Applied Catalysis B: Environmental 44 (2003) 117151

Formation and control of N2 O in nitric acid production

Where do we stand today?
J. Prez-Ramrez a, , F. Kapteijn a , K. Schffel b , J.A. Moulijn a
a Reactor and Catalysis Engineering, Delft University of Technology, Julianalaan 136, 2628 BL Delft, The Netherlands
b Hydrocarbon Processes and Catalysis, Oil and Energy Research Center, Norsk Hydro, P.O. Box 2560, N-3907 Porsgrunn, Norway
Received 23 September 2002; received in revised form 14 January 2003; accepted 19 January 2003

Abstract
Nitric acid production represents the largest source of N2 O in the chemical industry, with a global annual emission of
400 kt N2 O. The high impact of N2 O on the environment as greenhouse gas and stratospheric ozone depletor, and the ongoing
agreements and prospective regulations calls for the development of efficient and economical systems for N2 O mitigation, but
no mature commercial technology is yet available. In this review, the current state-of-the-art for N2 O control in the nitric acid
manufacture is presented. The formation of N2 O and the process are analyzed and several options for reducing its emissions
are discussed, depending on the position in the process. Primary abatement options deals with modifications in the ammonia
oxidation catalyst, secondary abatement with options between the ammonia converter and the absorber, tertiary abatement
with options in the tail-gas upstream of the expander, and quaternary abatement with options in the tail-gas downstream of the
expander. The abatement technologies are evaluated based on the technical advantages and disadvantages, and cost efficiency.
2003 Elsevier Science B.V. All rights reserved.
Keywords: Nitrous oxide; Laughing gas; N2 O; Nitric acid production; Abatement; Technology; Catalyst; Environment; Greenhouse gas;
Ozone layer; In-process-gas; Tail-gas

1. Environmental role of N2 O N2 O has been relatively constant for many centuries

Nitrous oxide (N2 O), also known as laughing
gas, has been long considered as a relatively harm-
less species and has suffered from a lack of interest
from scientist, engineers, and politicians. However,
during the last decade, a growing concern can be
noticed since N2 O is a harmful gas in our environ-
ment, contributing to the greenhouse effect and ozone
layer depletion. The atmospheric concentration of
Corresponding author. Present address: Oil and Energy Re-
search Center, Norsk Hydro, P.O. Box 2560, N-3907 Porsgrunn,
Norway. Fax: +47-35-92-4738.
E-mail address: javier.perez.ramirez@hydro.com
(J. Perez-Ramrez).
(270 ppbv). The present-day N2 O atmospheric con-
centration is 310 ppbv, which means a 9% increase
from pre-industrial levels (285 ppbv) at an annual
growth rate of 0.20.3% [13]. This mismatch be-
tween N2 O sources and sinks is due to anthropogenic
practices, especially during the second part of the
20th century (Fig. 1).
N2 O is the major source of NOx in the stratosphere,
and therefore an important natural regulator of strato-
spheric ozone [46]. The overall influence of N2 O on
the ozone layer is complex and very different from
that of substances covered by the Montreal protocol in
September 1987 (chlorofluorocarbons (CFCs), halons,
carbon tetrachloride, and methyl chloroform) [7]. A
great deal of uncertainty exists regarding the chemical

0926-3373/03/$ see front matter 2003 Elsevier Science B.V. All rights reserved.
doi:10.1016/S0926-3373(03)00026-2
118 J. Perez-Ramrez et al. / Applied Catalysis B: Environmental 44 (2003) 117151

Fig. 1. Evolution of the atmospheric N2 O concentration [3].

reactions involved. At this time, there are no doubts
on the negative effect of N2 O on the ozone layer, but
there is no consensus on a quantitative value for its
ozone-depleting potential [8].
N2 O also contributes to the greenhouse effect, a
phenomenon caused by strong absorbance of infrared
radiation in the atmosphere [9,10]. Although N2 O is
not the major contributor to global warming (6%)
[8], it is much more potent than either of the other
two most common anthropogenic greenhouse gases,
CO2 and CH4 . Due to its long lifetime of approx-
imately 150 years in the atmosphere, N2 O has 310
and 21 times the Global Warming Potential (GWP) of
CO2 and CH4 , respectively [11,12]. This makes that a
relatively limited emission (compared to other green-
house gases) is equivalent to about 10% of the CO2
emission. Until recently, strategies for addressing cli-
mate change have principally been focused on reduc-
ing emission from the main greenhouse gas carbon
dioxide, but the importance of other greenhouse gases

Fig. 2. Relative importance of the greenhouse gases in the EU for the fulfillment of the Kyoto target. Others: HFCs, PFCs, and SF6 . The
pie chart shows the distribution of the anthropogenic N2 O emission per sectors in the EU, averaged for the period 19901998 [14].
 J. Perez-Ramrez et al. / Applied Catalysis B: Environmental 44 (2003) 117151 119

and opportunities for their abatement have been more fossil fuels are also important sources. A rapidly in-
and more recognized in the last years. This culminated creasing source that also doubled between 1990 and
in an agreement at the Third Conference of the Parties 1998 is the transport sector, as a side-effect of the in-
(COP-3) to the United Nations Framework on Climate troduction of the catalytic converter to control NOx .
Change (UNFCC) in Kyoto (Japan) in December 1997 Control of N2 O emissions from agriculture is diffi-
to set legal binding targets for reducing emissions of cult to assess due to its diffuse data, leading to large
six greenhouse gases (CO2 , CH4 , N2 O, HFC, PFC, uncertainties of emission factors, on how various pa-
SF6 ) to be realized in the period 20082012 [13]. As rameters affect emissions, and scarce data on efficacy
a result, the EU agreed to reduce CO2 , CH4 , and N2 O of options to reduce emissions. N2 O emissions that
emissions by 8% of 1990 levels by 2010 (1995 levels can be reduced in the short term are associated with
for HFCs, PFCs, and SF6 ). Emissions targets for other chemical production and energy industry (35% in
industrialized nations are 6% for Japan and Canada the EU, Fig. 2) [11,1820]. This emission is concen-
and 7% for United States [14]. Unfortunately, not all trated in a limited number of large N2 O sources, which
the parties have ratified these targets as yet [15]. N2 O holds promise for an economic and efficient reduction
emissions account for 10% of the total greenhouse gas strategy to fulfil the Kyoto commitment. Emissions
emissions in the EU (Fig. 2), so its reduction plays an from chemical industry mainly apply to adipic acid
important role in reaching the global 8% target. and nitric acid production plants. Other newly iden-
tified sources are production plants of caprolactam,
glyoxal, acrylonitrile, and in general, processes us-
2. Sources of N2 O and emission control ing nitric acid as oxidizing agent or involving ammo-
nia oxidation [20,21]. Emissions from the latter N2 O
sources are less significant and not quantified as yet.
N2 O is produced by both natural and anthropogenic
Stationary combustion processes of coal (and in gen-
sources (Table 1). Biological processes in soils and
eral fossil fuels), biomass, and municipal and indus-
oceans are the primary natural source of N2 O. Within
trial waste also involves a significant N2 O emission,
human activities agriculture, through soil cultivation
and quantification is not accurate. Emissions from ve-
and the use of nitrogen-fertilizers, contributes to 57%
hicles also have a large uncertainty.
of the total emission in the EU (see Fig. 2). Chemi-
Prior to legislation a number of industries have vol-
cal production and the burning of organic material and
untarily initiated efforts to reduce N2 O emissions from
adipic acid production, with a global reduction from
Table 1 600 kt per year in 1994 to <100 kt per year currently
Global N2 O emissions [1618] (see Table 1). Different (thermal and catalytic) abate-
Type of source N2 O emissions
ment technologies have been successfully developed
(Mt N2 O per year) for this source, due the high N2 O concentration in
Natural 13 the tail-gas (typically 2540 vol.%) [11,18,22,23]. Due
Soils 10 to the exothermicity of the decomposition reaction, a
Oceans 2.9 large increase in the temperature occurs within the cat-
Atmospheric chemistry 0.2 alyst bed. For instance, the decomposition of 35 vol.%
Anthropogenic 7 N2 O in air leads to an adiabatic temperature rise of the
Agriculture (including fertilizers) 3.5 gas of 940 K. In this temperature window, a large
Nitric acid production 0.4 number of catalysts exhibit considerable activity. So,
Adipic acid production <0.1b
Fossil-fuel combustion (stationary)a 0.20.5
in this case the activity of the catalyst is not a critical
Fossil-fuel combustion (mobile)a 0.40.9 factor for the effectiveness of the technology.
Biomass combustion 1.0 Only in the particular case of adipic acid production,
Sewage treatment 1.5 technologies are commercially available, but their ap-
Total of all sources 20 plication (extrapolation or adaptation) in other sources,
a Emission is uncertain.
e.g. (single) nitric acid plants and stationary com-
b Reduced from 0.6. bustion processes, is not feasible [1820], due to the
120 J. Perez-Ramrez et al. / Applied Catalysis B: Environmental 44 (2003) 117151
dissimilar characteristics of the tail-gas:
diluted N2 O streams (in the 0.050.5 vol.% range),
relative low temperature (typically <800 K), and
presence of catalyst inhibitors (O2, H2 O, NOx , and
SO2 ).
An attractive variant of this abatement technology
is the integration of adipic acid and nitric acid plants
[24,25]. In this case, HNO3 is piped from the ni-
tric acid plant into the adipic acid plant. The adipic
acid tail-gas, containing around 20 vol.% of N2 O and
20 vol.% NO2 (plus CO2 , O2 , and N2 ) is pumped back
into the nitric acid plant. The loop is closed by absorb-
ing NO2 in water in the absorption column, recovering
nitric acid. What is left is a gas which is almost free
from NO and NO2 but still contains all of the N2 O
formed in the adipic acid process, which can be (cat-
alytically or thermally) decomposed in a unit inte-
grated in the nitric acid plant, downstream of the nitric
acid absorption column. This system has been suc-
cessfully implemented in the BASFs adipic acid and
nitric acid plants in Ludwigshafen (Germany).
A novel process option also for adipic acid plants,
conjointly developed by Solutia and the Boreskov In-
stitute of Catalysis, consists in the reutilization of N2 O
as a valuable oxidant to oxidize benzene to phenol
[26]. This process also takes advantage of the high
N2 O concentration in the tail-gas. The process has
been successfully tested in a pilot-plant facility erected
by Solutia in Pensacola (Florida, USA). The new sys-
tem is meant to replace the thermal off-gas treatment
currently used in this adipic acid plant. This applica-
tion is unfeasible in diluted tail-gases, e.g. in nitric
acid plants or combustion processes. Attempts to con-
centrate these streams by selective adsorption using
metal-exchanged zeolites has led to N2 O concentra-
tions of 5 vol.% [27], still insufficient to satisfy prac-
tical requirements.
3. Incentives for reducing N2 O emissions from
nitric acid production
During the manufacture of weak nitric acid (concen-
tration of 5065%), N2 O is produced as an unwanted
by-product of the catalytic oxidation of ammonia (see
Section 4). N2 O is currently emitted via the tail-gas
stack. Contrary to the well-developed technology for
de-NOx , i.e. recovery by extended absorption and
abatement by selective catalytic reduction (SCR) with
ammonia (see, e.g. [28]), no mature technology for
N2 O abatement has been installed as yet. Efforts in
developing de-N2 O technologies can be stimulated
by the following ways.
1. Moral pressure from other industries: Adipic acid
producers have led the way in making and ful-
filling a voluntary commitment to reduce N2 O
emissions to 100 kt N2 O per year. Now nitric acid
plants around the world collectively represent the
largest single source of N2 O in the chemical indus-
try (400 kt N2 O per year), and producers of weak
nitric acid (Table 2) can come under pressure to
develop and implement abatement technology.
2. New regulations: Even though specific emission
levels for N2 O have not been yet approved in
the EU countries, stringent limits probably are
expected in the near future. France has recently
adopted a pioneering attitude, imposing an emis-
sion tax of 37 per ton N2 O (0.5 per ton carbon)
[31]. The rate of this tax is quite low compared
with the cost of the proposed abatement options
(discussed later), but represents the commence-
ment of more severe regulations, and will surely
catalyze other governments to implement similar
measures.
3. Efficacy: In several European countries, e.g. Ger-
many, United Kingdom, and The Netherlands, re-
duction of N2 O emissions from nitric acid plants
would effectively contribute or even fulfill their Ky-
oto commitments, which represents an extra moti-
vation for development of technology. For instance,
application of the developed catalysts in the six ni-
tric acid production facilities in The Netherlands
(Table 3) could already contribute to the reduction
of 8.5 Mt CO2 -eq., which represents 35% of Ky-
oto commitment in this country.
4. Cost effectiveness: In general, the contribution of
non-CO2 greenhouse gases, especially N2 O, PFCs,
and HFCs is relatively high due to the promising
cost efficiency of the control measures [2933]. For
example, the cost of CO2 reduction measures by
changing coal-fired to gas-fired electricity plants
is 23 per ton CO2 -eq. removed, while the cost
efficiency of reduction of HFC is about 11 per
ton CO2 -eq. The cost of N2 O reduction measures
 J. Perez-Ramrez et al. / Applied Catalysis B: Environmental 44 (2003) 117151 121

Table 2
Major licensors of weak nitric acid production [29,30]
Licensor (former name) Country No. of plants in Europea,b

Technip (Grande Paroisse) France 30

Technip USA (KTI-Chemico) USA 8
Krupp Uhde Germany 51
Norsk Hydro Norway 19
Stamicarbon (DSM) The Netherlands 14
Giovanola (Bagmag) Switzerland 15
GIAP Russia 6 (>300 in Russia)
Montedison (Montecatini) Italy 4
Other known licensors, but with a small contribution (<4 plants) to the European market: Tecnicas Reunidas, Spain (former Espindesa);
Jacobs Engineering, UK (former Humphreys & Glasgow); Sumitomo Chemicals, Japan; Chematur Engineering, USA (former Weatherly);
Aventis, France (former Rhone-Poulenc); Dupont, USA.
a Capacities not considered.
b Excluding former Russia.

