Selective catalytic oxidation of ammonia to nitrogen at low temperature on Pt/CuO/Al2O3

Article abstract of DOI:10.1016/j.jcat.2004.11.023

Pt/CuO/Al₂O₃ catalysts with 20 wt% CuO and 0.5-4 wt% Pt were prepared and used for the selective catalytic oxidation of ammonia in wet and dry gases and in synthetic biogas with CO, hydrogen, and methane. Imaging of the catalysts with HREM and STEM, combined with elemental analysis, reveals that copper is well dispersed and occurs in the form of patches on the support. Platinum in the form of nanometer-sized particles are localised on copper-rich areas. The activity measurements show that Pt/CuO/Al ₂O₃ is active and selective for the oxidation of ammonia to form nitrogen. Compared with dry feeds, the catalyst shows higher selectivity for nitrogen when the feed is wet. Pretreatment of the catalyst with SO ₂-containing gas produces an improved catalyst with a selectivity for nitrogen of about 97-98% at complete ammonia conversion. Pt/CuO/Al ₂O₃ shows good performance also for the oxidation of ammonia to nitrogen in the presence of biogas components including water. However, the performance is not as good under dry conditions, giving either low selectivity for nitrogen or low conversion of the ammonia, depending on the partial pressure of oxygen.

Full text of DOI:10.1016/j.jcat.2004.11.023

Journal of Catalysis 230 (2005) 1–13
www.elsevier.com/locate/jcat
Selective catalytic oxidation of ammonia to nitrogen at low temperature
on Pt/CuO/Al2O3
a,∗
Göran Olofsson a, L. Reine Wallenberg b, Arne Andersson
a
Department of Chemical Engineering, Chemical Center, Lund University, Box 124, SE-221 00 Lund, Sweden
b
Department of Materials Chemistry/nCHREM, Chemical Center, Lund University, Box 124, SE-221 00 Lund, Sweden
Received 7 September 2004; revised 12 November 2004; accepted 16 November 2004
Available online 12 January 2005
Abstract
Pt/CuO/Al O catalysts with 20 wt% CuO and 0.5–4 wt% Pt were prepared and used for the selective catalytic oxidation of ammonia
2
3
in wet and dry gases and in synthetic biogas with CO, hydrogen, and methane. Imaging of the catalysts with HREM and STEM, combined
with elemental analysis, reveals that copper is well dispersed and occurs in the form of patches on the support. Platinum in the form of
nanometer-sized particles are localised on copper-rich areas. The activity measurements show that Pt/CuO/Al O is active and selective for
2
3
the oxidation of ammonia to form nitrogen. Compared with dry feeds, the catalyst shows higher selectivity for nitrogen when the feed is
wet. Pretreatment of the catalyst with SO -containing gas produces an improved catalyst with a selectivity for nitrogen of about 97–98% at
2
complete ammonia conversion. Pt/CuO/Al O shows good performance also for the oxidation of ammonia to nitrogen in the presence of
2
3
biogas components including water. However, the performance is not as good under dry conditions, giving either low selectivity for nitrogen
or low conversion of the ammonia, depending on the partial pressure of oxygen.
2004 Elsevier Inc. All rights reserved.
Keywords: Selective catalytic oxidation; SCO; Ammonia; Pt/CuO/Al O catalysts; Dry feed; Wet feed; Sulphating; Biogas components; HREM; STEM
2
3
1. Introduction
in the future concerns the use of biogas for synthesis pur-
poses. In the gasification of biomass, the fuel-bound nitrogen
is released predominantly as ammonia, which in combined
heat and power production should be removed prior to the
gas turbine [1]. Thus, a catalyst operating at high tempera-
tures (∼ 700 ◦C) would give the best process efficiency. On
the other hand, in other applications it is usually desirable
to have a catalyst operating at lower temperatures, for ex-
ample, in SCR facilities where ammonia is fed to react with
NOx on a catalyst to give nitrogen. Here the degree of NOx
conversion is limited because any trace of ammonia has to
be avoided. Therefore, a NH3/NOx molar ratio of about 0.9
is usually not exceeded. The NOx removal efficiency would
be improved if one could use a secondary SCO catalyst bed
to efficiently convert an added small excess of ammonia to
nitrogen under SCR conditions. However, the difficulty is
that the SCO catalyst has to be effective below 300 ◦C, as
in many other applications in the presence of high concen-
The abatement of ammonia from industrial waste gas
streams is becoming increasingly important because of en-
vironmental concerns. Depending on the application, var-
ious techniques can be used for the removal of ammonia,
such as adsorption, absorption, chemical treatment, catalytic
decomposition, and selective catalytic oxidation (SCO) to
form nitrogen and water. The last technique is attractive,
especially for treatment of large gas flows containing oxy-
gen and a low concentration of ammonia. The application
area for SCO includes the treatment of waste gases from
chemical production processes, SCR facilities, and gasifi-
cation/combustion of biomass for combined heat and power
generation. Another application area that might be important
*
Corresponding author.
E-mail address: Arne.Andersson@chemeng.lth.se (A. Andersson).
0021-9517/$ – see front matter
doi:10.1016/j.jcat.2004.11.023
2004 Elsevier Inc. All rights reserved.

