catal. B:,
5(1995)367.
[24] G. Lu, J. van Grondelle, B.G. Anderson and R.A. van Santen, J. Catal.,
186(1999)100.
35
NH3 oxidation to nitrogen and water at low temperatures
using transition metal catalysts
36
Chapter 3
Selective low temperature NH3 oxidation to N2 on copper-
based catalysts
Abstract
TPD, TPR, NEXAFS, UV-visible spectroscopy and HREM have been used to
characterize the state and reactivity of NaY and alumina-supported copper-based
catalysts for the oxidation of ammonia to nitrogen. The results of HREM and UV
spectra show that a CuAl2O4 like phase is more active than a CuO phase for the
ammonia oxidation reaction. Both surface oxygen and copper lattice oxygen can react
with NH3 to produce N2 but surface oxygen is much more active than lattice oxygen at
low temperature. For copper zeolite catalysts [Cu-O-Cu]2+-like species or small
copper oxygen aggregates are the likely forms of the catalytically-active centers at
low temperature. The activity of CuY was increased by treating the sample with
NaOH. This treatment presumably increases the amount of low temperature active
centers.
Selective low temperature NH3 oxidation to N2 on copper-based catalysts
1. Introduction
Gas phase ammonia oxidation to nitrogen may find application in the treatment of
ammonia containing industrial flue gases, in the SCR de-NOx process for ammonia
slipstream treatment, in the purification of reformates for fuel-cell systems, in the
deodorization of ammonia/amine containing gas and for the small scale production of
pure nitrogen as a safety gas. Previous studies show that noble metals such as Pt and
Ir are very active for this reaction but are less selective. Significant amounts of nitrous
oxide or nitric oxide are produced on these catalysts [1-3]. Supported molybdenum
and vanadium catalysts show higher selectivity but the reaction temperature needed is
too high for some of the industrial applications [4-6].
Copper oxide is one of the most active catalysts and has been considered as a
potential substitute for noble metal-based emission control catalysts. Sazonova et al.
[7] investigated the activity of various catalysts for ammonia oxidation at
temperatures between 250° and 400°C. They concluded that the most active and
selective catalysts are V/TiO2, Cu/TiO2 and Cu-ZSM-5. Tops∅e of Denmark [8]
recently patented a process for catalytic low temperature oxidation of ammonia in off-
gas at temperatures between 200° and 500oC. The catalysts used were silica-supported
Cu, Co and Ni oxides doped with small amounts of noble metals (100 to 2000ppm).
They claimed that the selectivity and activity of the catalysts was improved when the
catalysts were sulphated either during or prior to contact with the ammonia-containing
gas. Their data show that the selectivity can be improved from 53 % to 99 % by
sulphation. Wollner reported that high ammonia conversion (80-100 %) was obtained
over mixed copper-manganese oxides supported on titania catalysts at temperatures
greater than about 300 oC [9]. However the selectivity was not reported clearly. The
results of Hodnett [10] show that Cu/Al2O3, which is a very active SCR catalyst, is
also very active and selective for ammonia oxidation at a temperature of about 325
o
C. A pulse reaction study of various transition metal-exchanged zeolite Y catalysts
revealed CuY as the most active catalyst for ammonia oxidation at low temperature
[11].
The aforementioned studies show that copper oxides supported on alumina, SiO2,
TiO2 or zeolites are all active ammonia oxidation catalysts. The operating
temperature for these catalysts is still rather high (greater than 300 oC). Little is
known about the relationship between catalyst preparation and the active phase for
these copper-based ammonia oxidation catalysts. In the present study various kinds of
copper oxides supported on alumina and on zeolite Y have been prepared and tested
as catalysts for the ammonia oxidation reaction. TPD, TPR, UV-visible spectroscopy
38
Selective low temperature NH3 oxidation to N2 on copper-based catalysts
and HREM were used to characterize these catalysts in an attempt to shed light on the
optimal preparation for active and selective low-temperature ammonia oxidation
catalysts.
2. Experimental
The starting zeolite was a commercial NaY zeolite from Akzo with a Si/Al ratio of
2.4 (Na56Al56Si136O384.249H2O). NaY was ion-exchanged in an aqueous solution of
copper nitrate at room temperature for 24 hours. After filtration, it was washed with
deionized water three times, dried at 110 oC and then calcined at 400 oC for 2 hours.
The copper contents were 3.7 and 8.4 wt% respectively. The samples will be referred
to as CuY-3.7 and CuY-8.4 in the subsequent discussion.
39
Selective low temperature NH3 oxidation to N2 on copper-based catalysts
wavelength range. The Kubelka-Munk function F(R) was plotted against the
wavelength (in nm).
NH3 and O2 TPD experiments were performed using a fixed-bed flow reactor system
equipped with a computer interfaced quadrupole mass spectrometer. After adsorption
of NH3 or O2 at room temperature the TPD data were recorded by mass spectrometer
while the temperature was increased from 50° to 500 oC at a heating rate of 5 oC/min.
The TPR data were obtained on a typical computer-controlled temperature
programmed reduction apparatus. The hydrogen consumption was measured using a
thermal conductivity detector. The temperature was ramped from 25° to 800 oC at a
rate of 5 oC/min.
In situ NEXAFS (near edge X-ray absorption fine structure) measurements were
performed in BESSY I and BESSY II, the Berlin electron storage ring for synchrotron
radiation. The instrument consists of a two chamber UHV system, of which one
chamber with a base pressure of <5×10-9 mbar was connected to the U49/1-SGM
beamline. The second chamber used as reactor is separated from the first by a 300 nm
polyimide window, withstanding a pressure of at least 100 mbar in the reaction
chamber. The detector system, which is described in detail in refs. [31-34], mainly
consists of an aluminum plate facing the sample (collector) and a shielded gold mesh,
both working in total-electron-yield (TEY) mode. Due to their different shape and
spatial arrangement, the collector signal is a mixture of electrons from the sample and
the gas phase, whereas the gold mesh provides an almost pure gas phase signal.
Therefore the gas phase spectrum must be subtracted from the collector spectrum to
obtain the resonances only from the sample surface. The evaluation procedure for
copper L-edge is less difficult due to a lack of any resonances in the gas phase, so that
only a normalization to the beam intensity and a pre-edge background subtraction is
needed.
3. Results
The results of Table 1 show that the performance of the copper alumina catalysts is
comparable to copper zeolite catalysts at high temperature (>300 oC). At lower
temperature, the copper zeolite catalysts become superior. An optimal copper loading
exists for the copper alumina catalysts. For CuY catalysts, increased loading led to
higher activity.
40
Selective low temperature NH3 oxidation to N2 on copper-based catalysts
Cu-Al-10 250 12 96
300 90 97
350 100 90
Cu-Al-15 250 9 97
300 46 96
350 100 94
CuY-3.7 250 16 97
300 54 98
350 100 97
CuY-8.4 250 23 97
300 88 98
350 100 98
CuY-3.7 200 21 97
(with after-treatment) 250 56 98
300 100 98
CuY-8.4 200 35 95
(with after-treatment) 250 68 97
300 100 98
Reaction condition: NH3 = 1.14%; O2 = 8.21%;
flow rate = 74.7 Nml/min; cat. weight =0.2g
Ione et al.[12] and Schoonheydt et al.[13] reported that the polynuclear nickel or
copper ions were formed in the framework of zeolite Y when NaY was ion exchanged
with an aqueous Ni(NO3)2 or Cu(NO3)2 solution at pH 6-7. These polynuclear cations
gave a higher activity for CO oxidation than the mononuclear state. Suzuki et
al.[14,15] successfully prepared a highly dispersed nickel oxide catalyst, which was
also a very good CO oxidation catalyst, by hydrolysing the Ni2+ exchanged zeolite Y
with an aqueous solution of NaOH or ammonia at various pH values. In this research
the effect of NaOH aftertreatment has also been examined. We have used the
procedure reported by Suzuki et al [14,15] to study the effect of NaOH aftertreatment
on CuY catalysts. The results are included in Table 1.
As can be seen in Table 1, the activity of the two copper zeolite catalysts for ammonia
oxidation increased significantly following NaOH treatment. This significantly higher
activity was probably induced by the formation of small metal oxides in the zeolite. It
should be noted that the CuY catalysts are active at temperatures as low as 200 °C.
Temperatures of at least 250 °C are necessary for reaction on Cu alumina catalysts.
41
Selective low temperature NH3 oxidation to N2 on copper-based catalysts
0.008
2
MS(28) intensity (a.u.)
0.006 1
3
0.004
4
0.002
0
0 100 200 300 400 500 600
Temperature, oC
42
Selective low temperature NH3 oxidation to N2 on copper-based catalysts
0.16
0
0 100 200 300 400 500 600
o
Temperature, C
Fig.2 and Fig.3 compares NH3 and N2 desorption profiles measured during NH3 TPD
experiments on copper zeolite catalysts that had been subjected to NaOH
aftertreatment. The zeolite catalysts were pretreated in O2 at 400 oC for 2 hours prior
to the experiments. The amount of catalyst is same for all the samples used in these
experiments. Only one cycle of NH3 TPD data are shown in Fig.2 and Fig.3 since the
successive cycles measured on these zeolite catalysts showed no N2 production. This
means that either no oxygen existed on the copper zeolite catalysts or that the copper
was completely reduced after the first cycle of the NH3 TPD experiment. The
desorption of NH3 on NaY showed only one peak at 150 oC whereas two peaks (at
150° and 350 oC) were observed on the CuY-8.4 catalyst. The introduction of Cu into
NaY not only creates a more strongly bonded NH3 adsorption center but it also
increases the amount of weakly bonded NH3 centers. After NaOH treatment, a third
ammonia adsorption center seemed to be produced, although this center is not very
strong. There was almost no N2 produced on NaY whilst on CuY-8.4 there were two
N2 peaks at the temperature at which ammonia desorption occurred. This shows that
apparently there were some active ammonia oxidation centers created by the
introduction of Cu into NaY. The first active center (at around 150 oC) is very
interesting as it indicates that ammonia oxidation at low temperature on copper
catalysts is possible. A third peak (at around 190 oC) for ammonia oxidation is created
following NaOH treatment. Comparison of the N2 production profiles measured on
the alumina-based catalyst (Cu-Al-10) and on the zeolite-based catalyst shows that
CuY is apparently much more active than Cu-Al2O3 at low temperature (<200 oC).
43
Selective low temperature NH3 oxidation to N2 on copper-based catalysts
0.012
0.004
NaY
0
0 100 200 300 400 500 600
o
Temperature, C
Catalysts used in O2 TPD and H2 TPR were all treated in O2 at 400 oC for 2 hours and
then cooled to room temperature. The amount of catalyst used for all O2 TPD and H2
TPR experiments is same, so the peak areas of different catalysts can be compared
directly with each other. It can be seen from Fig.4 that two peaks exist in the O2 TPD
profiles on the CuY-8.4 catalyst. Since no O2 desorption was observed on NaY, these
two O2 desorption peaks must be caused by the Cu in the zeolite. After the NaOH
treatment the amount of O2 desorbed was greatly increased. Many
investigations[16,17] on the complete oxidation of CO and hydrocarbons have shown
that their reactivities directly depend on the capacity of oxygen adsorption. Here it
also appears that oxygen adsorption relates to the rate of ammonia oxidation. Fig.5
shows the TPR profiles of the CuY catalyst with and without the aftertreatment.
Integration of the peaks reveals that the amount of reducible copper ions increased
greatly when the CuY catalyst was treated with NaOH.
44
Selective low temperature NH3 oxidation to N2 on copper-based catalysts
6.E-03
CuY-8.4-af.tr.
MS intensity (a.u.)
4.E-03 CuY-8.4
2.E-03
0.E+00
0 100 200 300 400 500 600
Temperature, oC
7
6 C uY-8.4
5 C uY-8.4-af.tr.
TPR signal
4
3
2
1
0
0 200 400 600 800
o
T em p erature, C
45
Selective low temperature NH3 oxidation to N2 on copper-based catalysts
7.5 nm 12 nm
15 nm
15 nm
No metal oxide particles were observed on the alumina surface of the Cu-Al-5 and the
Cu-Al-10 catalysts with TEM (see Fig. 6). Metal particles with diameters ranging
between 5 and 10 nm were observed on the Cu-Al-15 catalyst (see Fig. 7). The Cu-
Al-15 catalyst was black in colour, probably due to the formation of these CuO
particles on the alumina surface. By contrast the catalysts with lower copper loadings
46
Selective low temperature NH3 oxidation to N2 on copper-based catalysts
No differences in the HREM images were observed on samples of the catalyst CuY-
3.7 before and after NaOH treatment. However, a very small amount of particles was
discovered on the outer surface of the zeolite crystals for CuY-8.4 samples (see Fig. 8
and Fig. 9). Again there was no big difference observed before and after NaOH
treatment.
The Cu2+ ion has a 3d9 electronic structure. In the presence of a crystal field generated
by ligands or oxygen ions, d-d transitions appear in the visible or near-infrared range.
For an octahedral environment, transitions appear between 600 and 800 nm depending
on the crystal-field strength. Table 2 shows the summary of our assignments of the
UV spectral bands measured on different kind of catalysts.
47
Selective low temperature NH3 oxidation to N2 on copper-based catalysts
3
Cu-Al-15
Cu-Al-10
Cu-Al-5
2
F(R)
0
200 300 400 500 600 700 800 900 1000
Wavelength, nm
12
CuO(II)
10 Cu2O(I)
8
F(R)
6
4
2
0
100 200 300 400 500 600 700 800 900 1000
Wavelength, nm
It can be seen from Fig.12 that the absorption band appears at about 840 nm for CuY
samples without NaOH aftertreatment. This indicates the existence of isolated Cu2+
ions of square pyramid symmetry [19-21]. After NaOH treatment, this band shifts to
48
Selective low temperature NH3 oxidation to N2 on copper-based catalysts
about 750 nm, similar to the bands observed on the alumina-based copper catalysts
(Cu-Al-5 and Cu-Al-10).
0,5
CuY-8.4-af.tr.
0,45 CuY-3.7-af.tr.
CuY-8.4
0,4 CuY-3.7
0,35
0,3
F(R)
0,25
0,2
0,15
0,1
0,05
0
200 300 400 500 600 700 800 900 1000
Wavelength, nm
12
CuY-3.7-af.tr.
10 CuY-3.7
8
F(R)
6
4
2
200 300 400 500 600 700 800 900 1000
Wavelength, nm
CuY-3.7 samples with and without aftertreatment, were reduced in H2/He at 500 oC
for 2 hours before measuring the UV spectra (see Fig.13). The appearance of an
absorption maximum at 560 nm indicates the creation of a metallic copper phase[21].
The absorption band at 840 nm measured on CuY-3.7 means that some Cu2+ ion can
not be reduced. The intensity of the absorption band at 560 nm (Cuo) did increase at
the expense of 840 nm band (Cu2+) indicating that some further reduction of copper is
possible following NaOH treatment.
49
Selective low temperature NH3 oxidation to N2 on copper-based catalysts
28 Cu L2,3-edge
2+ +
Cu Cu
Intensity, a.u.