Table 3
Potential market volume in The Netherlands [29,30,32]
Owner/location Plant capacity (t HNO3 per day) N2 O emission (t N2 O per daya ) N2 O emission (kt
CO2 -eq. per daya )
Hydro Agri/Sluiskil 1620 12.2 3.8
Hydro Agri/Sluiskil 2030 15 4.7
DSM/Geleen 580 16.622.2 5.16.9
DSM/Geleen 1390
DSM/Ijmuiden 710 10.3 3.2
DSM/Ijmuiden 680
Total 7010 54.159.7 16.818.6
a Figures based on production (not on capacity).

at nitric acid plants varies between 0.2 and 3 per
ton CO2 -eq. (see Section 6.1).
In view of these incentives, the large N2 O emission
reduction potential in the nitric acid industry combined
with the finite number of sources and the potential cost
efficiency offers a very attractive means to meet the
targets set in the Kyoto agreement, and existing and
prospective regulations.
4. N2 O formation in nitric acid production
4.1. General considerations and practical aspects
In order to develop N2 O reduction technologies in
nitric acid plants, it is important to analyze the causes
for formation of this gas in the overall process and the
factors influencing its emission. Production of weak
nitric acid is based on the Ostwald process and consists
of the following basic chemical operations [34]:
Catalytic oxidation (combustion) of ammonia with
air into nitric oxide.
Oxidation of nitric oxide into nitrogen dioxide.
Absorption of nitrogen dioxide in water to produce
nitric acid.
Fig. 3 shows a typical flowsheet of a weak nitric
acid plant according to a dual-pressure configuration,
where these operations are integrated. The emission
of N2 O depends exclusively on the ammonia com-
bustion process. Once formed, N2 O passes unreacted
through the plant and is not affected by the operat-
ing condition in the absorber or the eventual de-NOx
SCR after-treatment in the tail-gases (not required in
modern plants like that in the figure).
 122
J. Perez-Ramrez et al. / Applied Catalysis B: Environmental 44 (2003) 117151
Fig. 3. Flowsheet of a dual-pressure nitric acid plant, with the potential location of the different abatement measures (see Section 5).
 J. Perez-Ramrez et al. / Applied Catalysis B: Environmental 44 (2003) 117151 123

Table 4
Reactions during oxidation of ammonia
Reaction H (kJ mol1 ) Keq

NH3 + 1.25O2 NO + 1.5H2 O (1) 227 3.2 1012

NH3 + O2 0.5N2 O + 1.5H2 O (2) 275 1.3 1012
NH3 + 0.75O2 0.5N2 + 1.5H2 O (3) 318 7.5 1015
NH3 + 4NO 2.5N2 O + 1.5H2 O (4) 469 3.1 1010
NH3 + NO + 0.75O2 N2 O + 1.5H2 O (5) 324 5.0 1011
NH3 + 1.5NO 1.25N2 + 1.5H2 O (6) 453 8.3 1020
NO 0.5N2 + 0.5O2 (7) 90 2.3 103
NH3 0.5N2 + 1.5H2 (8) +46 4.6 103
N2 O N2 + 0.5O2 (9) 82 6.9 109
N2 O + 0.5O2 2NO (10) +96 6.4 100
Thermodynamic data (enthalpy and equilibrium constant) at 1173 K and 1 bar. Enthalpy expressed per mol of NH3 except for reactions
(7) (per mol of NO) and reactions (9) and (10) (per mol of N2 O).

Ammonia oxidation has been traditionally con-
sidered as a classic example of strongly exothermic
mass-transfer limited reaction(s) with critical points
of heat extinction and ignition [3538]. In indus-
try, platinum-rhodium gauzes (typically 90% Pt and
10% Rh) catalyze this process. A catchment facil-
ity typically made of Pd recovers 6080% of the
evaporated PtO2 as metallic platinum. From 3 to 50
gauzes are stacked in a reactor (burner) with diam-
eters up to 6 m. Operating temperatures range from
1073 to 1223 K, with pressures from 1 to 6.5 bar
(in dual-pressure plants) and from 1 to 13 bar (in
mono-pressure plants). Contact time in the gauze
pack area is in the range of 103 to 104 s (space
velocity of 60,000600,000 h1 ), being one of the
shortest-contact time commercial processes known.
The catalytic oxidation of ammonia can follow
several paths, depending on the temperature, yield-
ing N-products with different oxidation degree, in an
increasing order: N2 , N2 O, and NO (Eqs. (1)(3),
Table 4). The reaction leading to N2 is the most favor-
able thermodynamically. From the calculated equi-
librium concentrations in the range of 4731273 K,
it can be concluded that the non-catalytic combustion
of ammonia would lead exclusively to nitrogen. In
the presence of catalyst, a more complex situation
arises. At low temperatures (423473 K), N2 is the
only N-containing product. At higher temperatures
N2 O formation starts, passing through a maximum (at
675 K) with increasing temperature. The formation
of NO, the desired product in nitric acid production,
starts at 573 K and its yield continuously increases
with temperature. Hence, towards deep oxidation not
only the catalyst activity but also the selectively are of
essential importance for the commercial application
of this process.
Under industrial conditions (see above), NO yields
of 9597% are typically achieved with Pt gauzes, de-
pending on the pressure and temperature [39]. The se-
lectivity towards N2 O and N2 is 1.52.5 and 44.5%,
respectively. Lower catalyst temperatures for a cer-
tain pressure tends to be more selective towards un-
wanted products. From an engineering standpoint, the
oxidation process is considered as well optimized
and further prevention of by-products formation by an
improved catalyst performance is expected to be lim-
ited. However, a slightly improved NO selectivity is
greatly attractive, not only due to the corresponding
lower N2 O emission, but also due to the increased pro-
duction of nitric acid. Roughly, an increased NO yield
of 1% in a medium nitric acid plant would lead to a
profit of k 500 annually.
Generally, the amount of N2 O formed depends on
combustion conditions, catalyst composition and state
(age), and burner design. Fig. 4 can be used to con-
vert between various ways of expressing levels of N2 O
emission as a function of the yield of the correspond-
ing reactions in Table 4. There is a considerable vari-
ation of the emission of N2 O per ton of nitric acid
produced, as the current internationally accepted Inter-
governmental Panel on Climate Change (IPCC) guide-
line figures for estimating national N2 O emissions
indicate (Table 5).
Undesired reactions leading to N2 O and N2 involve
not only the reactants (ammonia and air (Eqs. (2)
and (3)) but also consecutive reactions between
124 J. Perez-Ramrez et al. / Applied Catalysis B: Environmental 44 (2003) 117151

Fig. 4. Different ways to express the N2 O emissions and concentrations in nitric acid plants vs. the ammonia oxidation selectivity to N2 O.

unreacted ammonia with nitric oxide (Eqs. (4)(6)). but the heterogeneous rate exceeds that of the ho-
The economical impact of these loss reactions mogeneous route [41]. Palladium in the catchment
is double negative: not only the conversion effi- packs also catalyzes the latter reactions [17]. In
ciency of the reactant (NH3 ) is decreased, but also practice, a dependency of the extent of these reac-
the yield of the desired product (NO). These reac- tions with the contact time of the gas in the burner
tions may occur via both homogeneous (gas phase) has been identified. As a rule of thumb, high flow
or heterogeneous (platinum surface) mechanisms, rates and linear velocities of 1 m s1 are necessary
to minimize these loss reactions [36]. Decompo-
sition of NO (Eq. (7)), which is metastable at the
Table 5
converter temperature, and NH3 (Eq. (8)) should be
IPCC emission factors for N2 O from HNO3 production [40]
also minimized. These processes have a significant
Production process N2 O emission contribution at temperatures >1273 K and relatively
factor (kg N2 O
long contact times.
t1 100% HNO3 )
Canada
Plants without NSCR 8.5
4.2. Kinetics and mechanism of ammonia
Plants with NSCR <2a oxidation
USA
Plants without NSCR 9.5 Understanding the mechanism and kinetics of am-
Plants with NSCR <2a monia oxidation, in particular towards N2 O produc-
Norway tion, would be of primary relevance to prevent or min-
Process-integrated N2 O destruction <2 imize its formation.
Atmospheric pressure plant 45 Never has a reaction system been studied so long
Medium pressure plant 67.5 with ultimately so little consensus on a mechanistic
Other countries explanation. Roughly, selectivity in ammonia oxida-
Dual-pressure plant (European design) 810 tion is determined by competition between NH3 and
Older plants (pre-1975), without NSCR 1019 O2 molecules for active sites, and the intermediates
a Uncertainty 10%. surface concentration. At low temperatures (473 K),
 J. Perez-Ramrez et al. / Applied Catalysis B: Environmental 44 (2003) 117151 125