2
G. Olofsson et al. / Journal of Catalysis 230 (2005) 1–13
trations of water [2]. Furthermore, it is necessary to achieve
almost 100% conversion of the ammonia to avoid odour [3].
Platinum is active for ammonia oxidation and is used as
a catalyst in the first step in nitric acid production to oxidise
ammonia to NO at 800–900 ◦C [4]. At temperatures be-
low 300 ◦C platinum produces nitrogen rather than NO [5],
but the selectivity is comparatively low since high levels
of N2O are co-produced [2]. Examples of more promising
catalyst systems for the SCO of ammonia are (i) the sup-
ported oxide systems MoO3/SiO2 [6,7], V2O5/TiO2 [2,8],
CuO/TiO2 [8], CuO/Al2O3 [9,10], NiO/Al2O3 [11], and
Fe2O3/TiO2 [12]; (ii) the ion-exchanged zeolite systems
Cu-ZSM-5 [8,13], Cu-Y-zeolite [9], Pd-ZSM-5 [2], and Fe-
ZSM-5 [13,14]; and the supported oxide systems with a
precious metal component, Pd–V2O5–WO3/TiO2–SiO2 [3],
Ir–V2O5–WO3/TiO2–SiO2 [3], Ag/CuO/Al2O3 [15,16], and
Au/CuO/Al2O3 [17]. Considering the desired target of both
complete conversion of ammonia and high selectivity for ni-
trogen, a problem with most of these catalyst systems is that
they have to be used at temperatures well above 300 ◦C, and
the selectivity for nitrogen formation decreases with increas-
ing temperature and conversion. For use in SCR applications
the target temperature is below 300 ◦C, and thus there is a
need for more efficient low-temperature catalysts. The ad-
dition of a precious metal component to supported oxides
seems beneficial [3,15–17]; for example, Ag/CuO/Al2O3
has been reported to give complete conversion of ammo-
nia at 250 ◦C with a high selectivity for nitrogen [15].
A problem here is that the space velocity has to be low,
around 450 ml/(min g) of catalyst, compared with 1000–
5000 ml/(min g) for the other catalyst systems. Therefore, it
is of interest that Pt/CuO/Al2O3 has been reported to be ac-
tive and selective for nitrogen formation at 200 ◦C at a space
velocity of about 5000 ml/(min g) when it is used for ammo-
nia oxidation in dry biogas with low oxygen content [18].
For combined heat and power generation a temperature of
200 ◦C is not optimal since higher temperatures are pre-
ferred [1], but doubtlessly the Pt/CuO/Al2O3 system has a
potential for low-temperature applications.
In the present study, to further explore the Pt/CuO/Al2O3
catalyst system, samples with various platinum contents
have been prepared and used for the selective catalytic ox-
idation of ammonia to nitrogen under both dry and wet
conditions and with different oxygen contents in the gas.
Moreover, as waste gases may contain SO2, the influence
of SO2 treatment on the performance of the catalyst has
been investigated. It has been reported for ammonia oxi-
dation on CuO/TiO2 [8], CuO/Al2O3 [10], Fe-ZSM-5 [13],
and Fe2O3/TiO2 [12] that sulphate species may improve the
selectivity for nitrogen. In a patent [19] data are presented
showing that sulphating of, for example, a Pt/CuO/SiO2 cat-
alyst has a remarkable effect on the selectivity for nitrogen.
Another aim of the present investigation was to collect data
for ammonia oxidation in the presence of biogas compo-
nents (CO, H2, CH4) by FTIR analysis of the product stream.
In the previous investigation [18] of the same catalyst sys-
tem, mass spectrometric analysis was performed without
separation of the components with overlapping mass signals
(CO2/N2O and CO/N2).