24
20
CuY-8.4
16
CuY-8.4
(after treatment)
12
927 932 937
Photon Energy, eV
Fig.14 NEXAFS spectra at the copper L-edge for CuY based catalysts
Both catalyst are pretreated at 400 oC for 2 hours in vacuum and then measured in vacuum
50
Selective low temperature NH3 oxidation to N2 on copper-based catalysts
1,6
Oxidised 1,6
1,4
Reduced 1,4 Oxidised
1,2 Reaction Reduced
1,2 Reaction
1,0
Intensity (a.u.)
1,0
0,8
Intensity (a.u.)
0,8
0,6
0,6
0,4
0,4
0,2
0,2
0,0
0,0
-0,2
-0,2
900 910 920 930 940 950 960
900 910 920 930 940 950 960
Photon energy, eV
Photon energy, eV
Fig.15 In situ NEXAFS Cu L-edge spectra of Fig.16 In situ NEXAFS Cu L-edge spectra of
CuY-8.4 catalyst CuY-8.4-af.tr. catalyst Reaction conditions:
Reaction conditions: O2/NH3=3, T=200 oC, O2/NH3=3, T=200 oC, Pabs=4 mbar
Pabs=4 mbar
Fig.15 and Fig.16 show a plot of a selection of in situ NEXAFS spectra of CuY-8.4
catalysts taken at the copper L-edge at 200 oC. A gas flow ratio O2 /NH3=3 is used.
The signals in these figures are very weak due to the pressure effect and old
equipment used in BESSY I. However, qualitative analysis of these spectra is still
possible. During ammonia oxidation reaction conditions, the copper on both CuY-8.4
catalysts with and without NaOH treatment is partially reduced, most probably in the
state of Cu+.
4. Discussion
51
Selective low temperature NH3 oxidation to N2 on copper-based catalysts
At least three types of copper species in CuY have been described in the literature: 1.
Isolated ions interacting with the framework Al, either without an extraframework
ligand, or with an extraframework O or OH ligand. These species show a different
reactivity [24]. 2. Polymeric chains or multinuclear species (often called small
copper-oxygen aggregates), for instance [Cu-O-Cu]2+ inside the zeolite supercage [25-
27]. 3. CuO particles on the external surface of zeolite crystals [28]. Although at high
loadings the HREM results show some CuO particles on the outer surface of the
zeolite, the UV spectra show no apparent CuO adsorption band intensity at 650 nm.
This means that only a small amount of copper is on the surface of the zeolite crystals
as CuO as compared with the amount of copper inside the pores of zeolite.
In Y-type zeolites X-ray diffraction studies by Gallezot et al. [37] indicated that Cu2+
ions occupy mainly SI and SI’ sites (see Fig.17). This means that at moderate
exchange levels, the Cu2+ ions are unavailable for catalysis, unless they can be
induced to take up positions in the supercages. Although ammonia may readily enter
the solidate cage, molecules such as molecular oxygen are apparently too large to
enter the 2.2 Å aperture at moderate temperatures. Therefore copper centers that are
accessible to ammonia oxidation have to be located in the supercages. According to
earlier studies [29] they are localized at SII sites i.e. near the centers of six-membered
rings of the sodalite cages. The results of NH3 TPD on different zeolite samples show
that only one weakly bonded NH3 adsorption center exists on NaY. The introduction
of Cu into NaY not only creates a more strongly bonded NH3 adsorption center but it
also increases the amount of weakly bonded NH3 centers. Isolated Cu2+ ions are
believed to be responsible for the strongly bonded NH3 adsorption centers since they
can easily react with ammonia forming comparatively stable [Cu(II)(NH3)4]2+
52
Selective low temperature NH3 oxidation to N2 on copper-based catalysts
This may explain the increase of weakly bonded NH3 adsorption centers after
introducing Cu into the NaY. At even higher temperatures the [Cu(II)(NH3)3]2+
complex can be completely decomposed. After NaOH treatment, another weakly
bonded NH3 adsorption centers appeared and the amount of strongly bonded NH3
adsorption centers decreased. This may attribute to the fact that some of the isolated
Cu2+ ion migrated from the sodalite cages to the supercages of the zeolite Y. Some
copper clusters such as [Cu-O-Cu]2+, can also be formed during this process. These
copper clusters might be another type of centers for weakly bonded NH3 adsorption.
The results of O2 TPD and TPR strongly indicate this migration. The UV spectra for
reduced CuY samples also indicate that more Cu2+ ions are reduced to Cu0 after
NaOH treatment. Comparison of the UV spectra for CuY samples with and without
aftertreatment shows that there are mainly isolated copper ions in the zeolite before
NaOH treatment. The shift of the absorption band from 840 nm to 750 nm after
treatment indicates the creation of [Cu-O-Cu]2+ inside the zeolite supercage or the
presence of small copper-oxygen aggregates [25-27]. Perhaps copper ions are first
deposited as Cu(OH)2 by hydroxylation both in the supercages and in the sodalite
cages. Part of the Cu(OH)2 then moves to the SII sites in the supercages during this
hydroxylation process. Upon dehydration of Cu(OH)2, [Cu-O-Cu]2+ complexes are
produced which are thought to be the active centers for low temperature ammonia
oxidation. Actually [Cu-O-Cu]2+ already exists in CuY prior to NaOH treatment.
According to prior studies [30], [Cu-O-Cu]2+ species are present even at low ion-
exchanged levels in CuY. The concentration of [Cu-O-Cu]2+ species increases with
the Cu-loading. This is in agreement with the finding that low temperature active
centers for ammonia oxidation exist on an untreated CuY catalyst.
The results of NEXAFS measurements also indicates that mainly isolated copper ions
exist in CuY before NaOH treatment. These isolated copper ions can be changed into
other copper species or clusters by NaOH treatment. In situ NEXAFS studies on
ammonia oxidation over un-supported copper catalysts showed that mainly N2 was
produced when copper was in the state of Cu+ during reaction. When copper was in
the state of Cu2+ during reaction, the main product was NO or N2O [36]. It has been
shown by our in situ NEXAFS measurements that copper on CuY catalysts is in the
state between Cu0 and Cu2+, most likely Cu+ during ammonia oxidation reaction. This
53
Selective low temperature NH3 oxidation to N2 on copper-based catalysts
5. Conclusions
HREM and UV-visible spectral measurements showed that no metal oxide particles
form on alumina at copper loadings below 10 % (by weight). At higher loadings a
CuO phase was detected. The fact that the performance of 10 wt.% Cu / Al2O3 was
better than 15 wt% (Cu-Al-15) supports the conclusion that a “CuAl2O4”-like surface
phase is more active than a CuO phase in the ammonia oxidation reaction. NH3 TPD
profiles on Cu-Al-10 indicate that both surface oxygen and lattice oxygen can react
with NH3 to produce N2. However, surface oxygen is much more active than lattice
oxygen at low temperature.
NH3 TPD on CuY catalysts shows three types of active centers. Two of these are
active at low temperature (below 200 oC) and one is active at a higher temperature
(above 300 oC). The existence of low temperature active centers indicates that
ammonia oxidation at low temperature on copper catalysts is possible. According to
the UV and NEXAFS spectra, the [Cu-O-Cu]2+-like species or small copper oxygen
aggregates are responsible to the low temperature active centers. However the amount
of low temperature active centers or the concentration of [Cu-O-Cu]2+ species is small
prior to NaOH treatment. The NaOH treatment of CuY increases the amount of low
temperature active centers.
The above results apparently indicate that the environment or the type of active
copper species is very important for low temperature ammonia oxidation and is
strongly related with different supports and preparation methods.
54
Selective low temperature NH3 oxidation to N2 on copper-based catalysts
References
55
Selective low temperature NH3 oxidation to N2 on copper-based catalysts
359(1998).
[29] Breck, D.W., Zeolite Molecular Sieves, A Wiley-Intersci. Pabl., New York-
London-Sydney-Toronto, 1974.
[30] Conesa, J.C. and Soria, J., J. Chem. Soc. Faraday Trans. I, 75, 406(1979).
[31] M. Havecker, A. Knop-Gericke, T. Schedel-Niedrig and R. Schlogl, Angew.
Chem. Int. Ed., 37 (1998) 1939.
[32] A. Knop-Gericke, M. Havecker, T. Neisius and T. Schedel-Niedrig, Nucl. Instr.
and Meth., A 406 (1998) 311.
[33] M. Havecker, A. Knop-Gericke and T. Schedel-Niedrig, Appl. Surf. Sci., 142
(1999) 438.
[34] A. Knop-Gericke, M. Havecker, T. Schedel-Niedrig and R. Schlogl, Topics in
Catalysis, 10 (2000) 187.
[35] J. Stohr., NEXAFS spectroscopy, Springer Series in Surface Science, Vol. 25
(Springer, Berlin, 1992).
[36] R.W. Mayer, M. Havecker, A. Knop-Gericke and R. Schlogl, to be published.
[37] P. Gallezot, Y. Ben Taarit and B. Imelik, J. Catal., 26 (1972) 295.
[38] W.B. Williamson, D.R. Flentge and J.H. Lunsford, J. Catal., 37 (1975) 258.
56
Chapter 4
Intermediate species and reaction pathways for the
oxidation of ammonia on powdered silver catalysts
Abstract
Ammonia oxidation reaction pathways on high surface area silver powder has been
studied by TPD, TPR, FT-Raman and by transient as well as by steady-state ammonia
oxidation experiments. NO was found to be the main reaction intermediate yielding
N2O as well as N2. NO could be formed even at room temperature and when oxidized
to NOx became adsorbed to the silver surface, thus blocking the active sites for
oxygen dissociation. The dissociation of oxygen is thus believed to be the rate-
controlling step for ammonia oxidation. The selectivity for N2, N2O and NO are
mainly determined by the surface oxygen coverage and by reaction temperature. The
adsorbed NOx, N2Ox species are actually inhibitors for ammonia oxidation but these
adsorbed species lower the surface oxygen coverage. Hence the selectivity to nitrogen
is improved with the increasing amount of these adsorbed species.
1. Introduction
The selective catalytic oxidation of ammonia to nitrogen and water (SCO process) has
been proposed as an effective ammonia abatement process [1-3]. Among the catalysts
published in the open literature noble metals such as Pt, Ir are the most active
catalysts at low temperatures. However the selectivity of these metal catalysts is not
satisfying, especially at the condition of very high O2/NH3 ratios. Silver was also
found in our laboratory to be one of the most active ammonia oxidation catalysts but
Intermediate species and reaction pathways for the oxidation
of ammonia on powdered silver catalysts
also showed a poor selectivity to nitrogen, also forming N2O as a main product. In
order to improve the selectivity of these catalysts it is essential to develop a clear
understanding of the reaction pathways for ammonia oxidation on these catalysts.
New insights in the reaction mechanism of ammonia oxidation were obtained from
surface science experiments [10-13]. Single crystals of platinum were used in these
experiments and the reaction was carried out under high vacuum conditions. The first
step of ammonia oxidation was believed to be the formation of NHx species through
oxydehydrogenation of ammonia. Mieher and Ho [11] concluded that NO and N2
were formed by the combination of nitrogen atoms with adsorbed oxygen atoms or
with two nitrogen atoms respectively. However Bradley and King [13] suggested that
NO was directly produced by reacting NHa with Oa. The formation of N2 resulted
from the consecutive dissociation of NO. Since only NO and N2 were produced under
the reaction conditions of these studies, the formation of N2O was not discussed. The
recent study of Van den Broek and van Santen [14,15] showed that NH and OH were
the main surface species adsorbed on Pt and Ir sponge catalysts after ammonia
oxidation. The reaction of NH and OH was thought to be the rate-determining step
under their reaction conditions. They argued that N2O was produced by NO reacting
with adsorbed N. The STM technique was used by the group of M.W. Roberts [16-18]
to study ammonia oxidation on copper surfaces. They observed that at very low
temperatures ammonia could be oxidized to adsorbed NH2 and NH species. At higher
temperature NO formation was also observed. On metal oxide catalysts a mechanism
based on the recombination of two NHx species giving rise to a hydrazinium
intermediate was also proposed [19-21].
58
Intermediate species and reaction pathways for the oxidation
of ammonia on powdered silver catalysts
Ammonia adsorption on Ag(110) has been studied previously by TPD, TPRS and by
EELS with and without coadsorption of molecular and atomic oxygen[22]. It was
concluded that a diffusing nitrogen adatom was the reactive intermediate in NO and
N2 formation. On the basis of a combination of XPS and VEEL spectra a dioxygen-
NH3 complex had been suggested to be a key intermediate in the oxidation of
ammonia on Ag(111) surfaces [35,36]. No report has been published to-date on the
intermediate species and reaction mechanisms on polycrystalline silver at atmospheric
pressure. In this study the ammonia reaction pathways were studied by TPD, TPR,
FT-Raman and by pulsed ammonia reactions on high surface area silver powder at
high O2/NH3 ratios.
2. Experimental
Polycrystalline silver was prepared from Ag2O (>99.99%) powder. The powder was
first reduced in flowing H2/He overnight at 400 oC and then oxidized in flowing
O2/He at 400 oC for 2 hours. This process was repeated three times to remove any
carbon species and to maintain a stable surface area. The surface area of reduced
silver powder determined by krypton BET was 0.67 m2/g with an average particle size
of 0.43 µm. Catalytic activity measurements were carried out in a quartz, fixed-bed
reactor (4 mm internal diameter). The amount of silver powder catalyst was about 0.2
g. Ammonia, oxygen and helium flow rates were controlled by mass flow meters.
59
Intermediate species and reaction pathways for the oxidation
of ammonia on powdered silver catalysts
NH3 TPD and TPR experiments were performed using a fixed-bed flow reactor
system equipped with a computer interfaced quadrupole mass spectrometer. After
adsorption of NH3 at room temperature the TPD or TPR data were monitored with a
mass spectrometer while the temperature was increased from 50 oC to 500 oC at a
heating rate of 10 oC/min. FT-Raman spectra were obtained with a Bruker FRA 106
Raman Spectrometer equipped with an in-situ reaction chamber. The standard laser is
an air-cooled, diode-pumped Nd:YAG operating at 1.064 µm with a maximum laser
power of 500 mW. Usually a laser power of 150 mW was used for measures under
atmospheric conditions. Detection was performed by a liquid-nitrogen-cooled InGaAs
detector. One hundred scans were co-added at a resolution of 2 cm-1.
3. Results
Ammonia oxidation at high O2/NH3 ratios was carried out on silver powder at
temperatures below 400 oC. Fig.1 shows the results of NH3 conversion and products
selectivity versus reaction temperatures. It can be seen that silver is very active for
ammonia oxidation and shows high conversion even below 160 oC. At low ammonia
conversion the N2 selectivity decreases with increasing temperature. At high ammonia
conversion the N2 selectivity increases with increasing temperature. When the
temperature is over 300 oC the N2 selectivity decreases again because of the formation
of NO.
120
100 Conv.
Conv. or Sel., %
80
Sel. N 2
60
40
Sel. NO x
20
Sel. N 2O
0
50 150 250 350 450
Temperature, oC
60
Intermediate species and reaction pathways for the oxidation
of ammonia on powdered silver catalysts
3.2 TPD and TPR of ammonia on reduced and on oxygen-covered silver poeder
Fig.2 shows the NH3 TPD profile measured on completely reduced silver powder.