N-coverage dominates the catalyst surface leading Table 6

to N2 . Upon increasing temperature, the relative Elementary reaction steps for ammonia oxidation proposed by
different authors
O-coverage increases by dissociative adsorption of
molecular oxygen, and more oxidized products are Reaction Reference
formed. Under industrial conditions, the oxygen-co- NH3 + NH3 (13) [58]
verage is dominant, leading to NO. O2 + 2 2O (14) [58]
All attempts to suppress the critical regimes and to NH3 + O NH2 + OH (15) [58]
NH2 + O NH + OH (16) [58]
obtain intrinsic kinetics for concentrations and tem-
NH + O N + OH (17) [58]
peratures used in industry failed [3537,42]. These N + O NO + (18) [57,58]
kinetic problems disappear at temperatures <573 K, NO NO + (19) [59]
when N2 and N2 O are the main products. For ex- N + N N2 + 2 (20) [58,60]
ample, the chemical kinetics of the low-temperature OH + OH H2 O + O + (21) [58]
NH + O NO + H (22) [57]
ammonia oxidation on platinum wire at ammonia
NH + 2O NO + OH (23) [60]
pressures ranging from 0.05 to 0.2 bar and at oxy- NO N + O (24) [60]
gen pressure ranging from 0.2 to 0.95 bar concluded NH2 + NO N2 + H2 O + 2 (25) [61]
that the experimental data are well described by the NO + N N2 + O + (26) [61]
first-order kinetics with respect to the surface cov- NO + N N2 O + (27) [61]
N2 O N 2 O + (28) [62]
erage by oxygen and ammonia pressure, with an
activation energy of 125 kJ mol1 [43]. For a more represents an active site of the catalyst.
oxidized product, N2 O, the selectivity is proportional
to the surface coverage by oxygen. With increasing stable intermediates:
contact time, the N2 O yield decreases, which was
explained by its interaction with adsorbed ammo- NH3 + O NO + H2 + H (11)
nia on the catalyst surface. Whereas mass-transfer NH3 + NO N2 + H2 O + H (12)
phenomena determine the ammonia conversion un-
der the process conditions of industrial reactors, the N2 was believed to be formed in a bimolecular sur-
selectivity of NO formation is determined by the face reaction between NO and NH3 , with the transition
kinetics of the individual reactions in Table 4. The from N2 to NO formation occurring at a surface tem-
reaction rates obtained from laboratory measurements perature when NO has too short a surface residence
at low pressures (10100 Pa) have been typically time to take part in reaction (12), i.e. when NO starts
described by a LangmuirHinshelwood mechanism to desorb from the surface. Nutt and Kapur [55,56]
[44]. suggested a very similar reaction mechanism, but with
Before 1960, three different general reaction mech- dissociative adsorption of ammonia on the platinum
anisms on platinum were proposed on the basis of pos- surface.
tulation, because of the lack of experimental evidence The above models address only the overall reaction
of intermediate species: the nitroxyl (HNO) mecha- mechanism and do not go into the details of the ele-
nism [45], the hydroxylamine (NH2 OH) mechanism mentary reaction steps. More recent developments in
[46,47], and the imine (NH) mechanism [48,49]. surface science techniques over single crystals of plat-
These mechanisms propose that the final products inum yielded new insights on the reaction mechanism.
(NO, N2 , N2 O, and H2 O) are formed via a number of Table 6 summarizes the different elementary reaction
stages involving one of such intermediates products steps proposed in literature sources. Asscher et al. [57]
[4553]. used a laser multiphoton technique, combined with
Fogel et al. [54] used SIMS (secondary ions mass molecular beams under UHV conditions, to probe NO
spectrometry) to study the ammonia oxidation at formation during reaction on a Pt(1 1 1) crystal. By an-
0.01 Pa and 2981473 K over polycrystalline Pt wire. alyzing the decay in the NO signal after pulsing NH3 ,
They concluded that the N2 O was not formed during they identified two different NO production kinetics
the oxidation reaction and that the reaction proceeds (Eqs. (18) and (22)). They found Eq. (22) to be the
via a LangmuirHinshelwood mechanism without faster mechanism and enhanced at high O2 /NH3 ratios
126 J. Perez-Ramrez et al. / Applied Catalysis B: Environmental 44 (2003) 117151
and had a higher activation energy than Eq. (18), the
slower reaction.
More recently, Mieher and Ho [58], using tempe-
rature-programmed reaction spectroscopy (TPRS),
electron-energy loss spectroscopy (EELS), and low-
energy electron diffraction (LEED) over Pt(1 1 1)
concluded that the reaction proceeded via the simple
stripping of adsorbed NH3 by adsorbed oxygen atoms
(Eqs. (15)(17)), followed by the combination of ni-
trogen atoms with oxygen, to form NO (Eqs. (18)
and (19)), or with nitrogen atoms, to form N2
(Eq. (20)). Recombination of adsorbed hydroxyl
groups leads to water formation (Eq. (21)). Bradley
et al. [60] found similar results using surface molecu-
lar beams under UHV conditions over Pt(1 0 0). How-
ever, the production of N2 and NO was found to occur
through other paths. They suggested that NO was
directly produced by reacting adsorbed NH species
with adsorbed O species (Eq. (23)), in a similar way
as proposed by Asscher et al. [57]. The formation of
N2 results from the consecutive dissociation of NO
(Eqs. (20) and (24)). This pathway is supported by
recent results from van den Broek et al. [63] showing
that NH and OH are the main surface species ad-
sorbed on Pt sponge catalysts during ammonia oxida-
tion at atmospheric pressure. Other alternative routes
for N2 production, decreasing the NO selectivity, are
the reaction of adsorbed NO with atomic N-species
(Eqs. (25) and (26)). However, in none of the low
pressure (<100 Pa) investigations nitrous oxide was
ever detected and thus formation of N2 O was not dis-
cussed, so no evidence for Eqs. (27) and (28) has been
gathered.
Concluding, although the mechanism of ammonia
oxidation, and more specifically the steps involved in
N2 O formation, has not been unambiguously unrav-
eled, it appears that N2 O (and probably N2 ) is not
a primary product of the reaction between ammo-
nia (or derived radical species (NHx )) and oxygen
on the catalyst surface. Current mechanistic inves-
tigations in vacuum using a Temporal Analysis of
Products (TAP)-reactor are contributing to further
elucidate this complex mechanism. The uncertainty
of extrapolating results from vacuum experiments
to the real process always exist, but it is an opti-
mal alternative in view of the experimental issues
involved in investigations under industrial conditions
[64].
5. N2 O abatement options during nitric acid
production
Development of N2 O abatement systems aims at
the achievement of high efficiency (>90% N2 O con-
version) and selectivity (<0.2% NO loss). The ap-
proaches followed by industry, research institutes, and
universities can be classified in four groups, according
to the position in the process (see different locations
in Fig. 3):
Primary abatement measures aim at preventing
N2 O being formed in the ammonia burner. This
involves modification of the ammonia oxidation
process and/or catalyst.
Secondary abatement measures remove N2 O from
the valuable intermediate stream, i.e. from the
NOx gases between the ammonia converter and
the absorption column. Usually this will mean in-
tervening at the highest temperature, immediately
downstream of the ammonia oxidation catalyst.
In tertiary abatement measures, the tail-gas leaving
the absorption column is treated to destroy N2 O.
The most promising position for nitrous oxide
abatement is upstream of the tail-gas expansion
turbine.
Quaternary abatement measures are the pure end-
of-pipe solution, where the tail-gas is treated down-
stream of the expander, on its way to the stack.
The technologies have been evaluated attending
to different levels of development: microlevel, aim-
ing at a suitable (catalytic) formulation, mesolevel,
which mainly focuses on the reactor selection, and
macrolevel, regarding the integration of the unit in
the plant.
5.1. Primary abatement measures
5.1.1. Optimal oxidation process
Optimizing the catalytic oxidation of NH3 to NO
both from reactor design and operation for a certain
type of plant can be considered as a preliminary
measure for a reduced N2 O emission. Good practices
for performing this practices have been collected
elsewhere [29,39]. The ammonia-to-air ratio in the
feed should be 1:9 (i.e. 10 vol.% NH3 ) for an op-
timal NO yield. Temperatures above 1173 K with a
relatively fresh gauzepack will lead to a lower N2 O
 J. Perez-Ramrez et al. / Applied Catalysis B: Environmental 44 (2003) 117151
production, but catalyst losses due to platinum va-
porization increase dramatically, which in fact reduce
the NO yield and increase by-product formation. Low
operation pressures in the burner yield high NO yields
(due to thermodynamics), but at the expenses of low
nitric acid productions. A compromise is operation
at 45 bar, as in most modern plants. A good mixing
of ammonia and air, and an even distribution of the
mixture across the gauze catalyst is also essential to
achieve good performances. Air pollution or contam-
ination from the ammonia can poison the catalyst.
This effect, as well as poor ammonia-air mixing and
uneven gas distribution across the catalyst gauze,
may reduce the yield by 10%. Of course, the plat-
inum gauze should be replaced regularly to keep an
optimum performance (312 months, depending on
the process variant). The noble metal losses cause a
reduction in mechanical strength of the catalyst gauze
(limiting the lifetime of the catalyst) and a gradual
increase of the rhodium content on the catalyst sur-
face. This leads to decreased activities and a higher
production of N2 O and N2 , at the expense of a lower
NO yield.
On the long-term, an optimal reactor design could
reduce N2 O formation and improve NO efficiencies.
An interesting approach is the use of reversed flow op-
eration, by which the dynamics of the catalytic process
(adsorption, desorption, and surface reaction rates) can
be manipulated using a dedicated procedure of flow
reversals in transient operation, aiming at optimal se-
lectivity and conversion with a significant energy sav-
ing. These reactor concepts will be applied only in
new plants due to the high retrofit costs.
A more efficient approach for existing plants is di-
rected to the improvement of the geometry of Pt-Rh
gauzes to optimize the NO yield or replacement of the
Pt-Rh gauzes by oxide-based catalysts.
5.1.2. Modification of the Pt-Rh gauzes
Every additional mol of NH3 converted to NO in the
platinum catalyst is a mol less that can be converted
to N2 O. Thus, measures that increase the conversion
efficiency of the catalyst gauzes will reduce N2 O for-
mation (position 1 in Fig. 3). The efficiency of Pt-Rh
gauzes in ammonia oxidation is mainly characterized
by parameters selectivity, noble metal losses, and life-
time. This is generally dependent on a large number
of (complex) interconnected factors, such as the spe-
127
cific operating parameters of a particular plant, as well
as the gauze catalyst geometry and configuration. The
gauze and/or gauzepack structure have a significant
impact on the heat and mass-transfer, as well as on
the hydrodynamic flow and gas mixture distribution
[44,65,66].
The use of Pt and later Pt-Rh alloys as a cata-
lyst changed little between 1909 and 1990. The cat-
alyst was originally manufactured as woven gauzes
(Fig. 5a). Johnson Matthey introduced knitted gauzes
in 1990, which have practically displaced the tradi-
tional woven gauze from the market, and has now
become the standard in industry (95% of the total
production). Several knitted gauzes with different ge-
ometry can be seen in Fig. 9bd, which are recom-
mended by manufacturers in view of their multiple
advantages compared to woven gauzes (see Fig. 5 and
[6770]). The three-dimensional geometry and high
porosity of knitted gauzes (up to three times higher
than woven gauzes) ensures an even temperature pro-
file across the catalyst. The improved heat distribution
along the gauzepack reduces the metal loss, com-
pared to woven gauzes (large temperature gradients),
and the resulting deactivation is slower, thus leading
to extended campaign lengths. A higher net NO se-
lectivity (therefore a lower N2 O production) has also
been claimed by manufacturers, although no quantifi-
cation is available.
Recently, new three-dimensional configurations
(e.g. Multinit from dmc2 ) have been proposed as
being even more effective than conventional knitted
gauzes [69]. Two models are shown in Fig. 6, where
two dense mesh-layer at the top and bottom are con-
nected by so-called pile (and weft) threads. The low
porosity of the mesh-layers results in higher reaction
rate than in conventional knitted gauzes, while the
improved heat conduction between the pile threads
prevents the presence of large temperature gradients
between gauze layers, reducing the precious metal
loss and catalyst deactivation. Weft threads placed in
the middle of the structure form a third, dense, and
highly active location for the reaction. As a result,
the high activity of woven gauzes with the low pre-
cious metal losses of knitted gauzes are combined.
As there are many varieties of knitting stitch and
configurations, without doubt all the possibilities for
improvement have not yet been exhausted and thus
the potential N2 O reduction factor is still uncertain.
128 J. Perez-Ramrez et al. / Applied Catalysis B: Environmental 44 (2003) 117151

Fig. 5. Commercial Pt-Rh catalysts of different geometries for NH3 oxidation: (a) woven gauze, and (bd) knitted gauzes [67]. Advantages
of knitted geometry according to manufacturers are listed in the figure.

5.1.3. Oxide-based combustion catalysts
The high costs of the Pt-Rh gauzes, subsequent
metal losses during operation, as well as metal recov-
ery and refining systems have stimulated the research
towards substitution of Pt gauzes by oxide catalysts.
Oxide catalysts offer the advantages of lower invest-
ment and longer campaign length compared to pre-
cious metal catalysts, as well as a simpler manufacture
[17,35,42]. Another important reason for the revival of
interest is the low N2 O emission generated over this
catalyst, since metal oxides, e.g. Co3 O4 , are active in
decomposing it at the burner conditions. Recent pub-
lications reported a conversion rate of NH3 to N2 O
as low as 0.5% over Co3 O4 catalysts with or without
addition of low amounts of cerium and lanthanum ox-
ides as promoters (stabilizers) [71,72].
A vast number of patent applications have claimed
a relatively high NO selectivity (9095%) using sin-
gle metal oxides or mixed oxides (perovskites and
spinels) based on Co, Fe, Bi, Mn, Cr, and Ni [42].
However, most of the tests were carried out in lab-scale
reactors with powder or small pellets and not under
conditions of industrial relevance, thus being of lim-
ited value. In addition, no data on catalyst stability
has been reported. In particular, Co-based systems
have been studied more intensively. Some examples
from the past are the commercial catalysts by Hedler
(cobalt-oxide stabilized by various admixtures) [73],
African Explosives (cobalt-oxide promoted by scan-
dium, yttrium, and other rare earth metals) [74,75], and
ICI (cobalt-oxide supported on magnesiumaluminum
spinel) [76].
There are serious drawbacks of oxide-based catal-
ysts that inhibit the replacement of the well-established
and proven Pt-Rh gauze catalyst. A crucial aspect is
the lower ammonia conversion efficiency to NO of
Co3 O4 (8892% in high-pressure plants) compared
with the Pt-Rh gauzes (9597%), which has been
attributed to the larger sorption capacity of oxides
[42,77]. Taking into account that under similar oper-
ating conditions, on the most active oxides the ammo-
nia conversion is complete at contact times of 102 s,
while on platinum gauzes it is achieved at contact
times of 104 s, the specific activity of oxide catalysts
 J. Perez-Ramrez et al. / Applied Catalysis B: Environmental 44 (2003) 117151
Fig. 6. Structure of novel Multinit gauzes, combining the prop-
erties of knitted and woven gauzes [69].
appears to be at least 2 orders of magnitude lower
than that of the platinum group metals. Indeed, the
apparent first-order rate constant is 24,600 s1 for the
platinum catalyst and 73 s1 for the iron oxide catalyst
[77]. Another disadvantage is the reversible deactiva-
tion of Co-systems, due to reduction of Co3 O4 by NH3
to the inactive CoO under reaction conditions in the
upper parts of the bed [78]. Lower activities can also
be associated to the sintering of the cobalt-oxide phase
and changes in the porous structure under reaction
conditions [79,80]. Furthermore, the optimal operating
temperature of Co3 O4 is >100 K lower than for Pt-Rh
catalysts, due to the lower heat of oxygen adsorption in
on-top position of the regular centers of cobalt-oxide
compared to platinum. This could result in difficul-
ties with the steam balance in a revamped plant [17].
The design of the ammonia burner is a crucial aspect
for the potential replacement of Pt gauzes by oxide
catalysts, and major changes will disable retrofit in
existing plants, making its application prohibitive.
Other disadvantages of oxide systems are the leaner
NH3 /air mixtures (<10 vol.%), a higher-pressure
drop, and also the higher poisoning susceptibility to
129
sulfur compounds in the air [35]. The drawbacks of
oxide systems have hitherto not been compensated
by the advantages mentioned above, although several
companies (e.g. PCS Nitrogen, JD Technologies, Joy
Industries) are commercializing cobalt-oxide systems
for ammonia combustion (see, e.g. [81]).
With regard to industrial process economics, which
is primarily determined by the nitric acid yield and
possibility to use the same reactors without serious
revamping, it was proposed to use oxide catalysts
together with a fewer number of platinum gauzes.
The combined bed idea was proposed by GIAP
(State Institute of Nitrogen Industry in Russia) in the
1950 s, with the installation of a dual-bed system (Pt
gauze + iron-chromium catalyst) in an atmospheric
pressure plant [82]. In the 1990s, GIAP in coopera-
tion with Boreskov Institute of Catalysis, have led to
an oxide catalyst based on FeMn oxides, in form of
granules (515 mm) or extruded monolith [42,83]. At
present, the dual-bed system is successfully operated
in 17 converters (mostly at high pressure) of different
plants in NIS and in other countries where the ammo-
nia oxidation technology developed in the USSR is
used. NO yields of 9495.6% (and no NH3 slip) have
been reported in high-pressure plants and 98% in
atmospheric plants, with no NH3 slip at 11231173 K
and campaign lengths of 1.5 years. These high yields
were achieved with a reduced platinum metal loading
(2025%), leading to decreased noble metal losses
(reduction of 15%) [42]. Nothing has been mentioned
concerning the N2 O emissions in this dual-bed system
compared to the pure platinum gauze bed.
5.2. Secondary abatement measures
The process options described here focus on the
decomposition of N2 O, either by thermal decompo-
sition or by catalytic decomposition, typically in the
high-temperature zone immediately downstream of the
ammonia burner (position 2 in Fig. 3).
5.2.1. Homogeneous decomposition in the burner
The extended chamber concept is based on the ther-
mal decomposition of N2 O in the hot zone downstream
of the ammonia combustion Pt-Rh gauzes at temper-
atures ranging from 1023 to 1223 K and pressures of
113 bar (Fig. 7a). This involves a novel design of the
ammonia oxidation unit so that the residence time of
130 J. Perez-Ramrez et al. / Applied Catalysis B: Environmental 44 (2003) 117151
Fig. 7. Secondary abatement measures: (a) extended reaction chamber for homogeneous N2 O decomposition, and (b) in-process-gas catalytic
N2 O decomposition.
the process-gas in the burner is extended ( = 12 s)
and the N2 O, which is metastable at these temper-
atures, may decompose. This technology, developed
and patented by Norsk Hydro [84], was successfully
implemented in the latest nitric acid plant of this com-
pany in Porsgrunn (Norway) in 1990, resulting in an
N2 O level in the tail-gas of about 200 ppmv (70% N2 O
reduction) [85].
The increased residence time in the reaction cham-
ber for decomposition requires extended space be-
tween the gauzes and the superheater. This process
option leads to increased capital costs of the plant by
about 56%, with no increase in operational costs.
However, it appears to be prohibitive for retrofitting,
due to the high costs involved in the redesign of the
existing ammonia burners. It is highly unlikely that
any new nitric acid plant will be built in Europe before
2010, so this option will not be implemented in this
decade. In the longer term, however, it could ensure
that all new plants have low levels of N2 O emissions.
Fig. 8 shows the simulated N2 O concentration pro-
file in the decomposition chamber for the case of an
ideal adiabatic reactor (no heat loss, dashed profile)
and for the more realistic case of a temperature gra-
dient along the decomposition chamber. Operational
input conditions (gas flow, pressure, temperature, and
temperature gradient) used in this simulation are rep-
resentative for modern dual-pressure plants. It is evi-
dent that high conversion levels (>90%) are difficult
to achieve. In addition, the method is not suited for at-
mospheric pressure plants as shown in the figure, due
to the low residence time of the gas in the reactor.
5.2.2. Catalytic decomposition in the burner
Instead of increasing the reactor chamber, the re-
duction of N2 O can be promoted using a catalyst for
selective N2 O decomposition. The homogenous de-
composition is a rather slow reaction under the prac-
tical conditions in the ammonia chamber (see Fig. 8).
This process can be accelerated by introducing a cat-
alyst behind the Pt-Rh gauze and the Pd catchment,
which enhances the reaction rate substantially such
that the necessary reactor volume can be reduced com-
pared to the extended reactor concept (Fig. 7b). The
inert Raschig rings, typically supporting the Pt-Rh
gauzes, can be replaced by the catalyst. Some plants
do not have a Raschig ring support, and baskets for
the catalyst have to be installed at minor costs.
In principle, this cost-efficient process option can be
easily applied in existing plants. The catalyst can be
introduced without making noticeable modification to
the existing burner. However, there are challenges for
the development and application of an in-process-gas
catalyst.
 J. Perez-Ramrez et al. / Applied Catalysis B: Environmental 44 (2003) 117151
Fig. 8. Homogeneous decomposition of N2 O as a function of the position in the reactor for an average modern dual-pressure plant in
adiabatic conditions, incorporating heat loss based on measured values, and in an atmospheric pressure plant [85].
5.2.2.1. Catalyst activity. The volume of the
Raschig ring layer in the space beneath the Pt-Rh
gauze (Fig. 7b) determines the available catalyst vol-
ume and thus the N2 O reduction. This volume depends
on the specific burner design. A typical catalyst bed
height ranges from 5 to 20 cm, which can be translated
into a gas-hourly space velocity (GHSV) in the order
of 50,000180,000 h1 . The figure means that in 1 s a
gas volume is passing through the catalyst bed which
corresponds up to 50 times the volume of the catalyst
bed. Since the reactor volume and gas flow are given
(and thus GHSV), the effective rate constant keff (as-
suming plug-flow behavior and first-order kinetics) is
a function of the required N2 O conversion (X(N2 O)):