Apparently a small amount of water in the helium or in the ammonia can be adsorbed
on Ag and is desorbed at about 150 oC. Trace ammonia also desorbs at this
temperature. This is mostly from water-coupled ammonia. This result shows that a
clean silver surface can not adsorb ammonia above room temperature. This is in
accordance with literature [22]. Fig.3 shows the NH3 TPD profile measured on
oxygen pre-covered silver powder. There are two water production peaks at 150 oC
and 280 oC respectively. The peak at 150 oC is similar to the water desorption peak
shown in Fig.2. The peak at 280 oC can be caused by the reaction of adsorbed OH
species according to the following reactions:
16.0
MS Signal (a.u.)
12.0
H2O
8.0
4.0
NH3
0.0
0 50 100 150 200 250 300 350 400 450
Temperature,oC
61
Intermediate species and reaction pathways for the oxidation
of ammonia on powdered silver catalysts
5.0E-09
3.0E-09
H2O
2.0E-09 N2
NH3
1.0E-09
O2
NO
N2O
0.0E+00
0 50 100 150 200 250 300 350 400 450
o
Temperature, C
The oxygen peak appears at the same position as the water peak. This strongly
suggests that reaction step (1) is rate-limiting. There is still a trace of NH3 desorbed at
150 oC. Significantly more N2 is produced at a temperature of 120 oC. This N2
production must come from adsorbed nitrogenous species on the silver surface. On
Ag(111) surfaces a dioxygen-NH3 complex has been detected by XPS and EELS at -
193 oC which is suggested to be a key intermediate in the oxidation of ammonia [35].
When temperature is increased to –63 oC the di-oxygen are not present and NH
species appear. A study on Ag(110) surfaces for ammonia oxidation shows that NH
groups persist on the surface at temperature corresponding with the water desorption
peak at 37 oC and possibly at significantly higher temperatures [22]. Thus the
adsorbed nitrogen species may be NHx(a) or NO(a) produced by oxygen abstraction
of hydrogen from NH3:
62
Intermediate species and reaction pathways for the oxidation
of ammonia on powdered silver catalysts
Besides N2 there is also some N2O produced. The maximum production temperature
for N2O production is higher than that of N2. This indicates that the selectivity of
ammonia oxidation depends on the surface N/O ratio. With the production of N2 the
amount of NHx decreases and the O/N ratio increases, leading to an increased N2O
production and the observed maximum.
4.5E-09
MS Signal (a.u.)
N2
3.0E-09
1.5E-09 H 2O
O2
NO
N 2O
0.0E+00
0 50 100 150 200 250 300 350 400 450
o
Temperature, C
Fig.4 shows the NH3 temperature programmed reaction (TPR) profiles on an oxygen
pre-covered Ag powder. NH3/He was flowing constantly over the silver powder
during the increase of temperature. With increasing temperature the rate of reaction
(3) and (4) are greatly increased and the surface N/O ratio increases accordingly. So
over the entire temperature range the only product is nitrogen. There are four N2
production maxima at 120 oC, 170 oC, 240 oC and 330 oC respectively. The first one is
apparently caused by adsorbed surface atomic oxygen. The later two appear
coincidentally after water production has reached a minimum. It is possible that water
inhibits the migration of subsurface oxygen onto the surface. When water desorbs,
more subsurface oxygen may become accessible for the ammonia oxidation reaction.
It also can be seen from the water production profile that a new peak at about 200 oC
appears in addition to the two already mentioned. This peak surely is not the result of
reaction (1). It may originate from the following reactions:
63
Intermediate species and reaction pathways for the oxidation
of ammonia on powdered silver catalysts
The possibility of reactions (7), (8), (9) cannot be excluded at low temperature but at
least one of these reactions, most probably reaction (9), will probably occur only at
higher temperatures (>150 oC) since NH(a) is most stable adsorbed species compared
with NH2(a) and NH3. There is also a new peak at about 330 oC in the H2O profile
which is in the nearly same position as the fourth N2 peak in the N2 profile. This may
be caused from the ammonia oxidation by bulk dissolved oxygen which is migrated
on to the silver surface at high temperature.
In order to further clarify the reaction mechanism pulse reactions were carried out at
different temperatures. Pulses of NH3 (circa 0.22 :mol) were made into a flow of
O2(10 vol%)/He. It can be seen from Fig.5 that during the first ammonia pulse large
amounts of N2O were produced simultaneously with N2 production. During the
following ammonia pulse the amount of N2O production was greatly reduced while
N2 production increased. When the reaction temperature was raised from 200 oC to
400 oC the situation changed completely. Large amounts of NO were produced during
each NH3 pulse but the N2 selectivity changed little (see Fig.6). These results indicate
that some adsorbed nitrogen species were produced during the first NH3 pulse that
greatly changed the reaction selectivity of the following NH3 pulse reaction at 200 oC
but had less effect on selectivity at 400 oC.
64
Intermediate species and reaction pathways for the oxidation
of ammonia on powdered silver catalysts
6.0E-04
MS Signal (a.u.)
N2
4.0E-04
N2O
2.0E-04
NO
0.0E+00
0 10 20 30 40 50
Time, min
o
Fig.6 NH3 pulse reaction profiles on silver powder at 400 C
(Reaction conditions were the same as in Fig.5)
In order to clarify the nature of the adsorbed species TPD experiments were done on a
silver powder catalyst after reaction under different conditions. Fig.7 shows the TPD
profiles measured on a Ag powder catalyst after reaction for 1 h in excess oxygen at
200 oC. It can be seen that the production profile patterns for N2O, N2 and O2 are
almost the same. This may indicate that these products originate from the same
species. It is well known that NO can react to give N2O, N2 and O2 according to the
following reactions [23,24]:
65
Intermediate species and reaction pathways for the oxidation
of ammonia on powdered silver catalysts
6.0E-04
H2O
MS Signal (a.u.)
4.0E-04
N2O
N2
2.0E-04
NH3 O2
NO
0.0E+00
200 300 400 500 600
o
Temperature, C
Fig.7 TPD profiles after NH3 oxidation reaction at 200 oC on silver powder
(Reaction conditions: NH3=1000 ppm; O2=10%; Flow rate=50 Nml/min;
Cat. Weight=0.1 g. TPD was performed in flowing He=20 Nml/min from
200 to 550 oC at 10 oC/min)
It also can be seen from Fig.7 that there is some water produced. This may indicate
the presence of some OH or NHx species but the quantity is small. The fact that there
is a trend to desorb N2, NO, N2O and O2 at higher temperature (>500 oC) suggests the
existence of more than one NO induced species. This species may be NO3- since it is
the most stable species on the silver surface.
When the reaction was carried out at low O2/NH3 ratios (eg. 1/1) the TPD profile after
reaction was totally different (see Fig.9). Much more water was produced and only N2
desorbs at low temperature. As most of the OH species can be desorbed as H2O at 280
o
C, according to reaction (1), the comparatively large amount of water must come
from NHx species. Most probably N2 and H2O result from the following reaction:
66
Intermediate species and reaction pathways for the oxidation
of ammonia on powdered silver catalysts
Thus, apparently there are more NHx species on the surface when the O2/NH3 ratio
decreases.
1.0E-03
MS Signal (a.u.)
O2
7.5E-04
5.0E-04
N2O
2.5E-04 N2
NO
0.0E+00
200 300 400 500 600
Temperature, oC
1.0E-03
H2O
MS Signal (a.u.)
7.5E-04
5.0E-04
N2
2.5E-04 NO
O2 NH3
N2O
0.0E+00
200 300 400 500 600
Temperature, oC
67
Intermediate species and reaction pathways for the oxidation
of ammonia on powdered silver catalysts
FT-Raman measurements were carried out on a silver powder with pre-adsorbed NO,
NO2 or NH3. For NO or NO2 adsorption a completely reduced silver powder was used
and the adsorption temperature was 50 oC. For NH3 adsorption the reduced silver
powder was first oxidized in O2 at 200 oC for one hour and then NH3 was adsorbed at
100 oC to increase the Raman signal. All of the measurements were done at room
temperature to avoid heat effects. The resulting spectra are shown in Fig.10. Bao et al.
have studied NO adsorption on Ag(110) by Raman and XPS [25]. They concluded
that NO was not stable on a silver surface. NO was quickly oxidized to NO2- and NO3-
which were strongly adsorbed on silver surface. SERS (Surface Enhanced Raman
Spectroscopy) spectra of NO2 adsorbed on silver powder were also reported in the
literature [26,27]. Peaks at 815, 1045 and 1285 cm-1 after exposure of silver powder to
several nitrogen dioxide pulses added into the He gas stream were observed. They
assigned the peaks at 815 and 1285 cm-1 to adsorbed NO2- and the peak at 1045 cm-1
to adsorbed NO3- on silver surface. Comparing the Raman spectrum of NH3
adsorption with NO and NO2 adsorption spectra in Fig.10, it is obvious that a similar
spectrum to that of NO, NO2 adsorption appears upon adsorption of NH3 on oxygen-
covered silver powder. This strongly indicates that NH3 can be quickly oxidized to
NO which will consequently be oxidized to NO2- and NO3- species. The remaining
peaks in Fig.10, such as in the positions of 615 cm-1, 670 cm-1 and 985 cm-1, are
assigned to the different oxygen species on the silver surface [29].
68
Intermediate species and reaction pathways for the oxidation
of ammonia on powdered silver catalysts
0.8
NO 2
NO 2
Intensity (a.u.)
0.6 NO 2
NO 3
0.4
NO
0.2 NH 3
0
500 1000 1500 2000
-1
Wavenumber, cm
Fig.10 Raman spectra for NO, NO2 and NH3 adsorbed on silver powder
(NO and NO2 adsorption at 50 oC on reduced silver powder; NH3 adsorption at 50 oC on oxygen-
covered silver powder)
1
N 2O 3 or N 2O 5
O 25 oC
0.8 100 oC
Intensity (a.u.)
200 oC
0.6
350 oC
0.4
250 oC
0.2
0
0 500 1000 1500 2000 2500
-1
Wavenumber, cm
Fig.11 Raman spectra measured at room temperature after ammonia oxidation on silver powder at
different temperatures
( Reaction conditions: NH3=1000 ppm O2=10 %; NH3/O2/He=50 Nml/min)
69
Intermediate species and reaction pathways for the oxidation
of ammonia on powdered silver catalysts
N2Ox [30-33], the peak at 245 cm-1 is assigned to surface adsorbed atomic oxygen and
the peaks at 266 and 1540 cm-1 are assigned to the N-N stretching Raman band of
N2Ox. The x can be 2, 3, 4 or 5. Unambiguous determination of x is difficult since
these N2Ox species often show intensities at nearly the same positions in Raman and
IR spectra. It can be seen from Fig.11 that the intensity of surface atomic oxygen
decreases with increasing temperature. The intensity of the N2Ox peaks increased
slightly as heating began but stayed unchanged afterwards. The NO2 peak intensity
decreased but NO3- adsorption signal stayed unchanged as the temperature increased.
When the temperature was increased to 300 oC the signals of N2Ox disappeared and
the surface atomic oxygen adsorption signal recovered again.
The reactions between NOx and NH3 were also tested on powdered silver catalyst to
prober the posssibilities of different reactions between these reactants.
80 250
Concentration, ppm
Concentration, ppm
NO 200
60
NH3
150
40
100
NO
20
50
N2 O
N2O NO2
0 0
0 200 400 600 0 100 200 300 400
o
Temperature, o
C Temperature, C
Fig.12 NO+H2 reaction on reduced silver powder Fig.13 NO+NH3 reaction on reduced silver
(NO=74 ppm, H2=5%, NO/H2/He=100 Nml/min (NH3=194 ppm, NO=74 ppm, Cat. Weight=
Cat. Weight=0.1 g) 0.1 g, NH3/NO/He=100 Nml/min)
It can be seen from Fig.12 that under steady state reaction conditions the NO cannot
be reduced by H2. This indicates that N2 and N2O cannot be directly formed from NO
alone on silver powder catalysts. Fig.13 shows the results of reaction between NO and
NH3. Only a small fraction of NO can react with NH3 directly to produce N2 and N2O
on silver powder catalysts. It can be from Fig.14 that the SCR reaction of NO with
NH3 is greatly increased by introducing O2. Fig.15 shows the results of reaction
between NO2 and NH3 on silver powder catalyst. The NO2 was first decomposed into
NO and then reacted with NH3 to produce N2 and N2O. At higher temperatures the
NO concentration increses again because of the direct ammonia oxidation. This
70
Intermediate species and reaction pathways for the oxidation
of ammonia on powdered silver catalysts
experiment clearly indicates that NO2 is not a reaction intermediate species to produce
N2O on silver catalysts.
180 200
NH3
160
135
Concentration, ppm
Concentration, ppm
120 NH3
90 NO N2O NO2
80
45 NO
40
NO2 N2O
0 0
0 100 200 300 400 0 100 200 300 400
o Temperature, oC
Temperature, C
Fig.14 NH3+NO+O2 reaction on silver powder Fig.15 NO2+NH3 reaction on silver powder
(NO=74 ppm, NO/NH3/O2/He=100 Nml/min (NH3=194 ppm, NO2=70 ppm, Cat. Weight=
NH3=168 ppm, Cat. Weight=0.1 g) 0.1 g, NH3/NO2/He=100 Nml/min)
4. Discussion
NH3 TPD and TPR results show that NH3 does not adsorb on clean silver surface
above room temperature. On an oxygen pre-covered silver surface NH3 dissociatively
adsorbs only as NHx with one or two hydrogen abstracted by oxygen according to
reaction (3) and (4). This is considered to be the initial step for ammonia oxidation.
Reaction (3) and (4) can be enhanced by either increasing the temperature or by
increasing the oxygen coverage on silver surface [17,18].
NHx species can subsequently react with either O or OH according to reaction (9) to
give N:
71
Intermediate species and reaction pathways for the oxidation
of ammonia on powdered silver catalysts
When there is pre-adsorbed oxygen on the silver surface the probability of reaction
(16) occurring is low due to the high rate of reaction (15). The Raman experimental
results clearly showed that NOx species appeared even at room temperature upon
ammonia adsorption. There was no evidence for the formation of molecular nitrogen
at room temperature on silver powder. The fact that reactions (14) and (15) can take
place at room temperature indicates that reaction (14) is not the rate-controlling step
for gas phase NO formation as suggested by many authors [11,15,21]. Actually the
NO formed can not desorb at low temperatures and quickly reacts with other adsorbed
oxygen to produce NOx which will strongly adsorb on the silver surface and block the
active sites for oxygen dissociation. Since reaction (16) is not significant at high
O2/NH3 ratio conditions, the formation of nitrogen may occur as follows:
The reaction (18) actually involves the NO dissociation and atomic nitrogen
combination as in reaction (16). This occurs only at low temperatures on reduced
silver and oxygen is a strong inhibitor of this reaction.