1
keff = GHSV ln (29)
1 X(N2 O)
Since the target conversion is typically set to >90%,
an effective rate constant in the order of 100 s1 is
required. At the high temperature of the ammonia
burner, the catalytic N2 O-decomposition reaction will
be highly mass-transfer controlled. In principle, the
high-pressure drop will rule out the application of a
packed bed, especially if small pellets (35 mm) are
used to achieve a higher catalyst effectiveness. Struc-
tured catalysts, e.g. monoliths, pair a low-pressure
131
drop to a high catalyst effectiveness. Even under
mass-transfer controlled conditions, the high effective
rate constant, keff (=kf a ), required can be achieved
under gas-phase operation. In this expression, a
the specific surface area of the catalyst (typically
10003000 m2 m3 for monoliths) and kf is the ex-
ternal mass-transfer coefficient (m s1 ). The latter
requires values less than 0.1 m s1 , which is easily
achievable.
5.2.2.2. Catalyst selectivity. An important issue
concerning process economy of the secondary cat-
alytic abatement is the potential loss of product
(HNO3 ) through the catalytic decomposition of NO
(Eq. (7) in Table 4) in the ammonia burner. The
in-process-gas N2 O-decomposition catalyst should
be inert towards the decomposition of NO. This is
however unlikely because NO is, as N2 O, thermody-
namically unstable with respect to decomposition into
N2 and O2 . The importance of product loss is exem-
plified in Fig. 9 for a nitric acid plant of 1000 t HNO3
per day. Due to normal changes in operational condi-
tions, it is difficult to identify NO losses below 0.5%,
with an annual economical loss of k 250. In order
to properly evaluate the feasibility of this abatement
option, it is essential to know the exact NO losses for
132 J. Perez-Ramrez et al. / Applied Catalysis B: Environmental 44 (2003) 117151
Fig. 9. Annual losses in k for a nitric acid plant of 1000 t HNO3 per day as a function of the NO losses in the in-process-gas
N2 O-decomposition catalyst [85]. Detection limit in commercial plants is 0.30.5%.
a given catalyst, since this will determine its impact
on the process economy.
5.2.2.3. Pressure drop, catalyst (mechanical) stabil-
ity, and product quality. The low catalyst utilization
obtained with catalyst particles in fixed-bed calls for
small catalyst particles, but these would lead to a se-
vere pressure drop at the high gas-hourly space ve-
locities applied. The mechanical construction of the
burner may set some hard-limits in this respect, since
the pressure drop generates an extra-load in addition
to the weight of the catalyst. The particles should also
have a minimum mechanical stability (crush-strength)
in order to be properly handled during installation
and operation (thermal expansion or contraction of the
burner during plant shut-down and start-up). Break-
age or abrasion of particles may also lead to an in-
creased pressure drop and dust formation, resulting
in the contamination of the acid product. This could
eventually lead to serious safety problems. These oper-
ational constraints can be effectively tackled by using
a monolithic-type reactor. Due to its open structure,
pressure drop is minimal, providing also remarkable
mechanical properties.
5.2.2.4. Catalyst (chemical) stability. The chemical
stability affects the level of N2 O reduction and the
NO selectivity. High-temperature catalytic processes
such as the catalytic reaction exhibit sintering phe-
nomena, which reduce the available surface area for
the chemical reaction resulting in a decrease of activ-
ity over time. The challenge is to develop materials
with are stable under the severe reaction conditions
(no chemical phase segregation). Efforts regarding cat-
alyst development are concentrated on perovskite or
spinel-like mixed oxides (vide infra). The high tem-
perature and composition of the process-gas (16 vol.%
H2 O, 10 vol.% NO, and 7 vol.% O2 ) put severe de-
mands on the catalyst, and thus more catalytic tests
are required to determine the long-term mechanical
stability, activity and selectivity.
5.2.2.5. Status of the in-process-gas N2 O decomposi-
tion. Several companies have made intensive efforts
to develop and install a catalyst in the space beneath
the ammonia oxidation gauzes in the burner. Build-
ing on their success in the catalytic abatement of N2 O
emissions from adipic acid plants, BASF has evalu-
ated promising catalyst to the conditions in nitric acid
plant ammonia converters. A CuZnAl spinel was
originally tested in lab-scale with artificially created
process-gases [86]. In March and July 1999, a sup-
ported oxide catalyst (CuO/Al2 O3 ) in the shape of
small pellets (denoted as O3-80) was successfully in-
stalled in commercial production plants of the com-
pany. Fig. 10 shows the procedure used for the catalyst
 J. Perez-Ramrez et al. / Applied Catalysis B: Environmental 44 (2003) 117151 133

Fig. 10. Installation of N2 O-decomposition catalyst (O3-80) in the ammonia burner in nitric acid plant [87]. Workmen stepping on the
(fragile) pellets will clearly induce catalyst breaking, leading to high-pressure drops during operation and dust formation, which will
eventually end up in equipment downstream or in the nitric acid product.

(pellets) installation in an industrial burner. In October
2000, BASF were able to report an 8090% reduction
in N2 O from 1000 to 100200 ppm in an atmospheric
combustion plant in Ludwigshaven (Germany) and a
70% reduction from 1200 ppm (and 1400) to 500 ppm
(and 250) in two burners of the 5.5 bar combustion
section of a dual-pressure plant in Antwerp (Belgium)
[87]. These plants are now running for 3 years with no
significant deactivation. NO losses were undetected,
i.e. <0.5%. An extremely important safety problem
may arise if copper leaches from the catalyst, ending
up in the fertilizer product, since copper catalyzes am-
monium nitrate decomposition.
Norsk Hydro is also active in the development of
a highly active (N2 O conversion >90%), selective
(NO loss <0.5%), and stable in-process-gas catalyst
[85]. More than 600 formulations based on spinel
and perovskite materials were tested in a multi-flow
laboratory reactor (up to 5 g of catalyst) under simu-
lated conditions. In a recent patent, the performance
of different supported and unsupported cobalt-oxide
spinels with composition Co3x Mx O4 (M = Fe, Al,
Co and x = 00.2) have been reported [88]. An ex-
ample is given in Fig. 11. Unsupported Co-systems
undergo fast deactivation. The use of MgO as catalyst
support leads to a deactivating system, probably due
to solid-state reactions at the burner conditions (e.g.
disappearance of active Co and formation of CoMg
solid solutions). Most promising results in terms of
activity, stability, and selectivity were obtained using
CeO2 as catalyst support for cobalt.
The most promising formulations in lab-scale were
tested in a pilot-plant of the company (up to 5 kg of
catalyst). This set-up is a unique test facility, being
attached to the full-scale nitric acid plant, and thus
giving the opportunity to evaluate catalytic perfor-
mance under identical real-plant conditions. Further-
more, it includes an accurate analysis system allowing
to study NO losses with an accuracy of 0.10.2% (ac-
curacy on commercial scale is 0.30.5% (see Fig. 9))
[85]. Norsk Hydros first generation catalyst, NH-1
satisfied the R&D criteria and was implemented in
a medium-sized plant of the company in Porsgrunn
(Norway) in March 2000 [85]. The pilot-plant results
were completely transferable to the real reactor in the
process, with a remarkable high activity and stability.
A catalyst lifetime of >5 years has been estimated for
this catalyst, and no adverse impact on the plant oper-
ation and plant performance was found. No leaching
of metal in the catalyst to the final product has been
detected. The performance of NH-1 was further im-
proved by Norsk Hydros second generation catalyst,
134 J. Perez-Ramrez et al. / Applied Catalysis B: Environmental 44 (2003) 117151

Fig. 11. Lab-scale testing for in-process-gas N2 O decomposition: N2 O conversion vs. time-on-stream of different Co-based catalysts [88].
Conditions: 1200 ppm N2 O, 1 vol.% NO, 4 vol.% O2 , 1.7 vol.% H2 O, balance N2 ; T = 1163 K; GHSV = 560,000 h1 ; P = 3 bar.

NH-2, aiming at lower NO losses (<0.2%) and im-
proved mechanical stability. This catalyst is currently
undergoing plant tests with successful results.
Hermsdorfer Institut fr Technische Keramik
(HITK) has proposed an active catalyst using bulk
metal oxides of transition metals (Cr, Mn, Fe, Cr,
Ni, and Cu) and lanthanum [89]. More recently
supported systems have been proposed, using mag-
nesium oxide (e.g. Co3 O4 /MgO, LaCoO3 /MgO, or
La1x Srx CoO3 /MgO). These materials showed 100%
N2 O conversion at temperatures above 1073 K in
lab-scale set-up with artificially created tail-gases
[90]. No data on catalyst stability was reported, but in
view of Norsk Hydros results with MgO-supported
systems (see Fig. 11), poor stabilities would be ex-
pected.
DuPont claimed a high catalytic activity of ZrO2 in
the form of cylindrical pellets installed as a fixed-bed
beneath the gauzes and catchment system [91]. Re-
cently, a modified formulation has been developed,
Fe/ZrO2 [92]. Catalytic tests with trilobes of 3 mm
in lab-scale were carried out using high feed N2 O
concentrations (10 vol.%), more similar to tail-gases
of adipic acid plants, and in the absence of gaseous
inhibitors. The size of the catalyst would presum-
ably lead to a high-pressure drop at typical burner
gas-loads.
In 2000, the German company L&C Steinmller
(BB Power), well-known as a supplier of waste-heat
boilers for nitric acid plants, patented the installation
of a wide range of N2 O abatement catalysts below the
ammonia oxidation gauzes [93]. A complex catalyst
composition was used (containing up to 11 metals) to
decompose N2 O in simulated tail-gases in lab-scale,
although no specifications (composition and pressure)
were given. The original patent claimed the ability
of the catalyst to oxidize N2 O (presumably to NO),
but this was disproven by BASF, who bought L&C
Steinmllers patent.
Oxidation of N2 O to NO (Eq. (10) in Table 4) would
be in any case an elegant solution, reducing the en-
vironmental impact of N2 O and increasing the HNO3
production simultaneously. This reaction is endother-
mic and although thermodynamic favorable, the equi-
librium constant is very low at the burner conditions.
Grande Paroisse claims to have developed a formu-
lation, consisting of a refractory oxide mixed oxide
(ZrO2 and Al2 O3 ) impregnated with a zirconium salt,
converting (partially) N2 O into NO [94]. From prelim-
inary lab-scale tests using simulated tail-gas at 1123 K,
 J. Perez-Ramrez et al. / Applied Catalysis B: Environmental 44 (2003) 117151
N2 O conversions of 7899% were obtained with NO
selectivities of 15% (N2 selectivity 85%). How-
ever, the degree of uncertainty of this option is impor-
tant. According to this company, the system requires
longer residence time of the process-gas in the burner
compared to those described above, which is limited
by the empty space available below the gauzes for cat-
alyst installation. This might involve major redesign
of the burner and higher catalyst cost, which is not
compensated for the extra HNO3 produced.
A ranking between the different manufacturers
would situate BASF and Norsk Hydro in a good mar-
ket position, since their catalysts are commercially
proven thus having less uncertainty. The danger of
copper leaching in the BASF catalyst is a major
drawback for commercialization. Though, most of the
patents and processes proposed in this application are
concerned with fixed-bed reactors, but in view of the
features of the process-gas and the feature of struc-
tured catalysts discussed above, monolithic reactors
are the most logical solution. This indicates options
for process development and optimization.
5.2.3. Catalytic decomposition downstream of the
burner
A novel process for secondary N2 O abatement has
been proposed by Krupp Uhde, consisting of the in-
stallation of a catalyst downstream of the ammonia
burner, in the line to the absorption column [95].
In this case, N2 O is catalytically decomposed using
iron-loaded zeolites (MFI, BEA, MEL, FER, MOR)
in the process-gas, i.e. very similar same catalytic
systems as used in tertiary abatement. Activity tests
in lab-scale have shown 100% N2 O conversions at
temperatures above 773 K over Fe-zeolites, but the
feed composition strongly differed from a real situa-
tion (7 vol.% H2 O and ppm levels of NO in the patent
versus 13 vol.% H2 O and 10 vol.% NO in the plant).
Furthermore, in this process line there is a complex
heat exchange systems with optimized operation con-
ditions (see Fig. 3), so the position of this catalytic
abatement unit is critical and space-limited. This op-
tion clearly interferes with the production process, so
revamping this option is by no means obvious in ex-
isting plants. In a real situation, the high steam con-
tent and relatively high temperature will lead to severe
dealumination of the zeolite framework during opera-
tion, which can seriously affect the catalyst stability.
135
Table 7
Composition of the tail-gas at the outlet of the NO2 absorber (N2
balance gas) for a nitric acid plant of 1500 t HNO3 per day and
a tail-gas flow of 200,000 Nm3 h1 [39]
Gas Concentration Typical value
NOx (ppm) 1003500 200a
N2 O (ppm) 3003500 1500
O2 (vol.%) 14 2.5
H2 O (vol.%) 0.32 0.5
a Modern nitric acid plants.
In addition, these zeolites may catalyze undesired re-
actions involving NO. The same catalyst can be used
to decompose N2 O in the tail-gas (see Sections 5.3.2
and 5.3.3) having numerous advantages: no process
interference, simplicity, similar operation temperature,
higher chemical stability due to milder conditions, i.e.
a significantly lower investment.
5.3. Tertiary abatement measures
Implementation of N2 O abatement in the tail-gas
train upstream of the tail-gas expander (position 3 in
Fig. 3) offers the great advantage of not influencing
the heart of the nitric acid plant, i.e. the ammonia
burner and the absorber, and can be compared with
the implementation of a de-NOx SCR unit.
The tail-gas, containing N2 O, NOx , O2 , and H2 O,
leaves the absorber column at temperatures close to
298 K. The composition of the gas at the outlet of the
absorber during stable operation may vary within the
limits shown in Table 7. The tail-gas passes various
heat recovery stages on its way to the expander. The
tail-gas expander recovers some energy and the ex-
panded gas is subsequently vented. The pressure and
especially temperature of the tail-gas are essential pa-
rameters to determine the optimal process option for
tertiary N2 O abatement. This is mainly influenced by
the process variant and the operating conditions in the
NH3 burner. The temperature also depends on the po-
sition in the tail-gas train (see Fig. 3). Typical tail-gas
temperatures range from 523 to 773 K at pressures of
313 bar. Low-temperature tail-gases are present in
(old) low-pressure plants (mono or dual). High pres-
sures in the ammonia burner lead to a high tail-gas
temperature. Accordingly, large dual-pressure or mod-
ern high-(mono)pressure plants lead to temperatures
136 J. Perez-Ramrez et al. / Applied Catalysis B: Environmental 44 (2003) 117151

Table 8
General characteristics and tail-gas conditions of the six Dutch (weak) nitric acid plants in 1996/1997 [29,30,32]
Plant

A B C D E F

N2 O concentration (ppm) 1250 1500 1150 1250 920 1500

Capacity kt1 HNO3 per year 584 500 210 730 255 245
Emission kt1 N2 O per year 4.4 4.56 1.52 5.4 1.7 2
Process variant Dual Mono Mono Dual Mono Dual
P upstream of the expander (bar) 10 11 5 11 4 10
T upstream of the expander (K) n.a. 773 723 n.a. 673 n.a.
T downstream of the expander (K) n.a. 220 220 n.a. 220 n.a.
H2 O concentration (vol.%) n.a. 0.5 1 n.a. 1 n.a.
n.a.: not available.

in the range of 673773 K (and even higher). Table 8
exemplifies the tail-gas characteristics of the six nitric
acid plants operating in The Netherlands. In this case,
high-temperature tail-gases are generally present,
which facilitates the development of tertiary options
for N2 O abatement (see below).
Tertiary measures are more flexible with respect to
the size of the reactor than secondary measures, en-
abling the achievement of high levels of N2 O removal
(>99%). A producer might even decide to install a
more generously sized reactor than that immediately
required, allowing for the addition of extra-catalyst to
cope with prospective N2 O emission regulations, or to
optimize the catalyst requirement against the level of
a potential greenhouse gas tax. The implementation of
a tertiary abatement upstream of the expander can be
limited by maximum allowable pressure drop (espe-
cially if fixed-bed reactors are used, see Section 5.3.5)
and requirements of the tail-gas expander with respect
to inlet temperature.
Technologies that can be (in principle) applied at
this level of the plant are:
Thermal decomposition.
Non-selective catalytic reduction (NSCR).
Catalytic decomposition.
Selective catalytic reduction (SCR).
5.3.1. Thermal decomposition
Thermal decomposition of N2 O is, at least in the-
ory, not restricted to secondary abatement measures
only. A process based on raising the temperature of
the tail-gas to the required 10231273 K in a recupera-
tive heat exchanger has been proposed to treat off-gas
streams containing at least 1500 ppm N2 O [96]. The
exothermal decomposition of 1500 ppm N2 O in He
causes only a small rise in temperature (6 K), so
a continuous heat supply from an external source to
compensate for heat losses, e.g. by injection of fuel
(natural gas, methane) is required. This process option
seems to be prohibitive, in view of the large investment
which can be expected for a high-temperature recu-
perative heat exchanger and the additional equipment
required to obtain the temperature at plant start-up.
Thermal decomposition could be applicable in specific
situations, e.g. if the tail-gas of the nitric acid plant
could be mixed with other high-temperature gases (or
highly concentrated in N2 O) from other nearby in-
dustrial processes, which would result in net N2 O
destruction.
5.3.2. Non-selective catalytic reduction (NSCR)
NSCR has been developed for NOx removal and has
shown to be capable of reducing N2 O as well. A re-
ducing agent like hydrogen, natural gas, or naphtha is
added to the tail-gas. First all O2 in the tail-gas is con-
sumed, followed by oxygen in components like NOx
and N2 O. Typical conversion efficiencies are about
90% for NOx and 70% for the N2 O, although under
pilot-plant trials, conversion efficiencies of 90% have
been achieved for N2 O [20,38]. Depending on the type
of catalyst (typically Pd/Al2 O3 ) and fuel used, the pro-
cess requires an ignition temperature of 473573 K
(with H2 ) up to 723753 K (with CH4 ) [17]. This
might need preheating of the tail-gas when the tail-gas
temperature is too low.
When the oxygen content in the tail-gas is <3 vol.%,
a single stage catalytic reactor can be used. For
 J. Perez-Ramrez et al. / Applied Catalysis B: Environmental 44 (2003) 117151
handling higher oxygen concentrations, a two stage
catalytic reactor is required to enable intermediate
cooling (Fig. 12a) [30]. Due to the strong exothermal
reaction, the outlet temperature can exceed 1073 K.
Additional cooling is required before the existing
expander, since replacing the expander is not econom-
ical. The surplus heat recovered from (intermediate)
cooling is converted to steam.
NSCR is a proven technology for N2 O reduction,
commercialized by GIAP, and has been applied in sev-
eral nitric acid industries of Russia and USA. How-
ever, the replacement of NSCR systems by SCR for
NOx reduction has a negative side-effect on its appli-
cation for N2 O reduction. NSCR is most likely not a
viable option anymore, due to the high fuel consump-
tion levels and high secondary emissions. In most of
the cases slip of fuel occurs and when using CH4
as a reductant, large amounts of CO are produced
[17,30,33,39]. These negative effects increase the op-
erational costs.
5.3.3. Selective catalytic reduction (SCR)
137
the ignition temperatures vary over a wide temper-
ature range (573773 K) (vide infra). A high igni-
tion temperature of the catalyst requires preheating
of the tail-gas. The SCR process is exothermic but
the amount of heat generated is limited compared to
NSCR. Most revamp cases also require extra-cooling
due to the limited allowable increase of the inlet tem-
perature of existing tail-gas expanders. For new plants
the tail-gas expander can be designed for higher in-
let temperatures, thus avoiding the installation of the
tail-gas cooler.
SCR of N2 O with hydrocarbons has been exten-
sively investigated over Fe-based zeolites (mainly
ZSM-5 and some BEA) [97107]. However, a quan-
titative comparison of activity data from different
authors cannot be made because of the different
(laboratory) conditions of feed composition, space
velocity, and total pressure (Table 9). In addition,
experiments on the pilot-scale are hardly available.
Co-addition of hydrocarbon to the tail-gas mixture
powers the temperature for N2 O reduction over
FeZSM-5 100 K with respect to direct N2 O de-
The schematic flowsheet of this abatement option is
visualized in Fig. 12b. A reducing agent, e.g. propane,
propene, natural gas, LPG (a mixture of propane and
butane), or ammonia, is added to the tail-gas leav-
ing the NO2 absorber. Depending on the operating
pressure, the type of reductant and the catalyst used,
composition (Fig. 13, from () to (); and Table 9),
leading to operation temperatures of 573648 K for
high N2 O-decomposition rates. Furthermore, a higher
stability in the presence of inhibitors like H2 O is also
achieved [98,100]. This behavior strongly contrasts
with other metal-zeolites (e.g. Cu-ZSM-5) or different

Table 9
Representative performances of Fe-zeolite catalysts for catalytic tail-gas N2 O abatement (reduction and decomposition)
Catalyst Reductant GHSV (h1 ), Feed composition T (K), X(N2 O) Reference
P (bar) (%)
FeZSM-5 C 3 H6 60,000, 3 1500 ppm N2 O, 1500 ppm C3 H6 , 2.3 vol.% O2 , 330 ppm NO, 650, 100 [21]
0.5 vol.% H2 O in He
1500 ppm N2 O, 2.3 vol.% O2 , 330 ppm NO, 0.5 vol.% H2 O in He 725, 95
FeZSM-5 C 3 H8 18,000, 1 500 ppm N2 O, 2 vol.% O2 , 3 vol.% H2 O, 1000 ppm C3 H8 in He 625, 80 [100]
FeZSM-5 C 3 H6 50,000, 1 500 ppm N2 O, 1000 ppm C3 H6 , 5 vol.% O2 in He 650, 90 [101]
500 ppm N2 O, 5 vol.% O2 in He 750, 100
FeZSM-5 C 3 H8 20,000, 4 1500 ppm N2 O, 1900 ppm C3 H6 , 2.5 vol.% O2 , 100 ppm NO, 625, 95 [103]
0.5 vol.% H2 O, in N2
1500 ppm N2 O, 2.5 vol.% O2 , 100 ppm NO, 0.5 vol.% H2 O in N2 725, 100
Fe-BEA CH4 60,000, 1 950 ppm N2 O, 500 ppm CH4 in He 575, 100 [106]
950 ppm N2 O in He 725, 100
Fe-BEA NH3 35,000, 1 2000 ppm N2 O, 2000 ppm NH3 , 3 vol.% O2 in He 713, 90 [107]
2000 ppm N2 O, 3 vol.% O2 in He 793, 90
GHSV: gas-hourly space velocity, P: total pressure, T: temperature, X(N2 O) = N2 O conversion.
138 J. Perez-Ramrez et al. / Applied Catalysis B: Environmental 44 (2003) 117151

Fig. 12. Tertiary abatement measures (upstream of the tail-gas expander) for catalytic N2 O removal: (a) dual-staged NSCR, (b) SCR, and
(c) direct N2 O decomposition.
 J. Perez-Ramrez et al. / Applied Catalysis B: Environmental 44 (2003) 117151 139