The mechanism of N2O formation from ammonia is not very clear. In low pressure
ammonia oxidation studies on Pt and Ag single crystals no N2O formation was found
over a large temperature interval and a large range of NH3/O2 ratios[13,22]. On Pt and
Ag polycrystals, however, large amounts of N2O are produced at low temperatures. At
higher temperatures NO, instead of N2O, is produced. The possible N2O formation
pathways proposed in the literature are as follows [4,15,23,34]:
Reaction (21) and (22) involve the intermediate HNO, HNO3 species which were not
observed on the silver surface. Thus, these two reactions are unlikely to occur on
silver catalysts. The difference between reaction (19) and (20) is that the dissociation
of NO is required in order that reaction (20) can proceed. This can only happen at
72
Intermediate species and reaction pathways for the oxidation
of ammonia on powdered silver catalysts
very low surface oxygen coverage. The chemisorption of nitric oxide at a Ag(111)
surface has been reported to form N2O at 80 K. However at 295 K this surface is
unreactive to NO(g) under the same condition [37]. Our experiments on silver powder
also show that N2O can be produced only on reduced silver at initial stage of NO
adsorption. When NO is introduced continuously to a reduced silver powder catalyst
at temperature from 50-400 oC no N2O is produced except for the initial time. Hence
reaction (19) is the most likely reaction candidate for the formation of N2O. To give
further evidence for this theory, the SCR reaction was also carried out on reduced
silver powder. The results show that the production of N2O is not significant when
NO reacts with NH3 alone on silver catalyst. In the presence of oxygen NO can react
quickly with NH3 to give rise to N2O and N2. These experiments clearly indicate that
N2 and N2O are not formed mainly from NO alone. NO cannot react directly with
NH3 on silver either. The role of O2 is actually to produce adsorbed NHx and N
species which can then react readily with NO to produce N2 and N2O.
Apparently NO2-, NO3- and N2Ox species are produced by NO reacted with O or O2:
N2
[NHx]
[O] [O] [N]
NH3 NHx NO N2O
[O]
73
Intermediate species and reaction pathways for the oxidation
of ammonia on powdered silver catalysts
It can be seen clearly from this scheme that NO is the key intermediate. Even at room
temperature NO can be quickly produced. Since NO cannot desorb at low
temperature, it blocks the active sites for oxygen dissociation. So the reaction can
continue only when the adsorbed NO is removed. At moderate temperatures (below
300 oC) NO can only be removed either as N2O or N2 through surface reaction (17)
and reaction (15) respectively. At even higher temperature NO can directly desorb as
one of the products. The rate-controlling step is thus the oxygen dissociation rate
which is determined by the rate of NO removal. The selectivities to N2 and to N2O are
also controlled by the relative rates of reaction (17) and (19). Actually, the relative
rates of these two reactions are controlled by the surface oxygen coverage. At high
surface oxygen coverage the reaction from NHx to NO and N is very fast so that the
surface NHx concentration will decrease and the rate of reaction (17) will decrease
accordingly. The selectivity is thus mainly towards N2O. On the contrary the reaction
becomes more selective to nitrogen when the surface oxygen coverage decreases.
As the surface NO2, NO3- and N2Ox species block some of the active sites for oxygen
dissociation, they become inhibitors for ammonia oxidation. However the adsorption
of these species on silver surface decreases the oxygen coverage. So the selectivity to
nitrogen increases with increasing amounts of these adsorbed species. The Raman
spectra for ammonia oxidation clearly showed that atomic surface oxygen intensity
decreased with increasing temperature from 50 oC up to 300 oC. This is due to the fact
that the amount of adsorbed N2Ox species increases with increasing temperature. At a
temperature higher than 350 oC N2Ox decomposes and is removed from the surface so
the oxygen coverage again increases. The steady-state tests for ammonia oxidation on
silver powder at different temperatures (Fig.1) also show that the nitrogen selectivity
increases with increasing temperature up to 300 oC. With further increase of
temperature, the selectivity to nitrogen decreases due to the increasing rate of NO
desorption. Ammonia pulse reactions again indicated the effect of adsorbed NOx,
N2Ox species on the product selectivity. After the first ammonia pulse at 200 oC some
adsorbed NOx, N2Ox species were produced which partially block the surface and
decreased the surface oxygen coverage. In subsequent ammonia pulse reactions the
selectivity to nitrogen was improved due to the lower surface O/N ratio. At higher
temperature (400 oC) the desorption rate of NO became high, so that NO is formed
instead of N2O. Since little NOx can be adsorbed on silver surface at this temperature
there is almost no selectivity difference among three consecutive ammonia pulses.
74
Intermediate species and reaction pathways for the oxidation
of ammonia on powdered silver catalysts
5. Conclusions
Silver is a very active catalyst for ammonia oxidation. At low temperature (below 300
o
C) mainly N2 and N2O are produced. At higher temperature NO instead of N2O
becomes one of the products. NO is an important reaction intermediate for this
reaction. Even at room temperature NO can be produced and adsorbed on silver
surface in the form of NOx. Since NO can not desorb at low temperature it blocks the
active sites for oxygen dissociation. The dissociation of oxygen is thus believed to be
the rate-controlling step for ammonia oxidation. The selectivity to N2, N2O and NO is
determined by surface oxygen coverage and temperature. Low surface oxygen
coverage favors nitrogen formation. Adsorbed NOx, N2Ox species are actually
inhibitors for ammonia oxidation but they also lower the surface oxygen coverage.
Hence, the selectivity to nitrogen is improved by increasing the amount of these
adsorbed species on silver surface.
Generally speaking, silver alone, like Pt, is not a good catalyst for selective
ammonia oxidation to nitrogen because too much N2O is produced. It follows from
the above conclusions that blocking of the sites for oxygen dissociation is an effective
way to improve the nitrogen selectivity, but also would result in a loss of catalyst
activity.
75
Intermediate species and reaction pathways for the oxidation
of ammonia on powdered silver catalysts
References
76
Intermediate species and reaction pathways for the oxidation
of ammonia on powdered silver catalysts
26. Von Raben, K.U., Dorain, P.B., Chen, T.T., and Chang, R.K., Chem. Phys. Letters
95, 269(1983).
27. Matsuta, H., and Hirokawa, K., Surf. Sci. 172, L555(1986).
28. Mcbreen, P.H., and Moskovits, M., J. Catal. 103, 188(1987).
29. Stencel, J.M., “Raman Spectroscopy for Catalysis”, Van Nostrand Reinhold, New
York, 1990.
30. Hisatsune, I.C., Devlin, J.P., and Wada, Y., J. Chem. Phys. 33(3), 714(1960).
31. Hisatsune, I.C., Devlin, J.P., and Wada, Y., Spectrochimica Acta 18, 1641(1962).
32. Kugler, E.L., Kadet, A.B., and Gryder, J.W., J. Catal. 41, 72(1976).
33. Bibart, C.H., and Ewing, E., J. Chem. Phys. 61(4), 1293(1974).
34. Andrussow, L., Z. Elektroch. 36, 756(1930).
35. Carley, A.F., Davies, P.R., Roberts, M.W., Thomas, K.K. and Yan, S., Chem.
Commun., 35(1998).
36. Carley, A.F., Davies, P.R. and Roberts, M.W., Current Option in Solid State
and Materials Science, 2(5), 525(1997).
37. Carley, A.F., Davies, P.R., Roberts, M.W., Santra, A.K. and Thomas, K.K.,
Surface Science, 406, L587(1998).
77
Intermediate species and reaction pathways for the oxidation
of ammonia on powdered silver catalysts
78
Chapter 5
Low Temperature Selective Oxidation of Ammonia to
Nitrogen on Silver-based Catalysts
Abstract
The low temperature gas phase oxidation of ammonia to nitrogen has been studied
over alumina-supported, silica-supported and unsupported silver catalysts. TPD,
TPR, TEM, XRD and FT-Raman were used to characterize the different silver
catalysts. The results showed alumina-supported silver to be the best catalyst due to
the interaction of silver with alumina. Pretreatment had a great effect on the catalyst
performance. Reduction in hydrogen at 200 oC without any pre-calcination gave the
best activity while reduction at higher temperatures showed little difference from
calcination pretreatment. At least four types of oxygen species were produced when
silver was oxidized at high temperature. These species are adsorbed molecular
oxygen, adsorbed atomic oxygen, strongly adsorbed atomic nitrogen and subsurface
oxygen respectively. Ammonia oxidation activity at low temperature is related to the
catalyst’s ability to adsorb oxygen either dissociatively or non-dissociatively. In
addition, a good correlation existed between the N2 selectivity for the SCO reaction
and the SCR performance of NO with NH3 for the silver-based catalysts, i.e., the
higher SCR yield of nitrogen, the higher the SCO selectivity to N2.
Low temperature selective oxidation of NH3 to N2 on silver-based catalysts
1. Introduction
2. Experimental
Catalytic activity measurements were carried out in a tubular quartz, fixed-bed reactor
(4 mm internal diameter). The amount of catalyst used was about 0.2 gram (250-425
µm particles). Ammonia, oxygen and helium flow rates were controlled by mass flow
controllers. The inlet and outlet gas compositions were analyzed by an on-line
Chemiluminescence NH3 and N2O Analyzer. A quadruple mass spectrometer was also
used to monitor the different products (Balzers OmniStar).
80
Low temperature selective oxidation of NH3 to N2 on silver-based catalysts
NH3 TPD or TPR experiments were performed using a fixed-bed flow reactor system
equipped with a computer-interfaced quadrupole mass spectrometer. After adsorption
of NH3 at room temperature the TPD or TPR data were recorded by mass
spectrometer while the temperature was increased from 50° to 500 oC at a heating rate
of 5 oC/min. The H2 TPR and O2 TPD data were obtained on a typical computer-
controlled temperature programmed reduction apparatus. The hydrogen consumption
was measured using a thermal conductivity detector. The temperature was ramped
from 25° to 800 oC at a rate of 10 oC/min.
FT-Raman spectra were obtained with a Bruker FRA 106 Raman Spectrometer
equipped with an in-situ reaction chamber. The standard laser is an air-cooled, diode-
pumped Nd:YAG operating at 1.064 µm with a maximum laser power of 500 mW.
Usually a laser power of 150 mW was used for measurements at atmospheric
pressure. Detection was performed by a liquid-nitrogen-cooled InGaAs detector. One
hundred scans were co-added at a resolution of 2 cm-1. High resolution transmission
electron microscopy (HREM) measurements were done on a Philips CM30-T electron
microscope at the National Centre for HREM at the Delft University of Technology.
X-ray powder diffraction (XRD) patterns were recorded on a Rigaku Geigerflex X-
ray powder diffractometer using Cu-Kα radiation. Prior to the experiment the
catalysts were ground and pressed into a sample holder containing vaseline. The
applied scanning speed was 1o per minute. Background subtraction was not applied.
81
Low temperature selective oxidation of NH3 to N2 on silver-based catalysts
3. Results
120
100 Conv.
Conv. or Sel., %
80
N 2 Sel.
60
40
NO x Sel.
20
N 2 O Sel.
0
100 200 300 400 500
o
Temperature, C
120
100 C o n v.
Conv. or Sel., %
80
S e l. N 2
60
40
S e l. N O x
20
S e l. N 2 O
0
100 200 300 400 500
o
T e m p e ra tu re , C
In order to elucidate the support effect on ammonia oxidation for the silver system,
pure silver powder, silica- and alumina-supported catalysts were prepared and tested
82
Low temperature selective oxidation of NH3 to N2 on silver-based catalysts
for ammonia oxidation at different temperatures. The results are shown in Fig.1, Fig.2
and Fig.3. It can be seen that pure silver powder catalyst behaves similarly to silica-
supported silver catalyst but quite differently from the alumina-supported silver
catalyst. The alumina-supported silver catalyst shows better selectivity to N2 than
silver powder or the silica-supported silver catalyst. It is interesting that mainly N2O
is produced on silver powder and silica-supported catalysts at lower temperatures. The
selectivity at low temperature (<300 oC) varies little for alumina-supported silver
catalyst but decreases abruptly at higher temperature due to the large production of
NO (see Fig. 3). Different catalyst pretreatment has a large effect on alumina-
supported silver catalysts (see Fig.4). Calcination at high temperature decreases the
activity of catalyst. This probably relates to a decrease in surface area at elevated
temperature. Reduction in hydrogen at 200 oC without any pre-calcination gives best
activity however reduction at higher temperatures shows little difference from the
calcined sample.
120
Conv.
100
Conv. or Sel., %
80
Sel. N 2
60
40
Sel. NO x
20
Sel. N 2 O
0
100 200 300 400 500
o
Temperature, C
83
Low temperature selective oxidation of NH3 to N2 on silver-based catalysts
The above results clearly show that silver-based catalysts are quite active for
ammonia oxidation. Alumina-supported silver catalysts are the most selective
compared with silver powder and silica-supported catalysts. However the N2
selectivity is still not very satisfying (around 80%).
120
Fig. 4 The effect of
100 pretreatment for 10
wt% Ag/Al2O3 catalyst
NH3 Conv., %
80 Reaction conditions:
NH3=1000 ppm;
60 O2=10 vol%;
Total flow rate=50
calc. 500 oC Nml/min;
40 calc. 800 oC
red. 200 oC Cat. weight=0.1 g
20 red. 500 oC
0
100 150 200 250
o
Temperature, C
84
Low temperature selective oxidation of NH3 to N2 on silver-based catalysts
Table 1 shows the performance of the Ag/Al2O3 catalyst compared with the noble
metal catalysts. It can be seen that the performance of silver-based catalyst is superior
to that of noble metal catalysts in both activity as well as selectivity. With further
improvement, this catalyst is promising and could potentially be used instead of noble
metal-based catalysts at the temperatures below 200 oC.
600
---- silver metal
500 ---- silver oxide
Fig. 5 XRD spectra
400 of supported silver
10%Ag/SiO catalysts calcinated
at 500 oC
2
CPS
300
200 5%Ag/Al 2 O3
10%Ag/Al 2 O3
100
15%Ag/Al 2 O3
0
20 40 60 80 100
2 Theta
XRD and TEM were used to examine the differences between silica- and alumina-
supported catalysts. All the catalysts were calcined in air at 500 oC for 24 hours before
testing. Fig.5 shows the XRD spectra of the different silica- and alumina-supported
catalysts. A distinct XRD pattern attributable to crystallized silver metal was observed
for 10 wt% Ag/SiO2 catalyst. However on 5 wt% Ag/Al2O3 catalyst no silver metal
phase was detected but specific XRD pattern for Ag2O appeared. As the silver loading
on alumina increased the metallic phase increased and the silver oxide phase
disappeared except for the peak at diffraction angle of 33.43 degree.
85
Low temperature selective oxidation of NH3 to N2 on silver-based catalysts
18 nm
10% Ag/Al2O3
33 nm
10% Ag/SiO2
Fig.6 shows the TEM photographs of different supported catalysts to demonstrate the
distribution of silver particles on the surface of different supports. As shown by Fig.6,
the silver was inhomogeneously distributed on both silica and alumina support. The
metal particle size distribution on silica-supported silver catalyst seems to be bimodal.
Many particles are about 15-40 nm in diameter, whereas many particles of about 1-2
nm diameter are also present. The silver particle size distribution on alumina is very
broad. The smallest particles are about 3 nm in diameter and the largest particles
found are about 40 nm in diameter. Table 2 shows the silver area and particle size of
three silver-based catalysts determined by different techniques.
86
Low temperature selective oxidation of NH3 to N2 on silver-based catalysts
0.6
O O2-
2-
O2 500 oC
0.5
Intensity(a.u.)
200 oC
0.4
50 oC
0.3
0.2
0.1
0 500 1000 1500 2000
-1
Wavenumber, cm
87
Low temperature selective oxidation of NH3 to N2 on silver-based catalysts
2.0E-11
1.6E-11
MS(32) Signal (a.u.)
1.2E-11
8.0E-12
a
4.0E-12
b
0.0E+00
0 200 400 600 800 1000
o
Temperature, C
1.6 3
TPR Signal (a.u.)
5
1.2 4
2
0.8
0.4 1
0
0 100 200 300 400 500
o
Temperature, C
88
Low temperature selective oxidation of NH3 to N2 on silver-based catalysts
cooled down to room temperature in O2. The O2 TPD profiles on 10wt% Ag/SiO2 (see
Fig.8) show that two O2 desorption peaks appear for a catalyst pretreated with oxygen
at 200 oC. One is at the temperature of about 280 oC and another at temperature over
800 oC. When the catalyst was pretreated in oxygen at 500 oC, four O2 desorption
peaks were observed at temperatures of 280, 450, 600 and 775 oC respectively. It is
well known that atomically adsorbed oxygen desorbs at around 280 oC on a silver
surface [19]. High temperature treatment in oxygen decreased the intensity of the
atomically adsorbed oxygen peak substantially. This is caused by silver surface
reconstruction as argued by many researchers [20-22,30].
The O2 TPD results above clearly show that pretreatment conditions are extremely
important to the silver oxygen interaction. Hence H2 TPR was carried out on silver
catalysts pretreated in oxygen at different temperatures. The catalysts were first
reduced at 500 oC for 2 hours before oxygen pretreatment. Fig.9 shows the H2 TPR
profiles on 10 wt% Ag/SiO2. It can be seen that three peaks exist at temperatures of
80, 97 and 460 oC respectively when the catalyst is pretreated in oxygen at a
temperature below 200 oC. When the pretreatment temperature increased to over 200
o
C a fourth peak appeared. The position of this peak along with the second peak
moved to higher temperatures with the increase of pretreatment temperature.
As H2 cannot directly go into the silver subsurface or bulk without reacting with the
surface oxygen, the first TPR peak must be surface adsorbed molecular or atomic
oxygen. As it is well known that adsorbed atomic oxygen can reach a maximum at
about 200 oC on silver surface, the second peak can thus be assigned to surface
adsorbed atomic oxygen and the first peak can therefore be attributed to molecularly
adsorbed oxygen. Compared with the Raman results the peak appeared after high
temperature treatment in O2 can be assigned to atomic oxygen strongly adsorbed on
the silver surface defects as discovered by L. Lefferts [30]. The peak at 460 oC thus
can be assigned to the subsurface oxygen since this oxygen is the most difficult one to
be reduced [23,30].
Fig.10 shows the H2 TPR profiles on 10 wt% Ag/Al2O3. It can be seen that there are
also three peaks at temperatures of 75, 130 and 470 oC that can be assigned to
molecularly adsorbed oxygen, adsorbed atomic oxygen and subsurface oxygen
respectively when catalyst is pretreated in oxygen at the temperature below 200 oC.
When the pretreatment temperature increased to over 200 oC, again strongly adsorbed
oxygen peak appeared. The position of this peak moved to higher temperatures with
the increase of pretreatment temperature.
89
Low temperature selective oxidation of NH3 to N2 on silver-based catalysts
2 .8
2 .4 5
TPR Signal (a.u.)
2
4
1 .6 3
1 .2
2
0 .8
0 .4 1
0
0 100 200 300 400 500
o
T e m p e ra tu re , C
For 5 wt% Ag/Al2O3 catalyst the TPR profile pattern was almost the same as that of
10 wt% Ag/Al2O3 but the second peak moved to higher temperatures when the
catalyst was treated in oxygen at higher temperature (see Fig.11). Apparently catalysts
with low silver loading are even more sensitive to oxygen pretreatment.
1
Fig.11 H2 TPR profiles
3 on 5 wt% Ag/Al2O3
0.8
catalyst
TPR Signal (a.u.)
4
0.6 5 1—oxidized at 50 oC;
2—oxidized at 100 oC;
0.4 2 3—oxidized at 200 oC;
4—oxidized at 400 oC;
5—oxidized at 500 oC
0.2
1
0
0 100 200 300 400 500 600
o
Temperature, C
90
Low temperature selective oxidation of NH3 to N2 on silver-based catalysts
NH3 TPR was carried out on different kinds of silver-based catalysts. The catalyst was
first reduced in H2/He flow at 400 oC overnight and then oxidized in O2/He flow at
1.6
MS(28) Signal (a.u.)
3
1.2
4
5
0.8
6
0.4
2
1
0
0 200 400 600
o
Temperature, C
5
MS(28) Signal (a.u.)
5
4
4
3
3
2
2
1 1
0
0 200 400 600
o
Temperature, C
different temperatures for 2 hours. Fig.12 shows the N2 production profiles on 5wt%
Ag/Al2O3 catalyst. The fact that the nitrogen production maximum moved to higher
temperature when the catalyst was treated in oxygen at higher temperature again
indicated the large effect of oxygen pretreatment. For the 10 wt% Ag/Al2O3 catalyst
91
Low temperature selective oxidation of NH3 to N2 on silver-based catalysts
the effect of the oxygen pretreatment on the position of the nitrogen production
maximum was observed to be less (see Fig.13).
2.5
2
MS Signal (a.u.)
1.5
0.5 N2
N 2O
NO
0
0 200 400 600
o
Temperature, C
Apart from the nitrogen production some N2O and NO are also produced during NH3
TPR. The typical profiles are shown in Fig.14. Apparently N2O and NO formation
peaks were behind the nitrogen formation peak. Studies of ammonia oxidation on Pt
showed that the selectivity to N2 or NO was mainly determined by the surface O/N
ratio [24,25]. A high O/N ratio favored NO formation. In the present case the surface
O/N ratio was always low because there was excess of ammonia in the flow. So a
large amount of N2 was initially produced in NH3 TPR process. The fact that there
was no NO or N2O formation below 200 oC probably indicates that the NO produced
was adsorbed on the silver surface. At higher temperature NO started to desorb and at
the same time N2O was formed from NO through a disproportionation [26] reaction at
the silver surface (NO + NO→N2O +O). The O formed from this reaction will then
react with ammonia to produce N2. That is why there is second nitrogen peak in N2
formation profile. To obtain evidence of the above hypothesis NO TPD was carried
out on reduced silver powder. NO was pre-adsorbed at 50 oC and then a TPD profile
was recorded by MS from 50 to 500 oC at a temperature increase of 10 oC/min.
NO TPD profiles in Fig.15 clearly show that NO started to desorb at about 200 oC and
N2O also formed during NO desorption. Since there is no NH3 in the stream N2
production is not observed.
92
Low temperature selective oxidation of NH3 to N2 on silver-based catalysts
8.0E-04
6.0E-04 NO
MS Signal (a.u.)
4.0E-04
N2
2.0E-04
N 2O
O2
0.0E+00
0 100 200 300 400 500 600
o
Temperature, C
The NH3 TPR profiles on 10 wt% Ag/SiO2 catalyst are very complicated (see Fig.16).
Clear explanation for these nitrogen production profiles seems to be impossible.
Generally this catalyst is also greatly affected by the high temperature pretreatment in
oxygen.
93
Low temperature selective oxidation of NH3 to N2 on silver-based catalysts
0.45
0.4 4
MS(28) Signal (a.u.)
0.35
0.3 3
0.25
2
0.2
1
0.15
0.1
0 100 200 300 400 500 600
o
Temperature, C
94
Low temperature selective oxidation of NH3 to N2 on silver-based catalysts
4. Discussion
Our Raman spectra and H2 TPR spectra clearly showed that there were at least four
types of oxygen species produced when silver was oxidized at high temperature.
These species are adsorbed molecular oxygen, adsorbed atomic oxygen, strongly
adsorbed atomic oxygen and subsurface oxygen respectively. It is surprising that at
room temperature the molecular oxygen, which is expected to desorb below room
temperature according to other researches at UHV conditions, still exists on our silver
powder catalyst. There may be two explanations: 1. The silver catalysts we used have
more structural defects on silver surface which can adsorb oxygen much more
95
Low temperature selective oxidation of NH3 to N2 on silver-based catalysts
strongly. 2. The atmospheric pressure, used in our study, may have great effect on the
adsorption of oxygen which is different from the other studies, using UHV conditions.
Ammonia temperature programmed reaction results showed that all oxygen species on
silver could react with NH3 to produce N2 or N2O but at low temperature only
adsorbed molecular oxygen and atomic oxygen were active. For low silver loading
catalyst the reactivity at low temperature was decreased when catalyst was pretreated
at high temperature in oxygen. Again less effect was found for high silver loading
catalysts pretreated by oxygen at high temperature.
The above results strongly indicate that the ammonia oxidation activity at low
temperature is related to the catalyst’s ability to adsorb oxygen either dissociatively or
non-dissociatively. This means that a high surface area silver catalyst with abundant
lattice defects is favored for low temperature ammonia oxidation.
The catalytic performance results showed that the activities of ammonia oxidation on
silver powder, silica-supported and alumina-supported silver catalysts were almost the
same but the selectivity varied remarkably. Though TEM results showed that the
silver particle sizes on both silica and alumina were in the same range with a wide
particle size distribution, the average particle size of silver on silica measured by O2
chemisorption was larger than that on alumina. This means that there exist many
smaller silver particles on alumina support that cannot be seen by TEM. Distinct XRD
pattern attributable to crystallized silver metal was observed for silica-supported silver
catalyst. The fact that silica-supported silver catalyst behaved similarly to silver
powder catalysts for ammonia oxidation indicated that there was no evidence of direct
interaction between silver and silica. However, the XRD spectra on low loading
alumina-supported silver catalyst showed the existence of a silver oxide phase. With
increasing silver loading the metallic silver phase appeared but the silver oxide phase
still existed. It should be pointed out that only metallic silver should be present on the
support after high temperature calcination just as in the case of silica-supported silver
catalyst. The presence of a silver oxide phase on alumina indicates again that there
exist very small silver particles on alumina support which can be oxidised easily even
at lower temperature. The fact that these small silver particles are very stable on
96
Low temperature selective oxidation of NH3 to N2 on silver-based catalysts
alumina support even at 500 oC may also indicate the present of a strong interaction
between silver and alumina. This type of interaction may involve Al-O-Ag bonding
and much more resistance to the high temperature migration. The silver- alumina
interaction was also mentioned in the study of Naoko Aoyama [27]. They observed
that an alumina-supported silver catalyst had a rather high catalytic activity for NOx
reduction, but Ag catalysts on other oxides were far less active for the reaction. They
concluded that the Ag compounds (not known) formed during catalysis are stabilized
on the alumina surface which in turn is responsible for the high catalytic activity
exhibited by Ag/Al2O3.
The previous Chapter about ammonia oxidation on silver powder catalyst revealed
that two steps were involved [28]. Ammonia was first oxidized to NO. This reaction
step was very fast on silver. Even at room temperature the NO could be quickly
produced and adsorbed on silver surface in form of NOx. At moderate temperature
(below 300 oC) the NO could be removed either as N2O or N2 through a surface SCR
reaction, the second step of ammonia oxidation. At even higher temperature NO could
directly desorb as one of the products. Apparently the SCR performance of the silver-
based catalysts was related closely to the ammonia oxidation performance of the
catalysts. It was reported that there existed a good correlation between the N2
selectivity for the SCO reaction and the SCR performance of NO with NH3 for the Fe-
exchanged zeolites, i.e., the higher the SCR yield of nitrogen, the higher the SCO
selectivity to N2 [29]. The results of the SCR reaction on silver-based catalysts
showed the similar good correlation between the SCR reaction and the SCO reaction.
It seems that the SCR performance (especially the selectivity to nitrogen) of an
alumina-supported catalyst was improved due to the interaction of silver with
alumina. Therefore the alumina-supported catalyst showed a better nitrogen
selectivity compared with silver powder and silica-supported silver catalysts for
ammonia oxidation.
97
Low temperature selective oxidation of NH3 to N2 on silver-based catalysts
5. Conclusions
Based on the above results, it can be concluded that silver-based catalysts are highly
active for ammonia oxidation to nitrogen at very low temperature. Due to the
interaction of silver with alumina, the alumina-supported silver catalyst showed better
activity and selectivity compared to the silica-supported silver catalyst and silver
powder catalyst. Different pretreatment had a large effect on alumina-supported silver
catalysts. Calcination at high temperature decreased the activity of catalyst. Reduction
in hydrogen at 200 oC without any pre-calcination gave the best activity while
reduction at higher temperatures showed little difference from calcination
pretreatment. There were at least four types of oxygen species produced when silver
was oxidized at high temperature. These species are adsorbed molecular oxygen,
adsorbed atomic oxygen, strongly adsorbed atomic oxygen and subsurface oxygen
respectively. Ammonia oxidation activity at low temperature was related to the
catalyst ability to adsorb oxygen either dissociatively or nondissociatively. This
means that a high surface area silver catalyst with abundant lattice defects is favored
for low temperature ammonia oxidation. In addition, there exists a good correlation
for the silver-based catalysts between the N2 selectivity for the SCO reaction and the
SCR performance of NO with NH3, i.e., the higher the SCR yield of nitrogen, the
higher the SCO selectivity to N2.
98
Low temperature selective oxidation of NH3 to N2 on silver-based catalysts
References
[1] M. Amblard, R. Burch, B.W.L. Southward, Appl. Catal. B, 22(3) (1999) L159.
[2] J.J.P. Biermann, Ph.D. Thesis, University of Twente, Twente, 1990.
[3] M. de Boer, H.M. Huisman, R.J.M. Mos, R.G. Leliveld, A.J. Dillen and J.W.
Geus, Catal. Today, 17 (1993) 189.
[4] J.J. Ostermaier, J.R. Katzer and W.H. Manogue, J. Catal., 41 (1976) 277.
[5] J.J. Ostermaier, J.R. Katzer and W.H. Manogue, J. Catal., 33 (1974) 457.
[6] J.E. Delaney and W.H. Manogue, in: Proceedings of the Fifth International
Congress on Catalysis, Vol. 1, 1973, p. 267.
[7] F.J.J.G. Janssen and F.M.G. van den Kerkhof, Selective Catalytic Removal of NO
from Stationary Sources, KEMA Sci. & Techn. Reports, 3(6) (1985).
[8] J.P. Chen and R.T. Yang, Appl. Catal., 80 (1992) 135.
[9] G. Tuenter, W.F. van Leeuwen and L.J.M. Snepvangers, Ind. Eng. Chem. Prod.
Res. Dev., 25 (1986) 633.