Fig. 13. N2 O conversion vs. temperature over ex-FeZSM-5 in simulated tail-gases from nitric acid plants. Conditions: () 4.5 mbar N2 O,
() 4.5 mbar N2 O + 70 mbar O2 , () 4.5 mbar N2 O + 1 mbar NO, () 4.5 mbar N2 O + 15 mbar H2 O, () 4.5 mbar N2 O + 70 mbar
O2 + 1 mbar NO + 15 mbar H2 O, () 4.5 mbar N2 O + 70 mbar O2 + 1 mbar NO + 15 mbar H2 O + 4.5 mbar C3 H6 ; balance He; P = 3 bar;
GHSV = 60,000 h1 (after [21]).
catalytic systems, which are inhibited or unaffected
by the hydrocarbon [98,102].
In general, for Fe-zeolite systems, the reduction
efficiency increases with (i) a higher operation tem-
perature, (ii) a higher total pressure, (iii) a higher
reductant concentration, and (iv) a lower gas-hourly
space velocity. Increased total pressure is beneficial for
N2 O destruction, lowers the slip of hydrocarbon and
also the emission of CO [103]. However, at increased
pressure the reducing agents are partly oxidized by
oxygen instead of N2 O, rendering the catalyst less
selective than at atmospheric pressure. This leads to
an increased reagent consumption for a required N2 O
reduction. Propane, propene, and LPG are the most
effective reductants to achieve a low-temperature ac-
tivity over FeZSM-5 [21,99,103,104]. The use of CH4
shifts the N2 O reduction to higher temperature and the
hydrocarbon slip is substantial [32,104]. This can be
remedied enhancing the oxidation performance of the
catalyst by impregnation of noble metals, like Pd or Pt
[99,103]. The major drawback of this process option
are the high costs associated with the permanent con-
sumption of reductant, which contributes importantly
to the total cost of the abatement unit (2060%), and
thus to the cost efficiency of the process [30]. In ad-
dition, it cannot be applied at high N2 O removal to
existing plants with tail-gases at T < 650 K.
N2 O abatement with ammonia as the reducing
agent, has been realized using different Fe-zeolites
(BEA, MFI, FAU, FER, etc.). Mauvezin et al. [107]
reported a temperature of 715 K for 95% N2 O
conversion (in the absence of H2 O and NO) over
Fe-BEA and Fe-FER, the most promising formu-
lations. Although ammonia is readily available in
a nitric acid plant, the high operation temperature
compared with the reduction by light hydrocarbons,
makes it non-competitive for N2 O reduction.
Interestingly, Kieger et al. [108] have reported the
simultaneous reduction of N2 O and NO with ammo-
nia over a Fe-FER catalyst in pilot-plant studies with
extrudates of 3 mm. In the temperature of 665725 K,
NOx and N2 O conversions of >90% and 40% have
been achieved. The efficiency of Fe-BEA for simul-
taneous SCR of NO and N2 O by NH3 has also been
proven by Coq et al. [109]. This option can be interest-
ing in tail-gases containing high concentrations of both
NO and N2 O at a relatively high temperature. How-
ever, the maximum N2 O conversions obtained so far
140 J. Perez-Ramrez et al. / Applied Catalysis B: Environmental 44 (2003) 117151
(40%) are insufficient, and make this technology not
competitive. Still, this process is advantageous com-
pared with simultaneous NON2 O reduction with light
hydrocarbons. In the last case, an efficient coupling
of both NO and N2 O reduction processes in tempera-
ture is not achieved [99]. We have recently found out
that the uncoupling of both reactions is mainly caused
by the inhibiting effect of NO on the N2 O reduction
and hydrocarbon oxidation processes [110]. Contrar-
ily, the N2 O reduction with NH3 over, e.g. Fe-BEA is
promoted in the presence of NO [109].
5.3.4. Direct catalytic decomposition
Direct catalytic decomposition of N2 O without ad-
dition of reducing agents is an attractive and economi-
cal option to reduce N2 O emissions in nitric acid plants
(Fig. 12c), since the high cost of the reductant and
the emissions involved (slip or undesired combustion
products) are avoided. However, none of the catalysts
proposed in the literature show a good activity and
stability in N2 O conversion under realistic conditions
of feed composition and space velocities. Although
there are numerous papers and patents regarding cat-
alytic systems for direct N2 O decomposition [1820]
(and references therein), laboratory results often devi-
ate from what is normally met within industrial prac-
tice, and in most cases severe stability tests under real-
istic conditions have not been taken into consideration.
Transition (Cu, Co, Ni) and noble metal-based
catalysts (Rh, Ru, Pd) on different supports (ZnO,
CeO2 , Al2 O3 , TiO2 , ZrO2 , calcined hydrotalcites, and
perovskites) are very active for N2 O decomposition
in N2 OHe feeds, but the presence of other gases in
the feed (O2 , NOx , H2 O) leads to strong inhibition
and/or deactivation and very low reaction rates un-
der practical conditions [18,19,111118]. From an
extensive screening program (200 catalysts) in a
six-flow reactor [119], it was concluded that no cat-
alyst shows reasonable activities under realistic con-
ditions of temperature, pressure, space velocity, and
feed composition below 700 K [20,21], which limits
the application of this process option to nitric acid
plants with high-temperature tail-gases (700773 K).
Only one-third of the nitric acid plants in Europe
have tail-gases of these characteristics, but since they
are dual-pressure or high-pressure plants with large
HNO3 productions, the contribution to the capacity is
>70%, with equal contribution to N2 O emission.
Of the catalysts evaluated, a novel ex-framework
FeZSM-5 catalyst led to remarkable performance in
simulated tail-gases compared to FeZSM-5 catalysts
prepared by ion exchange and sublimation methods
and noble metal-based catalysts [21,120,121]. The
ex-framework route starts with the preparation of
isomorphously substituted FeZSM-5, followed by
calcination and activation in steam. Details on the cat-
alyst system can be found elsewhere [122124]. The
effect of the feed composition in simulated nitric acid
tail-gases on the N2 O-decomposition activity over
this catalyst is shown in Fig. 13. Addition of oxygen
to the feed hardly affects the activity, while NO en-
hances the reaction rate considerably. Molecular oxy-
gen does not dissociate over FeZSM-5 and does not
compete with N2 O for the active sites. NO promotes
N2 O conversion by accelerating the desorption rate of
adsorbed O species (deposited by N2 O) [125,126].
Water severely inhibits the reaction, probably by hy-
droxylation of the active sites and adsorption in the
zeolite channels. Nevertheless, in the complete gas
mixture (N2 O + O2 + NO + H2 O), steam-activated
FeZSM-5 still shows a significantly higher activity
than in N2 O alone (100% conversion at 750 K). The
promoting effect of NO is thus stronger than the
inhibition by H2 O. This extraordinary behavior dis-
tinguishes FeZSM-5 from other N2 O-decomposition
catalysts. The remarkable behavior of ex-FeZSM-5
in simulated tail-gas mixtures is not limited to its
activity, but also includes stability during 600 h on
stream, at gas-hourly space velocities ranging from
36,000 to 120,000 h1 [21,120]. The performance of
ex-FeZSM-5 appears to enable a cost-effective and
efficient technology in different sources. The poten-
tial of this system has been further confirmed by
successful preliminary tests of coated monoliths in
pilot-scale. In a test for 100 days, and under relevant
conditions of feed composition, a stable N2 O conver-
sion (80%) at 723 K, 3 bar, and 50,000 h1 has been
achieved.
5.3.5. Reactor design
Laboratory screening and testing is carried out with
a catalyst powder or pellets. Small particles normally
do not apply in conventional industrial fixed-bed
reactor because of the high-pressure drop. The imple-
mentation of any catalytic N2 O abatement unit in sta-
tionary sources (chemical production or combustion
 J. Perez-Ramrez et al. / Applied Catalysis B: Environmental 44 (2003) 117151
Table 10
Comparison of reactors for direct N2 O decomposition over ex-FeZSM-5 in tail-gas of nitric acid plants (position 3 in Fig. 3)
Parameter Fixed-bed reactor
Type Extrudates
Geometry dp = 6 mm:
Pressure (bar) 4 10
Reactor volume (m3 ) 5 3
Bed height (m) 0.8 0.5
Catalyst effectiveness, 0.33 0.21
Pressure drop (mbar) 630 160
Basis of design: Nitric acid plant of 1500 t HNO3 per day, tail-gas flow = 200,000 Nm3 h1 , temperature = 750 K, pressure upstream the
expander = 4 or 10 bar, N2 O conversion = 80%, pressure drop = 200 mbar, reactor diameter = 2 m.
processes) requires a close look to the reactor type, in
order to achieve maximum yields under efficient and
safe operation. The practical form and shape of the
catalyst is a crucial aspect to obtain reliable design
data for full-scale implementation.
A numerical comparison between a fixed-bed re-
actor with shaped particles (extrudates) and mono-
lithic reactor with structured catalyst has shown very
advantageous aspects of the latter in the tail-gas up-
stream of the expander. For the calculation, the activity
data of ex-FeZSM-5 in the simulated tail-gas mixture
in Fig. 13 () were used. The basis of design and
the most relevant comparative results are shown in
Table 10, including the dependency of reactor volume,
effectiveness factor, and pressure drop of both geome-
tries.
Due to the high intrinsic reaction rates, the reac-
tion in common particles will exhibit severe diffusion
limitations (effectiveness factor of 2035% with ex-
trudates of 6 mm). Smaller particles to increase the
catalyst utilization would lead to a severe pressure
drop. However, monolithic (and in general structured)
catalysts structures allow decoupling of hydrodynam-
ics, kinetics, and transport phenomena. This allows the
combination of high intrinsic activity, high catalyst ef-
fectiveness, and a low-pressure drop. For the cases of
direct N2 O decomposition of selective catalytic reduc-
tion, we have proposed extruded monoliths with the
ex-FeZSM-5 catalyst. A standard monolith of 200 cpsi
and wall thickness of 0.4 mm leads to higher effective-
ness (90%), small reactor volumes (0.52.5 m3 , de-
pending on operating pressure), and low-pressure drop
(<100 mbar). This clearly indicates that the high cata-
lyst effectiveness in the monolith compensates for the
141
Monolithic reactor
Extruded monolith (square channels)
200 cpsi, d = 0.4 mm:
4 10
2.2 1
0.7 0.3
0.95 0.88
60 10
lower amount of catalyst as compared to a fixed-bed
reactor.
The monolithic catalyst for N2 O abatement in
stationary sources can be implemented as shown in
Fig. 14. The monolith catalyst, typically in square
units of ca. 150 mm 150 mm, can be packed in cat-
alyst baskets, which are inserted in catalyst layers. A
typical practice in the reactor design to allow flexible
operation (prospective or more stringent regulations
or operation changes in the plant) consists of leav-
ing spare space for an extra layer of catalyst. The
possible addition of reductant is also considered in
the figure. The vast commercial experience gained in
de-NOx NH3 -SCR, can be extrapolated to the design
and implementation of this system.
The structured reactor concept is optimal not only
for catalytic N2 O decomposition or reduction in the
tail-gas (upstream or downstream of the expander), but
also for N2 O decomposition in the process-gas, i.e.
below the platinum gauzes. In the latter case, much
higher reaction rates are expected due to the temper-
ature and pressure. Therefore, external mass-transfer
controls the process. From our calculations assuming
gas phase and laminar flow in the channels, it was con-
cluded that a monolithic geometry provides sufficient
high activities [20].
5.4. Quaternary measures
Finally, the N2 O abatement process can be located
downstream of the tail-gas expander before the stack
(position 4 in Fig. 3). These end-of-pipe applications
would not require modifications of the tail-gas ex-
pander in the existing plant. The low temperature
142 J. Perez-Ramrez et al. / Applied Catalysis B: Environmental 44 (2003) 117151
Fig. 14. Reactor and catalyst design for tail-gas N2 O abatement.
(473573 K) and pressure (1 bar) in this location
makes the development of a catalytic abatement very
challenging. In this case, the use of preheating equip-
ment and/or reductants is inevitable, which makes the
cost efficiency of the abatement option less favorable.
In spite of these difficulties, ECN claims feasibility
for this option [30,33].
5.4.1. Non-selective catalytic reduction (NSCR)
For NSCR, the heating capacity of the preheater
shown in Fig. 12a has to be enlarged due to the lower
inlet temperature. As this preheater is a gasgas heat
exchanger operated at atmospheric pressure, the vol-
ume required (and investment cost) will be higher.
Also the amount of heat produced has to be recov-
ered efficiently, possibly in the form of high-pressure
steam.
5.4.2. Selective catalytic reduction (SCR)
Due to the relatively low ignition temperature for
FeZSM-5, during N2 O reduction by hydrocarbons
(623 K) the extra-heat generated can be used to pre-
heat the tail-gas leaving the expander, thus limiting
the amount of utilities used, and production of excess
steam [30,33]. However, an optimized heat recovery
system does increase the investment costs. A flow-
sheet of this process is shown in Fig. 15a. The tail-gas
is preheated using the heat generated in the catalytic
reactor. A start-up in-line burner is required to start the
reaction and to correct for fluctuations in the tail-gas
flow and composition. As discussed in Section 5.3.3,
natural gas is not suitable as reducing agent for use
in an atmospheric pressure environment due to the
large hydrocarbon slip [99,103]. Therefore natural
gas is only used as fuel in a burner for start-up. An
N2 O reduction efficiency of at least 70% is claimed
by using C3 H8 [30,33]. Higher efficiencies can be
obtained by increasing reactor sizing or increased
addition of reducing agent, adding to the operational
costs.
5.4.3. Catalytic decomposition
This process is schematically shown in Fig. 15b.
As the ignition temperature of the catalyst (typically
>673 K) is well above the tail-gas temperature, pre-
heating is required. Heat recovery can be applied in
combination with additional firing, e.g. with natural
gas. A decomposition level of 70% seems feasible
 J. Perez-Ramrez et al. / Applied Catalysis B: Environmental 44 (2003) 117151 143

[30,33]. Using a larger catalytic reactor and/or a higher
operating temperature can increase the conversion lev-
els, at the expenses of increased pressure drop and/or
fuel consumption.
6. Evaluation of potential technologies
6.1. Cost estimation of the different options
Table 11 summarizes the different measures to
reduce N2 O emissions in nitric acid plants, with ad-
vantages, drawbacks, and costs for abatement per
ton of CO2 -eq. To enable cost comparison of the
different N2 O abatement techniques, all known costs
have been extrapolated to a base case. As base case
a nitric acid plant with a production capacity of
1500 t nitric acid (as 100 wt.%) per day, a tail-gas
flow of 200,000 Nm3 h1 and a N2 O concentration of
1500 ppmv in the tail-gas has been used. A distinction
is made between new and existing nitric acid plants.
For evaluation purposes the production cost of nitric
acid has been valued on 70 per ton of 100 wt.%
nitric acid. An operating time of 8000 h per year has

Fig. 15. Quaternary abatement measures (downstream of the tail-gas expander) for catalytic N2 O removal: (a) SCR, and (b) direct
decomposition.

Table 11
Summary of available techniques for the N2 O abatement from nitric acid plants (adapted from [30])
Technology
Primary measures
(1) Improvement of Pt-Rh
gauzes
(2) Oxide-based combustion
catalyst
Secondary measures
(3) Extended reactor chamber
(4) Catalytic decomposition in
the burner
Tertiary measures
(5) NSCR upstream of the
expander
(6) SCR upstream of
the expander
144
Advantages Disadvantages Cost efficiency ( per ton CO2 -eq.)
New plant Existing plant
J. Perez-Ramrez et al. / Applied Catalysis B: Environmental 44 (2003) 117151
Simple to install Uncertain technology (exploratory state) n.a. n.a.
Increase in NO yields
Low catalyst cost NO yield reduction, especially at high n.a. n.a.
combustion pressures
Lower optimal operation temperature
Strong deactivation
Simple process optimization Possible nitric acid production losses due to 0.31.1 0.92.0
NOx decomposition in extended reactor
chamber
No consumption of utilities High investment cost
No catalyst use Severe retrofit limitations for existing plants,
including downtime during construction
Operating experience available in commercial Limited N2 O conversion
scale plant
Retrofit in existing reactor chamber possible Possible HNO3 production losses due to 0.10.7 0.20.8
(replacement catalyst support) catalytic NOx decomposition
Less catalyst required due to increased Retrofit might be limited or impossible due to
reaction kinetics at high temperatures lack of available space for catalyst placement
and design of the burner
First tests on commercial scale have been Unknown thermostability of catalyst and
completed effects of fouling due to platinum deposition
Proven technology Consumption and slip of fuel 1.61.9 1.92.5
Combined NOx /N2 O reduction Production of CO/CO2
No influence on production capacity Heat production
Lifetime of catalyst
Not suitable as end-of-pipe option
Suitable for lower tail-gas temperatures Consumption of reducing agent 1.42.4 1.92.5
(>573 K)

(7) Direct decomposition
upstream of the expander
Quaternary measures
(8) SCR downstream of the
expander
(9) Decomposition
downstream of the expander
(1) The cost efficiency is influenced by the nitric acid production losses. The low range figure excludes production losses, the high figure range assumes 2% production
losses. (2) The cost efficiency is influenced by the LPG price (0.30.4 per kg); n.a.: not available.
Possible combined NOx /N2 O reduction
No influence on production capacity
Catalyst is exposed to milder conditions
compared to NSCR, resulting in expected
higher stability and lower catalyst costs
No consumption of utilities
Simple process c
No influence on production capacity
Catalyst is exposed to milder conditions
compared to in-process decomposition,
resulting in expected higher stability and
lower catalyst costs
Simple end-of-pipe use in existing plants
Possible combined NOx /N2 O reduction
No influence on production capacity
Catalyst is exposed to milder conditions
compared to NSCR, resulting in expected
higher stability and lower catalyst costs
Simple end-of-pipe use in existing plants
No influence on production capacity
Catalyst is exposed to milder conditions
compared to NSCR, resulting in expected
higher stability and lower catalyst costs
Retrofit limitations
Slip when using natural gas as reducing agent
Increased pressure drop and surplus steam
J. Perez-Ramrez et al. / Applied Catalysis B: Environmental 44 (2003) 117151
production due to steam cooler if compared
with direct decomposition
Not yet proven on commercial scale
Relative high tail-gas temperature; tail-gas 0.30.4 0.41.0
temperature <675 K requires preheating
Retrofit limitations
Not proven on commercial scale
Consumption utilities/reducing agent (LPG) n.a. 1.41.5
Pressure drop
Not yet proven on commercial scale
Fuel consumption for heating
Fuel consumption required for preheating n.a. 1.41.6
(>675 K)
Increased reactor size compared to (7)
(increased catalyst cost and pressure drop)
Not yet proven on commercial scale
145
146 J. Perez-Ramrez et al. / Applied Catalysis B: Environmental 44 (2003) 117151

Fig. 16. Cost of the different N2 O abatement options (in per ton CO2 -eq.) in nitric acid production plants.

been used. Investment and operation costs, and cost ef- 6.2.1. Influence of the NO loss
ficiency are indicative only to enable qualitative com- Based on data of Krupp Uhde [17], the amount
parison of the defined N2 O reduction techniques. For of NO loss that could be tolerated in a secondary
smaller plants it is expected that investment costs will abatement system can be determined. The results are
be lower but that the cost efficiency will decrease due shown in Fig. 17. The cumulative cost of a secondary
to scaling effects. An overview of the cost of the most in-process-gas catalytic abatement with different level
promising options is shown in Fig. 16. Alternatives of NO losses (0, 0.25, and 0.5%) has been compared
in an exploratory state (e.g. optimization of the Pt-Rh
gauzes and replacement by oxide systems) are very
unfavorable in economic terms and are not consid-
ered in this estimation. It should be noted that even
for the most expensive method the specific abatement
cost per ton CO2 -eq. is one order of magnitude lower
than those costs reported for the removal of CO2 from
combustion processes [2933].