[10] E.T.C. Vogt, A. Boot, J.W. Geus and F.J.J.G. Janssen, J. Catal., 114 (1988) 313
[11] J.J.P. Biermann, F.J.J.G. Janssen and J.W. Geus, J. Mol. Catal., 60 (1990) 229.
[12] M. de Boer, A.J. van Dillen and J.W. Geus, Catalysis Letters, 11 (1991) 227.
[13] A. Abe, N. Aoyama, S. Sumiya, N. Kakuta and K. Yoshida, Catal. Lett., 51
(1998) 5.
[14] T. Miyadera and K. Yoshida, 63th Annual Meeting of Chem. Soc. Jpn., 2C447
(1992), in Japanese.
[15] N. Irite, A. Abe, H. Kurosawa, K. Yoshida and T. Miyadera, 65th Annual
Meeting of Chem. Soc. Jpn., 3F135 (1993), in Japanese.
[16] P.H. McBreen and M. Moskovits, J. Catal., 103 (1987) 188.
[17] C. Backx, C.P.M. De Groot, P. Biloen and W.H.M. Sachtler, Surf. Sci., 128
(1983) 81.
[18] J. Eickmans, A. Godmann and A. Otto, Surf. Sci., 127 (1983) 153.
[19] C.T. Campbell, Surf. Sci., 157 (1985) 43.
[20] R.B. Grant and R.M. Lambert, Surf. Sci., 146 (1984) 256.
[21] V.I. Bukhtiyarov, I.P. Prosvirin and R.I. Kvon, Surf. Sci., 320 (1994) L47.
[22] X. Bao, M. Muhler, B. Pettinger and R. Schlogl, Catal. Lett., 22 (1993) 215.
[23] G.R. Meima, L.M. Knijf, A.J. van Dillen, J.W. Geus, J.E. Bongaarts, F.R. van
Buren and K. Delcour, Catal. Today, 1 (1987) 117.
[24] W.D. Mieher and W. Ho, Surf. Sci., 322 (1995) 151.
[25] J.M. Bradley, A. Hopkinson and D.A. King, J. Phys. Chem., 99 (1995) 17032.
[26] K. Otto and M. Shelef, J. Phys. Chem. 76 (1972) 37.
[27] N. Aoyama, K. Yoshida, A. Abe and T. Miyadera, Catal. Lett., 43 (1997) 249.
99
Low temperature selective oxidation of NH3 to N2 on silver-based catalysts
[28] Lu Gang, J. van Grondelle, B.G. Anderson and R.A. van Santen, J. Catal., 199
(2001) 107.
[29] R.Q. Long and R.T. Yang, J. Catal., to be published.
[30] L. Lefferts, Ph.D. Thesis, University of Twente, Twente, 1987.
100
Chapter 6
Bi-functional alumina-supported Cu-Ag catalysts for
ammonia oxidation to nitrogen at low temperature
ABSTRACT
1. Introduction
Most studies on ammonia oxidation focus on the high temperature (>800 K) process
that selectively produces NO because of its industrial importance. However the low
temperature reaction that mainly produces N2 and N2O is becoming more important
because of the environmental aspect. Various catalysts of different types have been
tested for the low temperature ammonia oxidation reaction: biological catalysts [1,2],
metal oxide catalysts [3-6], ion-exchanged zeolites [7-9] and metallic catalysts [10-
12]. When the various types of catalysts are compared it turns out that the metallic
catalysts such as Pt, Ir and Ag are the most active but are the least selective.
Significant amounts of N2O are produced on these catalysts. The metal oxide catalysts
such as CuO, V2O5 and MoO3 show very high selectivity to nitrogen, but the reaction
temperature needed is too high to be matched with some industrial applications.
We have previously reported that silver was very active for ammonia oxidation [13].
Alumina-supported silver catalysts were superior even to noble metal catalysts in both
activity and in selectivity to nitrogen. However the nitrogen selectivity of the silver-
based catalysts (circa 80%) is not sufficient as N2O and NO are still co-produced. As
these compounds are even more toxic than NH3, selectivity to nitrogen of near 100%
is required. Intermediate species and reaction pathways for ammonia oxidation on an
unsupported silver powder catalyst were studied by TPD, TPR, FT-Raman and by
transient as well as steady-state ammonia oxidation experiments [14]. NO was found
to be the main reaction intermediate that produced N2O as well as N2. NO could even
be formed at room temperature. Its subsequent oxidation to NOx led to adsorption on
the silver surface, blocking the active sites for oxygen adsorption. The pathway for
ammonia oxidation at low temperature was found to consist of a two-step consecutive
reaction. Ammonia was first oxidized to NO. This reaction step was very fast on
silver. At moderate temperatures (below 300 oC) NO could then be removed either as
N2O or as N2 through a surface SCR (selective catalytic reduction) reaction, the
second step of the reaction mechanism. At even higher temperatures NO could
directly desorb as one of the products. Apparently the ability of silver-based catalysts
to perform the SCR reaction was closely related to its ability to oxidize ammonia.
102
Bi-functional alumina-supported Cu-Ag catalysts
The bi-functional catalyst required must first contain a component active for ammonia
oxidation at low temperatures. The co-production of N2O can be avoided, or at least
further reduced, by the second component that performs the SCR of the intermediate
NO species. As stated above, Ag is capable of performing the first step. Copper oxide
appeared to be an excellent choice for the second component as it is known to be both
active and selective in the selective catalytic reduction of NO by ammonia. Supported
copper oxide catalysts are commonly reported in the literature for this purpose [25]. In
addition, as previously mentioned, CuO yields a high selectivity to nitrogen in the
SCO reaction. Alumina-supported Ag-Cu catalysts were thus prepared and screened
for their ability to catalyze both the SCO and the SCR reactions. These materials were
also characterized by several spectroscopic techniques in order to determine their
physico-chemical compositions and these were compared with catalytic performance.
The results are discussed in this article.
2. Experimental
103
Bi-functional alumina-supported Cu-Ag catalysts
Low Energy Ion Scattering (LEIS) experiments were performed in co-operation with
Calipso b.v. at the Eindhoven University of Technology. The newly developed ERISS
system was used as the ion scattering apparatus. In this apparatus ion doses can be
reduced to such a low level that the damage of the surface is negligible, thus enabling
the performance of static LEIS [15-20]. To prevent charging of insulating materials, a
neutraliser was available to spray low energy electrons to the sample. Samples were
prepared by pressing the ground sample with a load of 1600 kg into a tantalum cup (Ν
1 cm) and subsequently loaded into the pretreatment chamber that was connected to
the main UHV chamber. All the samples were pretreated at 300 oC in 214 mbar of
oxygen for 20 min. Most LEIS measurements were performed using 5 keV Ne+ ions;
these provided a good mass resolution for the heavy elements (Ag and Cu). A number
of samples were also analyzed using 3 keV 4He+ ions. This yielded a good sensitivity
for the light elements (C, O, Al, Si). In most cases, the samples were mechanically
scanned over an area of 1×1 mm2. During the measurements, the pressure in the main
chamber was determined by the noble gas used in the ion source and was typically
around 1×10-8 mbar. The base pressure of the system was in the 10-10 region. Analysis
of the outermost surface layer and at different depths was performed on both the
catalysts and on the gamma–alumina carrier. Depth analysis was performed by
sputtering the samples with Ne+ ions or 4He+ ions, thus removing up to 8 ML (Ne) and
up to 0.8 ML (He).
104
Bi-functional alumina-supported Cu-Ag catalysts
3. Results
Fig.1 shows the dependence of both the ammonia conversion level and the selectivity
to nitrogen on the catalyst composition at 250 oC. It can be seen that the conversion of
ammonia decreases with increasing copper weight ratio in the catalyst, but that the
nitrogen selectivity changes in the opposite direction. It has previously been shown
that selectivity is a very weak function of the ammonia conversion level [23]. The
optimum Ag/Cu weight ratio seems to be in the range between 1 and 3. Both activity
and selectivity to nitrogen are high in this range. Apparently, the nitrogen selectivity
can be greatly improved by addition of small amounts of copper to the silver catalyst
system with only a small loss in catalyst activity.
100 100
90
N2 Selectivity (mole%)
95 80
70 Conversion (mole%)
90 60
50
85 40
Con., %
30
Sel., %
80 20
10
75 0
Ag 9 3 1 0.33 Cu
Ag/Cu weight Ratio
Fig. 1 The effect of Ag/Cu weight ratio on ammonia conversion and on selectivity to nitrogen for Ag-,
Cu-, and AgCu/γ-Al2O3 catalysts.
(Reaction conditions: NH3: 1.14 vol.%; O2: 8.21 vol.%; flow rate=74.7 Nml/min; catalyst weight=0.2
g; T=250 oC)
105
Bi-functional alumina-supported Cu-Ag catalysts
10 wt%Ag/Al2O3 200 11 88
250 98 86
300 100 83
Ag/Al2O3+Cu/Al2O3 200 8 89
(Ag/Al2O3 : Cu/Al2O3 =3:1) 250 96 87
300 100 82
Reaction conditions: NH3 = 1.14%; O2 = 8.21 vol%; flow rate = 74.7 Nml/min;
catalyst weight = 0.2 g
100 Conv.
Conv. or Sel., %
80
Sel. N2
60
40
Sel. NOx
20
Sel. N2O
0
100 200 300 400 500
o
Temperature, C
Fig. 2 Ammonia oxidation on 7.5 wt% Ag-2.5 wt% Cu/Al2O3 catalyst at various temperatures.
Reaction conditions: NH3=1000 ppm; O2=10 vol%; total flow rate=50 Nml/min; catalyst weight=0.1 g.
Fig.2 shows the conversion of ammonia and the selectivity to various products on an
alumina-supported 7.5 wt% Ag-2.5 wt% Cu catalyst as a function of temperature.
Between 200 -300 oC the catalyst converted 100% of the ammonia with a selectivity
to nitrogen of circa 95%. Above 300 oC, more NOx is produced. The catalyst was
quite stable. There was no significant deactivation during a one-day test under these
experimental conditions. However, the selectivity did change somewhat with time
(see Fig.3). It can be seen from Fig.3 that the selectivity to nitrogen increased with
reaction time during the first 40 minutes on stream, whereafter it remained
106
Bi-functional alumina-supported Cu-Ag catalysts
unchanged. This is mainly caused by adsorbed NOx, N2Ox species produced during
reaction which can lower the surface oxygen coverage as discussed previously [14].
100
N2
75
Selectivity, %
50
25
N 2O
0
0 20 40 60 80
Time, min
Fig. 3 Selectivity change with time on 7.5 wt% Ag-2.5 wt% Cu/Al2O3 catalyst
Reaction conditions: NH3=1000 ppm; O2=10 vol%; total flow rate=50 Nml/min; catalyst weight=0.1 g.
10 wt%Cu/Al2O3 150 53 98
200 97 97
250 99 98
10 wt%Ag/Al2O3 200 5 67
250 14 74
300 25 72
Reaction conditions: NO = 500 ppm; NH3 = 1000 ppm; O2 = 8 vol%; flow rate = 100 Nml/min; catalyst
weight = 0.2 g
107
Bi-functional alumina-supported Cu-Ag catalysts
Fig.4 shows the XRD diffractograms of 5 wt% Cu/Al2O3, 5 wt% Ag/Al2O3 and 7.5
wt% Ag-2.5 wt% Cu/Al2O3 catalysts. All the catalysts were calcined at 500 oC in air
for 24 hours before measuring. No XRD diffraction lines attributable to crystallized
copper metal or other copper compounds were observed on the copper-containing
samples. However, diffraction lines attributable to AgO appeared in the XRD spectra
of both silver-containing samples. Low intensity diffraction lines due to metallic
silver were also observed on the 7.5wt%Ag-2.5wt%Cu catalyst.
300
CPS
200
5%Ag/Al 2O3
7.5%Ag-2.5%Cu
100 /Al2O3
5%Cu/Al2O3
0
20 40 60 80 100
2 Theta
108
Bi-functional alumina-supported Cu-Ag catalysts
20k
19k
15k
380 375 370 365 360 370 360 350 340
Ag 3d binding energy [eV] Kinetic energy [eV]
4.5k
4.0k
5.0 wt% Ag and 5.0 wt%
Auger intensity [Counts/s]
Cu on Al2O3 3.5k
XPS intensity [A.U.]
10 wt% Cu on Al2O3
Fig. 6(a) shows XPS spectra in the Cu 2p region [925- 960 eV]. The most intense
peak occurs at about 932.5 eV. Comparing this result with literature data [21] it seems
that Cu is in an oxidization state of +1 if copper is present as copper oxide since the
109
Bi-functional alumina-supported Cu-Ag catalysts
2p electrons in Cu+2 have a much higher binding energy (933.6 eV). Auger bands in
the Cu region are shown in Fig. 6(b). The kinetic energies are circa 914 eV, much
lower than that of CuO (918.1 eV) and Cu2O (916.2 eV). The UV-spectra of our γ-
alumina-supported copper catalysts [9] and EXAFS studies on other γ-alumina-
supported copper catalysts [24] clearly showed that copper was in an oxidization state
of +2. The shift of the binding energy may be caused by the interaction of copper with
the support. New copper species other than copper oxide are formed due to this
interaction. A second possibility is an interaction between Ag and Cu.
In order to be able to exclude any interaction between silver and copper two solid
solutions were made of 5 wt% copper in silver and 5 wt% silver in copper. A special
procedure was needed to fabricate these solid solutions, as at low temperatures these
metals do not mix at all, and will phase-separate into micro-crystallites when cooling
down from the melt. Such rough samples were prepared by melting the appropriate
amounts together in an alumina crucible by induction in a high-frequency furnace,
after which the samples were placed in two evacuated quartz capsules. These were
heated for 96 hours at a steady temperature of 800 oC, at which temperature solid
solutions of these compositions are possible according to the phase diagram [27]
Subsequent quenching prevented the formation of micro-crystallites. The XPS spectra
of the Ag 3d and Cu 2p did not show any difference from the pure metals after
polishing. Any direct interaction between copper and silver, either through formation
of an alloy or at boundaries is thus very unlikely.
The samples measured with LEIS are listed in Table 3. In addition to the Ag/Cu/γ-
alumina catalysts, we analysed a bare (-alumina carrier material and Ag- and Cu foil
reference samples. Furthermore, in order to analyse the influence of surface roughness
on the results we measured both a quartz sample and a silica carrier (Grace 332).
Although these samples had the same elemental composition, quartz had a very
smooth surface whereas the carrier material exhibits a much larger surface roughness
(specific area of 235 m2/g). Finally, an “ultra-pure” CuO powder sample was analyzed
for quantification purposes. Figures 7(a) and 7(b) show the LEIS spectra of 5 wt%
Cu/γ-alumina and 10 wt% Ag/γ-alumina samples as obtained with 5 keV Ne+ ions. At
the outermost atom layers of the samples significant Ag- and Cu- signals were
detected.
110
Bi-functional alumina-supported Cu-Ag catalysts
10 8
Cu
Ag
8 1 ML
6 0.4 ML
5 ML surface
surface
Yield (counts/nC)
Yield (counts/nC)
6 2.4 ML
2
2
0 0
1000 1500 2000 2000 2400 2800
Energy (eV) Energy (eV)
Fig.7(a) 5 keV Ne+ spectra of 5 wt% Cu/Al2O3 Fig.7(b) 5 keV Ne+ spectra of 10 wt% Ag/Al2O3
at different depths at different depths
The removal of the first (fraction of a) monolayer (ML) results in increased Ag- and
Cu signals. After a few ML’s the intensities returned to a magnitude on the order of
that of the outermost surface (Cu/γ-alumina) or slightly lower (Ag/γ-alumina). Most
likely, the first ML removed consisted of contamination that had been deposited on
111
Bi-functional alumina-supported Cu-Ag catalysts
the surface during the in-situ calcination. The subsequent decrease of the signals
suggests that prolonged sputtering resulted in smaller “accessible” silver and copper
surfaces. Fig. 8 shows one of the 5 keV Ne+ spectra recorded for the bi-metallic
catalyst (10 wt% Ag-2.5 wt% Cu/Al2O3). In all cases, the removal of the first (fraction
of a) ML resulted in increased intensities for both Ag and Cu. Again, we assume that
we are dealing with surface contamination that, most likely, originated from the
calcination of the samples.
8
Ag
1 ML
Yield (counts/nC)
6
Cu surface
5 ML
4
0
1000 2000 3000
Energy (eV)
Fig.8 5 keV Ne+ spectra of Cu/Ag/γ-alumina (2.5%w Cu, 10%w Ag) at different depths
Going deeper, down to 3-5 ML’s, the respective Cu signals decreased in a similar
fashion for all the copper-containing samples. Concerning Ag, we observed a
relatively large intensity drop for the sample shown in Figure 8. The 3 keV 4He+
spectrum recorded for the catalyst 10 wt% Ag-2.5 wt% Cu/Al2O3 also shows the F
peak originating from the bare carrier (Figure 9).
Table 4 Calculation of CuO platelets and Ag particles dispersion from LEIS measurements
Catalysts CuO platelets Ag particles
depth: 1 ML depth: 1 ML
Cu Ag Ag/Cu Cu LEIS Cu visible CuO platelets Ag LEIS Ag visible Ag particles
signal atoms signal atoms
wt% wt% at/at cnts/nC atoms/cm2 thickness, cnts/nC atoms/cm2 av. size
x103 #ML's x103 (nm)
112
Bi-functional alumina-supported Cu-Ag catalysts
12 O
Yield (counts/nC)
Al
F
8 surface
0.2 ML
1.6 ML
4 Ag
Cu
0
750 1750 2750
Energy (eV)
113
Bi-functional alumina-supported Cu-Ag catalysts
47 nm
47 nm
114
Bi-functional alumina-supported Cu-Ag catalysts
47 nm
3
2
1
Table 5 Relative amounts of copper and silver measured by EDX on different areas of 7.5wt%Ag-
2.5wt%Cu/Al2O3 (black-particle fraction). Data points refer to numbered arrows in Fig. 11(a).
Data point atom % atom %
Cu Ag
1 6.90 93.10
2 3.82 96.18
3 3.52 96.48
4 5.97 94.03
5 59.70 40.30
(area without dark
particles)
Two samples of mixed Ag-Cu/alumina catalysts were examined using TEM and EDX
elemental analysis. Both samples contained a total metal loading of 10 wt%. One
sample contained 7.5 wt% Ag and 2.5 wt% Cu. The other sample contained 2.5 wt%
Ag and 7.5 wt% Cu.
This sample clearly consisted of two different macroscopic phases: one containing
black particles and the other containing green particles. These two phases were
mechanically separated (by hand) and were analyzed separately. Fig. 11(a) shows a
representative TEM image that was obtained for the fraction containing black
particles. EDX analysis was performed at several different locations on the sample
(see numbered arrows). Data points 1 to 4 were chosen to analyze areas containing
dark particles. Data point 5 measured an area in which no dark particles were present.
115
Bi-functional alumina-supported Cu-Ag catalysts
As the relative amounts of copper and silver are of importance these are shown in
Table 5. Clearly the areas that include particles visible by TEM (points 1-4) contain
very high amounts of silver relative to copper (circa 95:5). By contrast the area
containing no visible particles (point 5; and others not shown here) contains relatively
more copper than silver (circa 60:40).
47 nm 7
2
6
5
3
Table 6 Relative amounts of copper and silver measured by EDX on different areas of 7.5wt%Ag-
2.5wt%Cu/Al2O3 (green-particle fraction). Data points refer to numbered arrows in Fig. 11(b).
Data point atom % atom %
Cu Ag
1 1.02 98.98
2 1.10 98.90
3 2.10 97.90
4 1.10 98.90
5 1.68 98.32
6 0.86 99.14
7 73.40 26.60
(area without dark
particles)
Fig. 11(b) shows a TEM image that was obtained for the macroscopic green particles
of this Ag-rich sample. The relative amounts of copper and silver (by EDX analysis)
are listed in Table 6. Similar to the black-coloured particle fraction, this phase also
contained a wide distribution of particle size (up to 50 nm). These particles (data
points 1-6) were slightly richer in silver relative to copper than were the black
116
Bi-functional alumina-supported Cu-Ag catalysts
particles. Again although data point 7 shows no particles visible by TEM, both copper
and silver were present, copper again being the more abundant.
12nm
3
1
8
6
7
2
4 5
Fig.12 TEM image of green particles of 2.5%Ag-7.5%Cu/Al2O3 catalyst
(the arrows point to the area where the EDX measurement is performed)
Fig. 12 shows a representative TEM image that was obtained for this Cu-rich sample.
Again particles of a large distribution were observed. Table 7 shows the relative
Table 7 Relative amounts of copper and silver measured by EDX on different areas of 2.5wt%Ag-
7.5wt%Cu/Al2O3. Data points refer to numbered arrows in Fig. 12.
Data point atom % atom %
Cu Ag
1 3.26 96.74
2 4.56 95.44
3 6.61 93.39
4 4.62 95.38
5 12.89 87.11
6 11.85 88.15
7 13.75 86.25
8 83.28 16.72
(area without dark
particles)
amounts of copper and silver measured at various locations. Again the darker particles
visible by TEM (points 1-7) contain a larger abundance of silver. However in several
locations the relative amount of copper to silver is higher than was observed in the
117
Bi-functional alumina-supported Cu-Ag catalysts
previous samples. Data point 8 again shows that copper and a smaller amount of
silver are present even in an area without visible dark particles.
4. Discussion
Ammonia oxidation experiments have shown that silver alone supported on alumina
has a very high ammonia oxidation activity but with a low selectivity to nitrogen.
Alumina-supported copper catalyst has a very high nitrogen selectivity but the activity
is low at low temperature. As discussed previously the aim of this study was to
determine if alumina-supported binary Cu-Ag catalysts yield a higher selectivity to
nitrogen without decrease in activity. The results show that this is indeed the case.
Addition of Cu to Ag/Al2O3 catalysts, by incipient wetness co-impregnation, showed
an increase in N2 selectivity from 87% to circa 95% at 250 oC whilst the activity was
not significantly reduced at Ag/Cu weight ratios greater than unity. An optimum ratio
of between 1-3 was observed. The activity and selectivity to nitrogen for the SCR of
NO by NH3 were also increased following the addition of Cu. This is consistent with
the previous observation that materials capable of catalyzing the SCO of NH3 are also
able to catalyze the SCR of NO by NH3. This is due to the fact that the former
reaction proceeds through an NO intermediate. Once formed the NO can be
selectively reduced by NH3 to nitrogen rather than being further oxidized to N2O.
The fact is that a mechanical mixture of Cu/Al2O3 and Ag/Al2O3 catalysts, with an
overall Ag/Cu ratio of 3:1, did not produce an increase in the selectivity to nitrogen as
observed in the catalysts prepared by incipient wetness co-impregnation. The above
result indicates that either close physical proximity or intimate chemical interaction
between the silver and the copper components is necessary in order to increase
selectivity. This can be understood by the fact that the intermediate NO cannot desorb
from the catalyst surface at temperatures below 300 oC, as shown previously [9]. Thus
at low temperatures, transport of NO between Ag and Cu must occur by surface
diffusion.
118
Bi-functional alumina-supported Cu-Ag catalysts
Based on the absence of distinct Cu particles in the TEM micrographs, it has been
suggested that Cu might have reacted with (surface layers of) γ-alumina and formed
CuAl2O4. However, the Cu atoms of this “normal spinel” component are known to
occupy tetrahedral, i.e. sub-surface, positions. Consequently, we would not expect
CuAl2O4 to yield the large copper LEIS signals we actually have obtained. Former
LEIS experiments performed on normal spinels yielded much smaller signals.
Furthermore, small Cu (oxide) species may simply exhibit insufficient contrast (in
TEM) when imaged on γ-alumina. The XRD results also show no diffraction line of
CuAl2O4 phase for all the copper containing samples. For these reasons, we judge the
occurrence of spinels to be rather unlikely. Excluding the presence of spinels, Cu may
either be present in metallic- or in oxidic form. Following surface tension
considerations, metallic Cu species would form distinct particles whereas CuO
species would wet the γ-alumina surface much better. The preparation of the catalysts
(through wet impregnation of a mixture of Ag- and Cu nitrates, followed by a
reduction step) is expected to yield CuO species rather than metallic Cu particles.
Furthermore, the catalysts were calcined prior to be analysed with LEIS analysis. For
these reasons, it is very likely that we are dealing with CuO species. Although the
expected form of copper oxide is Cu(II) oxide, XPS analysis of Cu/Al2O3 catalyst
employing Wagnen plots revealed that copper is in an oxidation state of +1, thus
suggesting the presence of Cu2O.
Analysis of 5 and 10 wt% Ag/Al2O3 catalysts by TEM and EDX showed that Ag was
always present as particles on the alumina surface. This is consistent with
expectations based on surface tension between Ag and Al2O3. TEM revealed that a
very wide particle distribution was always present (1-50 nm). XPS analysis of
Ag/Al2O3 employing Wagnen plots showed that Ag was present in an oxidation state
between +1 and +2. XRD analysis also showed the presence of AgO besides metallic
silver. This is surprising as Ag2O is very unstable and AgO decomposes at 200 oC, so
that following calcination at 500 oC metallic silver was expected.
TEM analysis of binary Cu-Ag catalysts revealed that a very large particle distribution
was present (1-50 nm). EDX analysis revealed that the particles contained very large
amounts of Ag relative to Cu. Areas in which no particles were present contained
relatively more copper but still contained some silver. This suggests again that Cu was
highly dispersed and that the particles were of Ag. Silver particles of very small
diameters (<1 nm) must also be present. The observations by TEM are consistent with
the results of LEIS. A model generated suggested that copper oxide was highly
dispersed, as a monolayer, and that the silver particles sit on top of this layer. The
average Ag particle size for all silver-containing samples determined by LEIS was
119
Bi-functional alumina-supported Cu-Ag catalysts
between 2.5 and 5.5 nm. The simple model predicts that the silver particle size
distribution is quite narrow and does not vary a lot amongst all the silver-containing
catalysts. This distribution is much smaller than was observed by TEM. This probably
arises from the fact that LEIS measures the silver content of the entire surface thus
including the small particles invisible to TEM.
XPS results show that the valence states of copper and silver are the same for single
component catalysts and bi-component catalysts. The TEM images show very broad
Ag particle size distributions for all these catalysts. There is no indication for the
formation of a new Cu-Ag phase. Indeed, as shown by the phase diagram [27], Ag
and Cu have very low solubilities together at temperatures below 780 °C. They tend
to phase-separate at lower temperatures. There is no stable region in which a Cu-Ag
alloy is formed.
Since the NO produced in the first step of ammonia oxidation cannot desorb at low
temperature, the copper and silver components should be in intimate contact with
each other so that the NO produced on the silver surface can migrate to the copper
surface easily. Otherwise, NO produced on the silver cannot reach the copper surface
and the chances are greater that N2O is produced rather than nitrogen. The EDX
results showed that copper and silver existed together everywhere on the alumina
surface. This proved that there was an intimate contact between copper and silver on
alumina-supported Cu-Ag catalysts. Our experiments show that the performance of
the catalyst, which is a physical mixture of 10 wt% Ag/Al2O3 and 10 wt% Cu/Al2O3,
behaves similarly to the 10 wt% Ag/Al2O3 catalyst alone. This again confirms the
importance of intimate contact between copper and silver.
In summary, copper was highly dispersed in all catalysts analyzed by LEIS and TEM.
Most likely, the catalysts contain copper oxides platelets with a thickness of roughly 1
ML. At increasing copper loading, relatively more (or more extended) rather than
thicker platelets are formed. However, silver has a broad particle size distribution.
The average size of the Ag species ranges between 2.5 and 5.5 nm measured from
LEIS. These numbers increase with increasing silver loading. From TEM images
120
Bi-functional alumina-supported Cu-Ag catalysts
silver is present as relatively large particles with sizes ranging from 1 nm up to 50 nm.
The EDX results also show the existence of silver in the areas without dark particles.
This indicates that a fraction of silver is dispersed in the form of platelets or very
small particles that cannot be observed by TEM. This fraction of silver may be most
interesting from a catalytic point of view. Unfortunately we still cannot
unambiguously determine which fraction of silver plays the major role in ammonia
oxidation at this stage.
Multi-functional catalysis is not a new concept, but catalyst design deliberately based
on the principle of multifunctionality will be inevitable to develop high-performance
catalysts which can cope with such difficult problems as in ammonia oxidation. This
study only gives one example of such a design. Actually there are more alternatives
available. Besides silver, many noble metals such as Pt, Ir are very active for
ammonia oxidation with low selectivity to nitrogen due to the low efficiency for SCR
reaction. There are also many good SCR catalysts such as vanadium oxide, iron oxide.
Opportunities still exist for preparing high-performance bi-functional catalysts from
these catalytic components.
121
Bi-functional alumina-supported Cu-Ag catalysts
5. Conclusions
Alumina-supported silver catalysts are very active for ammonia oxidation but N2O is
also co-produced on these catalysts. The addition of copper to these silver-based
catalysts can greatly improve the catalytic selectivity to nitrogen with a negligible loss
in activity for ammonia oxidation within the optimum Ag / Cu weight ratio range of
between 1 and 3. XPS results show that the valence states of copper and silver are the
same for single component catalysts and bi-component catalysts. LEIS and XPS
measurements show no indication of formation of a new Cu-Ag phase. Such Cu-Ag
alloys have been shown earlier to be thermodynamically unstable.
TEM images show the very broad particle size distributions for all these catalysts.
EDX results showed that copper and silver existed together everywhere on the
alumina surface thus intimate contact between copper and silver on alumina-
supported Cu-Ag catalysts exists. All results seem to be consistent with a model that
suggests that the catalyst structure consists of a highly dispersed (monolayer) of
copper oxide on alumina upon which silver particles of a very wide particle size
distribution sit.
122
Bi-functional alumina-supported Cu-Ag catalysts
REFERENCES
[1] M.C.M. van Loosdrecht and S.J. Heijnen, TIBTECH, 11 (1993) 117.
[2] J. Cole, TIBTECH, 11 (1993) 368.