6.2. In-process-gas versus tail-gas catalytic N2 O

decomposition

Catalytic N2 O decomposition directly after the am-

monia oxidation catalyst in the converter or in the
tail-gas upstream of the expander are the most attrac-
tive options from a technical and economical point of
view in existing nitric acid plants. The NO loss and
Fig. 17. Cumulative total costs vs. time in operation for () N2 O
the catalyst lifetime are decisive factors in determin- tertiary catalytic abatement (catalyst lifetime 3 years) and N2 O
ing the optimal solution for a specific application, as secondary catalytic abatement at three different NO loss rates ()
concluded from the comparison below. 0%, () 0.25%, and () 0.5%.
 J. Perez-Ramrez et al. / Applied Catalysis B: Environmental 44 (2003) 117151
Fig. 18. Cumulative total costs vs. time in operation for N2 O
tertiary catalytic abatement at three different catalyst lifetimes:
() 1 year, () 3 years, and () 5 years, and () NO loss of
0.25% is assumed for secondary abatement catalyst.
to a tertiary tail-gas catalytic abatement (catalytic N2 O
decomposition). The calculations are based on a 1500 t
HNO3 per day plant and the same price per ton was
assumed for both catalysts. A catalyst life of 1 and
3 years was considered for the secondary and tertiary
catalyst, respectively. Losses of NO were translated
into wasted NH3 and rated at January 2000 prices. The
secondary abatement measure provides the more eco-
nomic solution if there are no losses of NO over the
catalyst. If there is only 0.25% NO loss, the costs of
secondary catalyst and tertiary abatements coincide in
the long-term, although in the beginning the secondary
treatment is more economic (due to the initial invest-
ment for the reactor in the tertiary measure). If NO
losses amount to 0.5% the tertiary abatement method
is more economical after 5 years operation.
6.2.2. Influence of the catalyst lifetime
A similar calculation with the same assumptions as
above was carried out to investigate the effect of the
catalyst lifetime. Fig. 18 shows a single dashed curve
for the secondary abatement catalyst, with a lifetime
of 1 year and NO losses of 0.25%. The three solid
lines represent the total cumulated costs for tertiary
N2 O abatement catalysts with lifetimes of 1, 3, and
5 years, respectively. A tertiary catalyst with a 3-year
life is in the long-term comparable in cost to the sec-
ondary catalyst at the assumed rate of NO loss. The
147
short-lived, 1-year tertiary abatement catalyst is not
able to compete, because the volume of catalyst re-
quired per catalyst change is larger for the tertiary cat-
alyst. On the other hand, a 5-year catalyst life for the
tertiary catalyst would give it a clear edge over the
secondary abatement catalyst.
It should, of course, be emphasized that the calcu-
lations behind these curves involve a number of as-
sumptions and are therefore indicative only. However,
it is clear that the NO loss in the secondary, and the
catalyst lifetime in the tertiary catalytic abatement can
dramatically influence the economics of the chosen
process option. A possible future greenhouse tax has
not been considered in these calculations, but if there
turns out to be a difference between competing tech-
nologies in the N2 O reduction level (higher in the ter-
tiary abatement), this aspect needs to be entered into
the assessment of the most economic solution.
6.3. Evaluation
From the technical data and cost figures presented in
Table 11, the following evaluation can be put forward.
6.3.1. New nitric acid plants
Based on cost efficiency, an extended reactor cham-
ber, catalytic decomposition in the reactor chamber,
and catalytic decomposition upstream of the expander
are preferred. From these three options the homoge-
neous decomposition of N2 O in the process-gas has
the lowest risk for a negative impact on process per-
formance and leads to no additional operational costs.
However, because of the heat loss in the extended
chamber, high reduction levels (>80%) are difficult to
achieve. If an N2 O-decomposition catalyst in the re-
actor chamber proves to be stable enough with only
a minor effect on NOx decomposition, this technol-
ogy is preferred on the basis of high cost efficiency.
If maintenance has to be minimized, the extended re-
actor chamber may be preferred. Costs for catalytic
decomposition in the reaction chamber can be in the
order of magnitude of 0.10.7 per ton CO2 -eq., but
can increase dramatically for applications in the reac-
tor chamber when NOx also decomposes. Also the life-
time of the catalyst in the ammonia burner has a large
effect on the operating costs. Therefore, for stable
operation reasons catalytic decomposition upstream
of the expander is preferred (no influence on NOx
148 J. Perez-Ramrez et al. / Applied Catalysis B: Environmental 44 (2003) 117151
decomposition and nitric acid production), with 0.3
0.4 per ton CO2 -eq. A new plant can be designed in
such a way that the temperature before the expander
allows catalytic decomposition of N2 O (673723 K).
6.3.2. Existing nitric acid plants
Cost efficiency decreases for an existing nitric acid
plant. Only catalytic decomposition in the reactor
chamber could result in costs well below 0.45 per
ton CO2 -eq., but this is strongly depending on life-
time and selectivity of the catalyst. Furthermore, not
every nitric acid plant has available space for catalyst
placement. Using an extended reactor chamber causes
a drastic interference with the production process, be-
sides the uncertainties defined above (high operational
risks) and production losses during construction (not
incorporated in the assessment of cost efficiency). An-
other option is catalytic decomposition upstream of
the expander, with costs of 0.41.0 per ton CO2 -eq.
Application is however limited to a minimum tail-gas
inlet temperature of 673723 K. Only one-third of all
nitric acid plants fall into this category, but they are
responsible for a larger part of the capacity (>60%).
Otherwise, SCR and catalytic decomposition up-
stream of the expander should be applied, with related
costs of 1.92.5 per ton CO2 -eq., due to investment
and operational costs, which makes the method less
attractive, even if the operational risks can be consid-
ered as low. A major drawback of the tail-gas options
is the absence of full-scale implementation and scarce
testing in pilot-scale.
6.3.3. Cost efficiency
The cost for N2 O reduction is in most cases for a
considerable part determined by depreciation of the
investment. Due to scale effects, the impact of depreci-
ation will increase for smaller units. Therefore, appli-
cation of N2 O reduction measures in nitric acid plants
with a considerable production capacity (new plants)
will be more cost-effective than application in nitric
acid plants with a relatively low production capacity
(mainly old plants).
6.3.4. Cost impact
Up to now the emission of N2 O is hardly restricted.
Initiatives of nitric acid producers to limit N2 O emis-
sions are driven by individually set targets. Based
on economics of the fertilizer branch, costs for au-
tonomous application of de-N2 O technology have to
be reduced below a maximum of about 0.13 per ton
CO2 -eq. (equivalent to about 0.36 per ton nitric acid
produced). Therefore, most of the techniques men-
tioned above will only be applied when other driving
forces (legislation or joint implementation to reduce
CO2 emission) are introduced.
7. Concluding remarks
Nitric acid production is at present the largest source
of N2 O emission in the chemical industry (400 kt per
year). Emissions of this gas require urgent reduction
due to its role as a strong greenhouse gas (310 times
stronger that CO2 ) and as ozone layer depletor. Al-
though no governmental legislation has been officially
issued with respect to N2 O emission, the voluntary
commitment of adipic acid producers to reduce emis-
sions has put pressure on nitric acid producers. The
large N2 O (and thus CO2 ) reduction potential com-
bined with the finite number of nitric acid plants makes
N2 O abatement in this source a very attractive means
to meet the ongoing process of new agreements and
regulations, in particular the Kyoto Protocol on green-
house gas emissions. However, no mature technology
to reduce N2 O emissions from this source has been
installed as yet.
The European market for de-N2 O in nitric acid
production will concentrate on existing installations
(124 plants), since installation of new plants is not
likely is not expected in the next 10 years. Currently,
two de-N2 O technologies are available at commercial
scale: homogeneous decomposition in the enlarged
chamber and NSCR in the tail-gas. Homogeneous de-
composition in the reactor chamber can be effectively
used for new plants, but it is not competitive for ex-
isting plants due to high retrofit costs, while NSCR
is regarded as being outdated. A number of catalytic
de-N2 O technologies are currently under development.
The choice of a system depends on the plant char-
acteristics, on the location in the process, and on the
technical requirements of the abatement option. The
catalytic decomposition inside the reaction chamber
and the direct decomposition upstream of the tail-gas
expander show the highest cost efficiency (0.21.0
per ton CO2 -eq. in existing plants) and are expected
to play a major role on the future de-N2 O market. The
 J. Perez-Ramrez et al. / Applied Catalysis B: Environmental 44 (2003) 117151
influence of the in-process-gas catalytic decomposi-
tion on the nitric acid production capacity, strongly af-
fects the cost efficiency of this abatement option, and is
not yet well established. Catalytic N2 O-decomposition
upstream of the tail-gas expander may yield a high
N2 O reduction levels at moderate costs, but it cannot
be applied in low-temperature tail-gas facilities (T <
673 K). Selective catalytic N2 O reduction upstream
of the tail-gas expander can be applied for low tem-
perature tail-gases, but operation costs are prohibitive
due to the permanent consumption of the reductant. If
SCR of N2 O can be efficiently combined with SCR of
NOx , the SCR technology will gain attractiveness. The
cost efficiency of the abatement options downstream
of the tail-gas expander (end-of-pipe options) is even
less favorable, due to the inevitable use of preheating
equipment and/or reductants.
References
[1] M. Kavanaugh, Atmos. Environ. 21 (1987) 463.
[2] T. Machida, T. Nakazawa, M. Tanaka, Y. Fujii, S. Aoki, O.
Watanabe, in: Proceedings of the International Symposium
on Global Cycles of Atmospheric Greenhouse Gases, Sendai,
Japan, 1994, pp. 113116.
[3] J.T. Houghton, et al. (Eds.), Climate change 2001, the
scientific basis. Contribution of the Working Group I to the
Third Assessment Report of the IPCC, Cambridge, 2001.
[4] P.J. Crutzen, J. Geophys. Res. 76 (1971) 7311.
[5] M.B. McElroy, J.C. McConnell, J. Atmos. Sci. 28 (1971)
1095.
[6] P.J. Crutzen, C.J. Howard, Pure Appl. Geophys. 116 (1978)
497.
[7] http://www.montrealprotocol.org.
[8] J.T. Houghton, B.A. Callander, S.K. Verney (Eds.), Climate
change 1992, The Supplementary Report to the IPCC
Scientific Assessment, IPCC, Cambridge, 1992.
[9] L. Donner, V. Ramanathan, J. Atmos. Sci. 37 (1980) 119.
[10] V. Ramanathan, R.J. Cicerone, H.H. Singh, J.T. Kiel, J.
Geophys. Res. 90 (1985) 5547.
[11] G. Centi, S. Perathoner, F. Vazzana, CHEMTECH 29 (1999)
48.