[3] M. Amblard, R. Burch, B.W.L. Southward, Appl. Catal. B, 22(3) (1999) L159.
[4] J.J.P. Biermann, Ph.D. Thesis, University of Twente, Twente, 1990.
[5] M. de Boer, H.M. Huisman, R.J.M. Mos, R.G. Leliveld, A.J. Dillen and J.W.
Geus, Catal. Today, 17 (1993) 189.
[6] A. Wollner and F. Lange, Appl. Catal. A, 94 (1993) 181.
[7] Y. Li and J. N. Armor, Appl. Catal. B, 13 (1997) 131.
[8] N.N. Sazonova, A.V. Simakov and H. Veringa, React. Kinet. Catal. Lett., 57(1)
(1996) 71.
[9] Lu Gang, J. van Grondelle, B.G. Anderson and R.A. van Santen, J. Catal., 186
(1999) 100.
[10] J.J. Ostermaier, J.R. Katzer and W.H. Manoque, J. Catal., 41 (1976) 277.
[11] N.I. II’chenko, Russian Chemical Reviews, 45 (1976) 1119.
[12] A.C.M. van den Broek, Ph.D. Thesis, Technical University of Eindhoven,
Eindhoven, 1998.
[13] Lu Gang, J. van Grondelle, B.G. Anderson and R.A. van Santen, Appl. Catal. B,
Submitted in 2001.
[14] Lu Gang, J. van Grondelle, B.G. Anderson and R.A. van Santen, J. Catal., 199
(2001) 107.
[15] G.J.A. Hellings, H. Ottevanger, S.W. Boelens, C.L.C.M. Knibbeler and H.H.
Brongersma, Surf. Sci., 162 (1985) 913.
[16] C.L.C.M. Knibbeler, G.J.A. Hellings, H.J. Maaskamp, H. Ottevanger and
H.H. Brongersma, Rev. Sci. Instrum., 58 (1987) 125.
[17] P.A.J. Ackermans, P.F.H.M. van der Meulen, H. Ottevanger, F.E. van Straten
and H.H. Brongersma, Nucl. Instr. Meth., B35 (1988) 541.
[18] O. van Kessel, H.H. Brongersma, J.G.A. Hölscher, R.G. van Welzenis, E.G.F.
Sengers and F.J.J.G. Janssen, Nucl. Instr. Meth., B64, (1992) 593.
[19] J.P. Jacobs, A. Maltha, J.G.H. Reintjes, J. Drimal, V. Ponec and H.H.
Brongersma, J. Catal., 147 (1994) 133.
[20] J.C. Fullarton, J.-P. Jacobs. H.E. van Benthem, J.A. Kilner, H.H. Brongersma,
P.J. Scanlon and B.C.H. Steele, Ionics, 1 (1995) 51.
[21] J.F. Moulder, W.F. Stickle, P.E. Sobol and K.D. Bomben, “Handbook of X-ray
Photoelectron Spectroscopy” (J. Chastain, Ed.), Perkin-Elmer Corporation, Eden
Prairie, USA, 1992.
[22] D. Briggs and M.P. Seah (Editors), “Practical Surface Analysis by Auger and
X-Ray Photoelectron Spectroscopy”, Wiley, New York, 1983.
123
Bi-functional alumina-supported Cu-Ag catalysts
[23] Lu Gang, J. van Grondelle, B.G. Anderson and R.A. van Santen, Catal. Today,
61 (2000) 179.
[24] R.M. Friedman, J.J. Freeman and F.W. Lytle, J. Catal., 55 (1978) 10.
[25] J.N. Armor, Appl. Catal., 1 (1992) 221.
[26] C.D. Wagner, L.H. Gale and R.H. Raymond, Anal. Chem., 51(4) (1979) 466.
[27] P.M. Hansen, “Constitution of Binary Alloys”, McGRAW-HILL BOOK
COMPANY, INC., 1958.
124
Summary
The selective catalytic oxidation (SCO) of ammonia with oxygen to nitrogen and
water recently is considered to be an efficient method to abate ammonia pollution.
Previous studies show that noble metals such as Pt, Ir are very active for this reaction
at low temperature but less selective. Significant amounts of nitrous oxide or nitric
oxide are produced on these catalysts, especially under the condition of very high
O2/NH3 ratios. Metal oxide catalysts such as Co3O4, MnO2, CuO, Fe2O3, V2O5 and
MoO3 are also studied in the literature but the efficiencies are not enough at low
temperature for a practical use. The research described in this thesis was aimed at the
development of new, active and selective catalysts for low temperature selective
ammonia oxidation to nitrogen.
lattice oxygen can react with ammonia to produce N2. However, surface oxygen was
much more active than lattice oxygen at low temperature. NH3 TPD on CuY catalysts
showed three types of active centers. Two of these were active at low temperature
(below 200 oC) and one was active at a higher temperature (above 300 oC). The
existence of low temperature active centers indicated that ammonia oxidation at low
temperature on copper catalysts was possible. According to the UV spectra, the [Cu-
O-Cu]2+-like species or small copper oxygen aggregates were responsible to the low
temperature active centers. However the amount of low temperature active centers or
the concentration of [Cu-O-Cu]2+ species was small prior to NaOH treatment. The
NaOH treatment of CuY increased the amount of low temperature active centers.
These studies apparently indicate that the environment or the type of active copper
species is very important for low temperature ammonia oxidation and is strongly
related with different supports and preparation methods.
In Chapter 4 ammonia oxidation reaction pathways on high surface area silver powder
has been studied by TPD, TPR, FT-Raman and transent as well as steady-state
ammonia oxidation experiments. Silver is a very active catalyst for ammonia
oxidation. At low temperature (below 300 oC) mainly N2 and N2O are produced. At
higher temperature NO instead of N2O becomes one of the products. NO is an
important reaction intermediate for this reaction. Even at room temperature the NO
can be produced and adsorbed on silver surface in the form of NOx. Since NO can not
desorb at low temperature it blocks the active sites for oxygen dissociation. The
dissociation of oxygen is thus believed to be the rate-controlling step for ammonia
oxidation. The selectivity to N2, N2O and NO is mainly determined by surface oxygen
coverage and temperature. Low surface oxygen coverage favors nitrogen formation.
Adsorbed NOx, N2Ox species are actually inhibitors for ammonia oxidation but they
also lower the surface oxygen coverage. Hence, the selectivity to nitrogen is improved
by increasing the amount of these adsorbed species on silver surface.
conclusions that blocking of the sites for oxygen dissociation is an effective way to
improve the nitrogen selectivity, but also would result in a loss of catalyst activity.
In Chapter 5 low temperature gas phase oxidation of ammonia to nitrogen has been
studied over alumina-supported, silica-supported and unsupported silver catalysts to
distinguish the support effect on silver-based catalysts. TPD, TPR, TEM, XRD and
FT-Raman were used to characterize the different silver catalysts. The results showed
alumina-supported silver to be the best catalyst due to the interaction of silver with
alumina. Pretreatment had a great affect on the catalyst performance. Reduction in
hydrogen at 200 oC without any pre-calcination gave the best activity while reduction
at higher temperatures showed little difference from calcination pretreatment. At least
four types of oxygen species were produced when silver was oxidized at high
temperature. These species are adsorbed molecular oxygen, adsorbed atomic oxygen,
subsurface oxygen and bulk dissolved oxygen respectively. Ammonia oxidation
activity at low temperature is related to the catalyst’s ability of dissociatively or non-
dissociatively adsorption of oxygen. In addition, a good correlation existed between
the N2 selectivity for SCO reaction and the SCR performance of NO with NH3 for the
silver-based catalysts, i.e., the higher SCR yield of nitrogen, the higher the SCO N2
selectivity.
Multi-functional catalysis is not a new concept, but catalyst design deliberately based
on the principle of multifunctionality will be inevitable to develop high-performance
catalysts which can cope with such difficult problems as in ammonia oxidation. This
study only gives one example of such a design. Actually there are more alternatives
available. Besides silver, many noble metals such as Pt, Ir are very active for
ammonia oxidation with low selectivity to nitrogen due to the less efficiency for SCR
reaction. There are also many good SCR catalysts such as vanadium oxide, iron oxide.
Opportunities still exist for preparing high-performance bi-functional catalysts from
these catalytic components.
128
Samenvatting
De selectieve katalytische oxidatie (SCO) van ammoniak met zuurstof naar stikstof en
water is bekend als een doelmatige oplossing voor de vermindering van de NH3
uitstoot. Tussen de katalysatoren die genoemd worden als mogelijke materialen voor
dit proces, zijn edelmetaal katalysatoren zoals Pt, Ir heel aktief bij lage temperatuur
maar niet selectief naar stikstof. Bovendien produceeren ze signifikante hoeveelheden
N2O. Metaaloxides zoals Co3O4, MnO2, CuO, Fe2O3, V2O5 en MoO3 zijn ook aktief
voor ammoniak oxidatie maar de activiteit is niet hoog genoeg om te worden
toegepast als industriele katalysator. Het onderzoek, beschreven in dit proefschrift, is
gericht op de ontwikkeling van een nieuwe, aktieve en selectieve katalysator voor de
lage temperatuur selectieve oxidatie van ammoniak naar stikstof.
In hoofdstuk 4 is een studie naar het reaktiepad van ammoniakoxidatie over hoge
oppervlakte zilverpoederkatalysator beschreven. De zilverpoederkatalysator werd
bestudeerd met TPD, TPR, FT-Raman en transient alsook steady-state ammoniak
oxidatie. Zilver is een zeer actieve katalysator voor ammoniakoxidatie. Bij lage
temperatuur (beneden 300 oC), werd voornamelijk N2 and N2O gevonden. NO werd
gezien als een bijproduct bij hogere temperatuur (boven 300 oC). NO was een
belangrijk reaktie intermediair voor deze reaktie. Zelfs bij kamertemperatuur kan NO
worden geproduceerd en op het oppervlak van de katalysator als NOx geadsorbeerd.
Omdat NO niet bij lage temperatuur desorbeert, wordt de actieve site voor zuurstof
dissociatie geblokkeerd. De snelheidsbepalende stap voor ammoniak oxidatie moet
dus de zuurstofdissociatie zijn. De selectiviteit naar N2, N2O en NO wordt
voornamelijk bepaald door de temperatuur en de oppervlakte bedekking van
zuurstofatomen. Een lage oppervlaktebedekking met zuurstofatomen bevordent de
vorming van stikstof. Geadsorbeerde NOx, N2Ox species verhinderen feitelijk de
ammoniak oxidatie, maar ze verlagen ook de oppervlaktebedekking van
zuurstofatomen. Daarom wordt de selectiviteit naar stikstof verbeterd door
toenemende bedekking van deze species op het zilveroppervlak.
130
Samenvatting
geproduceerd. In het algemeen is zilver, evenals Pt, geen goede SCO katalysator
omdat te veel N2O wordt geproduceerd. De belangrijkste conclusie van bovenstaande
resultaten is dat de blokkering van actieve sites voor zuurstofactivering een
doelmatige wijze is om de selectiviteit naar stikstof te verbeteren, maar met enig
verlies van activiteit.
Omdat SCO uit verschillende reaktiestappen bestaat is het erg moeilijk om een enkele
aktieve fase te vinden die alle processen simultaan katalyseert. Multi-functionele
katalysatoren proberen het probleem op te lossen door meerdere aktieve katalytische
fasen te combineren. Een voorbeeld van zo’n aanpak wordt beschreven in hoofdstuk 6
met de Cu-Ag/Al2O3 katalysatoren. Op zilver gebaseerde katalysatoren zijn bekend
om hun grote activiteit voor de SCO reaktie, echter, een minpunt is hun grote
selectiviteit voor de vorming van N2O. Een Cu-Ag/Al2O3 katalysator bleek niet alleen
aktief maar ook selectief. Karakterisatie van deze katalysatoren door XRD, XPS,
LEIS, TEM en EDX hebben aangetoond dat het Ag aanwezig is als fijn verdeeld Ag
deeltjes met 2.5-5.5 nm in diameter, terwijl Cu aanwezig is als koperoxide plaatjes
met ongeveer 1 monolaag dikte. Ook werd geen indicatie gevonden voor de vorming
van een nieuwe Cu-Ag fase of Cu-Ag legering. Bovendien is door EDX gevonden dat
waar dan ook op het alumina oppervlak koper en zilver samen schijnen te bestaan. Dit
suggereert dat er een intiem contact tussen zilver en koper op de Cu-Ag/Al2O3
katalysatoren is opgetreden. Het mechanisme van bi-functionele katalysator kan dus
het promotor effect van kopertoevoeging verklaren. Zoals al eerder gezegd zijn
131
Samenvatting
Ag/Al2O3 katalysatoren erg actief voor de vorming van NO, de eerste stap van
ammoniak oxidatie. De Cu/Al2O3 katalysatoren, echter, zijn erg actief voor de SCR
reaktie van NO met NH3, de tweede stap van ammoniak oxidatie.
132
Publications
NH3 oxidation to nitrogen and water at low temperatures using supported transition
metal catalysts
Lu Gang, B.G. Anderson, J. van Grondelle, R.A. van Santen, Catalysis Today, 61,
179(2000)
133
Distribution of mixed alcohols in products synthesized over molybdenum sulfide
catalysts
W. Gao, Lu Gang, J.L. Pi, Z.B. Zhu and C.F. Zhang, J. Fuel Chem. & Tech.(Chinese),
21(1), 1993.
Rate model for phase separation of W/O emulsion in liquid surfactant membrane
process at high A.C. voltage
Lu Gang, Q.H. Luo and P.S. Li, Chinese J. of Chem. Eng., 1(4), 1993
134
Acknowledgement
This dissertation would never have been completed successfully without the whole-
hearted support of my thesis supervisors. I am very grateful to my first promoter,
Professor R. A. van Santen, who offered me the opportunity to pursue my PhD
program described in this thesis and provided valuable technical and personal
guidance. My special thanks also go to Professor van Veen, my copromoter, for his
vital support and generous guidance for the writing of this thesis.
I would thank all other reading committee members, Prof. L. Kapteijn, Prof. L.
Lefferts and Dr. J. Hoebink, who reviewed the manuscript and made many valuable
suggestions as well as offering many technical comments.
In the group of Laboratory of Inorganic Chemistry and Catalysis, many people have
provided their invaluable help. Many thanks are due to Dr. B.G. Anderson who spent
many hours on my experimental details, paper corrections and made contributions to
the project. It was so nice to work with you in BESSY of Berlin. Distinctive thanks
are due to Joop van Grondelle and Jos van Wolput, who are so kind to offer me
continuously a great deal of help, both technically and personally. The other
colleagues I should mention here are Arian, Pieter, Noud, Darek, Zhu, Ionel and
Nicolae. Sonia, thanks for the enthusiasm you put in the work during your stay in
Eindhoven. This was of great value for me.
I am indebted to those who also involved in the research work, especially to the group
of Prof. Brongersma for the LEIS measurement, W. Gennip for the XPS and Auger
measurement, P.J. Kooyman (National Centre for High Resolution Electron
Microscopy) for the TEM measurement, A. Knop-Gericke (Berlin, BESSY II) for
NEXAFS analysis.
135