[12] http://www.epa.gov/globalwarming/emissions/national/gwp.
html.
[13] http://www.unfccc.int/resource/convkp/kpeng.pdf.
[14] http://www.unfccc.int/resource/convkp/kpco2.pdf.
[15] http://www.unfccc.int/resource/convkp/ratlist.pdf.
[16] C. Kroeze, Nitrous Oxide Emission Inventory and Options
for Control in The Netherlands, Report no. 773001004,
National Institute for Public Health and the Environment,
Bilthoven, The Netherlands, 1994.
[17] M. Schwefer, R. Maurer, M. Groves, in: Proceedings of the
International Conference and Exhibition on Nitrogen 2000,
Vienna, 1214 March 2000, pp. 6081.
149
[18] F. Kapteijn, J. Rodrguez-Mirasol, J.A. Moulijn, Appl. Catal.
B 9 (1996) 25.
[19] G. Centi, S. Perathoner, F. Vanazza, M. Marella, M.
Tomaselli, M. Mantegazza, Adv. Environ. Res. 4 (2000) 325.
[20] J. Prez-Ramrez, Catalyzed N2 O Activation. Promising
(New) Catalysts for Abatement and Utilization, Ph.D. Thesis,
Delft University of Technology, Delft, 2002.
[21] J. Prez-Ramrez, F. Kapteijn, G. Mul, X. Xu, J.A. Moulijn,
Catal. Today 76 (2002) 55.
[22] R.A. Reimer, C.S. Slaten, M. Seapan, T.A. Koch, V.G. Triner,
in: J. van Ham, A.P.M. Baede, L.A. Meyer, R. Ybema (Eds.),
Non-CO2 Greenhouse Gases: Scientific Understanding,
Control and Implementation, Kluwer Academic Publishers,
Dordrecht, 2000, pp. 347358.
[23] R.A. Reimer, C.S. Slaten, M. Seapan, M.W. Lower, P.E.
Tomlinson, Environ. Prog. 13 (1994) 134.
[24] G. Kuhn, V. Schumacher, E. Wagner, Int. Fertil Soc. Proc.
435 (1999) 1.
[25] G. Kuhn, in: Proceedings of the International Conference
on Industrial Atmospheric Pollution, NOx N2 O Emission
Control: Panel of Available Techniques, Session 4, ADEME
Editions, Paris, 2001.
[26] G.I. Panov, CATTECH 4 (2000) 18.
[27] G. Centi G, P. Generali, L. dallOlio, S. Perathoner, Z. Rak,
Ind. Eng. Chem. Res. 39 (2000) 131.
[28] http://www.enviro-chem.com/plant-tech/3rdtier/noxtop.html.
[29] Dutch Notes on BAT for the Production of Nitric Acid,
Ministry of Housing, Spatial Planning, and the Environment,
The Hague, The Netherlands, 1999.
[30] A.W. Smit, M.M.C. Gent, R.W. van den Brink, Market
Potential for Reduction of N2 O Emission at Nitric
Acid Plants, Document 63578-700302, Jacobs Engineering
Nederland, 2001.
[31] C. Ewald, Workshop on Best Practices in Policies and
Measures, Copenhagen, 1113 April 2000.
[32] W. van der Lans, Prospects for Reducing N2 O Emissions
from the Production of Nitric Acid in The Netherlands,
Utrecht, The Netherlands, 2000.
[33] R.W. van den Brink, M.J.F.M. Verhaak, J.W.N. van Lijssel,
M.M.M. Gent, Catalytic Destruction of N2 O in Nitric
Acid Industry (Executive Summary), Report ECN-C-00-003,
Energieonderzoek Centrum Nederland, 2000.
[34] S.I. Clarke, W.J. Mazzafro, Kirk-Othmer Encyclopedia of
Chemical Technology, M. Howe-Grant (Ed.), fourth ed., vol.
17, Wiley, New York, 1997, pp. 8494.
[35] E. Wagner, T. Fetzer, in: G. Ertl, H. Knzinger, J. Weitkamp
(Eds.), Handbook of Heterogeneous Catalysis, vol. 4, VCH,
Weinheim, 1997, pp. 17481761.
[36] R.J. Farrauto, C.H. Bartholomew, Fundamentals of Industrial
Catalytic Processes, Chapman & Hall, London, 1997, p. 481.
[37] H.F. Rase, Handbook of Commercial Catalysts. Heterogen-
eous Catalysts, CRC Press, Boca Raton, 2000, p. 234.
[38] S.P.S. Andrew, in: C. Keleti (Ed.), Fertilizer Science and
Technology Series: Nitric Acid and Fertilizer Nitrates, vol.
4, Marcel Dekker, New York, 1985, pp. 3140.
[39] Best Available Techniques for Pollution Prevention and
Control in the European Fertilizers Industry, Booklet no.
2 of 8: Production of Nitric Acid, 2000, available at
http://www.efma.org/publications.
150 J. Perez-Ramrez et al. / Applied Catalysis B: Environmental 44 (2003) 117151
[40] Intergovernmental Panel on Climate Change (IPCC), Good
Practice, Guidance, and uncertainty Management in National
Greenhouse Gas Inventories: N2 O Emissions from Adipic
Acid and Nitric Acid Production, 2000 (Chapter 3.2).
[41] V.I. Atroshchenko, Proc. Kharkov Politechnical Inst. 6
(1959) 63.
[42] V.A. Sadykov, L.A. Isupova, I.A. Zolotarskii, L.N.
Bobrova, A.S. Noskov, V.N. Parmon, E.A. Brushtein, T.V.
Telyatnikova, V.I. Chernyshev, V.V. Kunin, Appl. Catal. A
204 (2000) 59.
[43] N.I. Ilchenko, G.I. Godolets, I.M. Avilova, Teor. Experim.
Khim. 11 (1975) 56.
[44] T. Pignet, L.D. Schmidt, J. Catal. 40 (1975) 212.
[45] L. Andrussow, Z. Ang. Ch. 39 (1926) 321.
[46] M. Bodenstein, Z. Elektroch. 41 (1935) 466.
[47] M. Bodenstein, Z. Elektroch. 47 (1935) 501.
[48] F. Raschig, Z. Ang. Ch. 40 (1927) 1183.
[49] J. Zawadzki, Discuss. Faraday Soc. 8 (1950) 140.
[50] L. Andrussow, Z. Elektrock. 36 (1930) 756.
[51] M. Bodenstein, Z. Ang. Ch. 40 (1927) 174.
[52] B. Neumann, H. Rose, Z. Ang. Ch. 33 (1920) 41.
[53] L. Andrussow, Z. Ang. Ch. 40 (1927) 166.
[54] Ya.M. Fogel, B.T. Nadykto, V.F. Rybalko, V.I. Shvachko,
I.E. Korobchanskaya, Kinet. Catal. 5 (1964) 431.
[55] C.W. Nutt, S. Kapur, Nature 220 (1968) 697.
[56] C.W. Nutt, S. Kapur, Nature 224 (1968) 169.
[57] M. Asscher, W.L. Guthrie, T.H. Lin, G.A. Somorjai, J. Phys.
Chem. 88 (1984) 3233.
[58] W.D. Mieher, W. Ho, Surf. Sci. 322 (1995) 151.
[59] S. Schneider, D. Bazin, F. Garin, G. Maire, M. Capelle, G.
Meunier, R. Noirot, Appl. Catal. A 189 (1999) 139.
[60] J.M. Bradley, A. Hopkinson, D.A. King, J. Phys. Chem. 99
(1995) 17032.
[61] T. Curtin, F.O. Regan, C. Deconinck, N. Knuttle, B.K.
Hodnett, Catal. Today 55 (2000) 189.
[62] S.H. Oh, H. Fisher, J.E. Carpenter, D.W.J. Goodman, J.
Catal. 100 (1985) 360.
[63] A.C.M. van den Broek, A.C.M. van Grondelle, R.A. van
Santen, J. Catal. 185 (1999) 297.
[64] E.V. Kondratenko, M. Baerns, J. Prez-Ramrez, in
preparation.
[65] A.P. Oele, Chem. Eng. Sci. 8 (1958) 146.
[66] J. Neumann, in: Proceedings of the Krupp Uhde Techno-
logies Users Group Meeting 2000 for Nitric Acid and
Ammonium Nitrate, Vienna, March 2000.
[67] http://www.noble.matthey.com/product.
[68] http://www.engelhard-clal.fr/dpi/gb/nitroprocess1.html.
[69] http://www.dmc-2.com/omg/en/index.html.
[70] http://www.ravindraheraeus.com/products.
[71] K. Schmidt-Szalowski, K. Krawczyk, J. Petryk, Appl. Catal.
A 175 (1998) 147.
[72] J. Petryk, E. Kolakowska, Appl. Catal. B 24 (2000) 121.
[73] F. Janssen, R. Meier, Catal. Today 16 (1993) 157.
[74] DE 334975 (1915), to BASF.
[75] DE 337568 (1916), to BASF.
[76] M.M. Karavayev, D.V. Zharov, N.V. Dobrovolskaya, T.I.
Nazarova, O.I. Kulish, NIITEchim, Moscow, 1975, p. 40.
[77] M.M. Karavayev, A.P. Zasorin, N.F. Kleshchev, Catalytic
oxidation of ammonia, Khimia, Moscow, 1983.
[78] K. Schmidt-Szalowski, K. Krawczyk, J. Petryk, Pol. J. Appl.
Chem. XLIII (1999) 29.
[79] K. Krawczyk, J. Petryk, K. Schmidt-Szalowski, in: G.
Poncelet, J. Martens, B. Delmon, P.A. Jacobs, P. Grange
(Eds.), Preparation of Catalysts VI, Study of Surface Science
and Catalysis, vol. 91, Elsevier, Amsterdam, 1994, p. 683.
[80] J. Petryk, E. Kolakowska, K. Krawczyk, Z. Kowalczyk,
in: B. Delmon, P.A. Jacobs, R. Maggi, J. Martens, P.
Grange, G. Poncelet (Eds.), Preparation of Catalysts VII,
Study of Surface Science and Catalysis, vol. 118, Elsevier,
Amsterdam, 1998, p. 341.
[81] http://www.joyindustries.com/cobalt oxide.htm.
[82] D.A. Epshtein, N.M. Tkachenko, M.A. Miniovich, Dokl.
AN SSSR 122 (1958) 874.
[83] I.A. Zolotarsky, A.S. Noskov, V.A. Kuzmin, L.N. Bobrova,
E.A. Brushtein, V.A. Sadykov, L.A. Isupova, V.A.
Chernyshev, A.I. Potekha, A.A. Khazanov, US 6451278 B1,
2002, to Boreskov Institute of Catalysis.
[84] E. Fareid, G. Kongshaug, L. Hjornevik, . Nirisen, EP
0359286, 1993, to Norsk Hydro.
[85] T. Hallan, O. Nirisen, K. Schffel, D. Waller, in: Proceedings
of the International Conference on Industrial Atmospheric
Pollution, NOx N2 O Emission Control: Panel of Available
Techniques, Session 4, ADEME Editions, Paris, 2001.
[86] V. Schumacher, G. Brger, T. Fetzer, M. Baier, M. Hesse,
WO 9955621, 1999, to BASF.
[87] G. Kuhn, in: Proceedings of the Krupp Uhde Technologies
Users Group Meeting 2000, Vienna, 1416 March 2000.
[88] . Nirisen, K. Schffel, D. Waller, D. vreb, WO
0202230 A1, 2002, to Norsk Hydro.
[89] W. Buckhardt, G. Winterstein, F. Seifert, M. Voigt, WO
0013789, 2000, to HIFTK/PKV.
[90] M. Voigt, W. Buckhardt, EP 1147813, 2001, to HIFTK/PKV.
[91] T. Koch, US 5478549, 1995, to DuPont.
[92] A. Vernooy, Dupont, WO 0051715, 2000, to DuPont.
[93] R. Kubisa, M. Klein, WO 9907638, 1999, to L&C
Steinmller.
[94] B. Neveu, C. Hamon, WO 9964139, 1999 to Grande-
Paroisse.
[95] M. Schwefer, R. Maurer, T. Turek, WO 0151415, 2001, to
Krupp Uhde.
[96] M. Galle, D.W. Agar, W. Watzenberger, Chem. Eng. Sci. 56
(2001) 1587.
[97] C. Pophal, T. Yogo, K. Yamada, K. Segawa, Catal. Lett. 44
(1997) 271.
[98] C. Pophal, T. Yogo, K. Yamada, K. Segawa, Appl. Catal. B
16 (1998) 177.
[99] M. Kgel, R. Mnnig, W. Schwieger, A. Tissler, T. Turek,
J. Catal. 182 (1999) 470.
[100] G. Centi, F. Vanazza, Catal. Today 53 (1999) 683.
[101] K. Yamada, S. Kondo, K. Segawa, Micropor. Mesopor.
Mater. 3536 (2000) 227.
[102] J. Prez-Ramrez, J.M. Garca-Corts, F. Kapteijn, M.J.
Illn-Gmez, C. Salinas-Martnez de Lecea, J.A. Moulijn,
Appl. Catal. B 25 (2000) 191.
 J. Perez-Ramrez et al. / Applied Catalysis B: Environmental 44 (2003) 117151
[103] R.W. van den Brink, S. Booneveld, J.R. Pels, D.F. Bakker,
M.J.F.M. Verhaak, Appl. Catal. B 32 (2001) 73.
[104] Z.S. Rak, M.J.F.M. Verhaak, A. Bos, G. Centi, WO
99/49954, 1999, to ECN.
[105] M. Schwefer, R. Mauer, T. Turek, M. Kgel, WO 0151182,
2001, to Krupp Uhde.
[106] S. Kameoka, K. Kita, T. Takeda, S. Tanaka, S. Ito, K. Yuzaki,
T. Miyadera, K. Kunimori, Catal. Lett. 69 (2000) 169.
[107] M. Mauvezin, G. Delahay, F. Kiich, B. Coq, S. Kieger,
Catal. Lett. 62 (1999) 41.
[108] S. Kieger, L. Navascues, P. Gry, in: Proceedings
of the International Conference on Industrial Atmos-
pheric Pollution, NOx N2 O Emission Control: Panel of
Available Techniques, Session 4, ADEME Editions, Paris,
2001.
[109] B. Coq, M. Mauvezin, G. Delahay, J.-B. Butet, S. Kieger,
Appl. Catal. B 27 (2000) 193.
[110] J. Prez-Ramrez, F. Kapteijn, 2003, submitted for
publication.
[111] J.N. Armor, T.A. Braymer, T.S. Farris, Y. Li, F.P. Petrocelli,
E.L. Weist, S. Kannan, C.S. Swamy, Appl. Catal. B 7 (1996)
397.
[112] T.W. Dann, K.H. Schulz, M. Mann, M. Collings, Appl.
Catal. B 6 (1995) 1.
[113] X.F. Wang, H.C. Zheng, Appl. Catal. B 3 (1993) 55.
151
[114] J. Oi, A. Obuchi, G.R. Bamwenda, A. Ogata, H. Yagita, S.
Kushiyama, K. Mizuno, Appl. Catal. B 12 (1997) 277.
[115] G. Centi, A. Galli, B. Montanari, S. Perathoner, A. Vaccari,
Catal. Today 35 (1997) 113.
[116] J. Oi, A. Obuchi, A. Ogata, G.R. Bamwenda, K. Tanaka, T.
Hibino, S. Kushiyama, Appl. Catal. B 13 (1997) 197.
[117] J. Prez-Ramrez, J. Overeijnder, F. Kapteijn, J.A. Moulijn,
Appl. Catal. B 23 (1999) 51.
[118] H.C. Zheng, X.Y. Pang, Appl. Catal. B 13 (1997) 113.
[119] J. Prez-Ramrez, R.J. Berger, G. Mul, F. Kapteijn, J.A.
Moulijn, Catal. Today 60 (2000) 93.
[120] J. Prez-Ramrez, F. Kapteijn, G. Mul, J.A. Moulijn, Chem.
Commun. (2001) 693.
[121] J. Prez-Ramrez, F. Kapteijn, G. Mul, J.A. Moulijn, Appl.
Catal. B 35 (2002) 227.
[122] A. Ribera, I.W.C.E. Arends, S. de Vries, J. Prez-Ramrez,
R.A. Sheldon, J. Catal. 195 (2000) 287.
[123] J. Prez-Ramrez, F. Kapteijn, G. Mul, J.A. Moulijn, J. Catal.
207 (2002) 113.
[124] J. Prez-Ramrez, F. Kapteijn, J.C. Groen, A. Domnech,
G. Mul, J.A. Moulijn, J. Catal. (2003), in press.
[125] J. Prez-Ramrez, F. Kapteijn, G. Mul, J.A. Moulijn, J. Catal.
208 (2002) 211.
[126] G. Mul, J. Prez-Ramrez, F. Kapteijn, J.A. Moulijn, Catal.
Lett. 77 (2001) 7.