2. Experimental
2.1. Catalyst preparation
Four different Pt/CuO/Al2O3 catalysts were prepared
with 20 wt% CuO and with 0.5, 1, 2, and 4 wt% Pt, respec-
tively. CuO/Al2O3 was synthesised first, and, in a second
step, the desired amount of Pt was deposited. CuO/Al2O3
was prepared by incipient wetness impregnation of Al2O3
(Condea, 119 m2/g) with a water solution of Cu(NO3)2 ·
3H2O (Merck, p.a.). The sample was dried for 5 h under
ambient conditions, followed by another 24 h at 120 ◦C.
Subsequently, the sample was calcined in air with ramp-
ing temperature. Primarily, the temperature was increased
by 10 ◦C/min up to 300 ◦C, where the sample was held for
1 h. In this step most of the nitrate decomposed to form ni-
trogen oxides. The temperature was thereafter increased by
10 ◦C/min up to 500 ◦C, where it was kept constant for 6 h.
For preparation of Pt/CuO/Al2O3, 3 g of the as-prepared
CuO/Al2O3 sample was immersed in the desired amount
of Pt(II) nitrate solution (Heraeus, 15.46 wt% Pt) diluted
with water to give a total volume of 40 ml. The slurry was
stirred repeatedly for 1 h to ensure that practically all of the
Pt was adsorbed on CuO/Al2O3. The Pt/CuO/Al2O3 precur-
sors were dried and calcined with the same procedure as that
used for CuO/Al2O3, with the exception that the final cal-
cination temperature was 450 ◦C instead of 500 ◦C. After
calcination, the samples were crushed and sieved. The frac-
tion of particles with diameters in the 250–425-µm range
was collected and used in the experiments.
2.2. Activity measurements
The activity measurements were carried out in a stainless-
steel plug-flow reactor with an inner diameter of 1.9 mm.
The reactor was placed in a GC-type electric furnace
equipped with a temperature control. An Environics series
2000 computerised multicomponent gas mixer was used to
prepare the inlet gas mixtures, consisting of NH3 and O2
in N2 with or without water vapour and biogas components
(CO, H2, and CH4). For analysis of the inlet and product
gases a GASMET FT-IR gas analyser was used, which per-
mitted an analysis of NH3, N2O, NO, NO2, CO, CO2, CH4,
and H2O. The formation of N2 and the consumption of H2
were obtained from the mass balance for N and H, respec-
tively, over the reactor. All feed gases (1.48 0.04 vol%
NH3 in He, O2, N2, CO, H2, and CH4) were high purity
grade supplied by AGA. The conversion of NH3 and the se-
lectivity for N2, N2O, NO, and NO2 were measured on the
catalysts for different gas compositions and temperatures. In
the experiments the total inlet gas flow was 1000 Nml/min,

G. Olofsson et al. / Journal of Catalysis 230 (2005) 1–13
3
the temperature was varied between 150 and 335 ◦C, and the
amount of catalyst was 0.20–0.35 g. Before the measure-
ments the catalyst was pretreated at 235 ◦C for 16 h under
a gas flow of 1000 Nml/min containing 700 ppm NH3 and
8.0 vol% O2 in N2. After this treatment, in some experi-
ments, the catalyst was subjected for another 30 min to the
former gas with an additional 100 ppm SO2.
and 44 amu), NO (30 amu), and NO2 (30 and 46 amu) were
registered and used for calculation of the desorption profiles.
X-ray diffraction (XRD) analysis was carried out on a
Seifert XRD 3000 TT diffractometer with monochromatic
Cu-Kα radiation (50 kV/30 mA). The scanning range was
2.5–40◦ (Θ) with a step size of 0.01◦ and a step time of 1.0 s.
Compound identification was accomplished by comparison
with the JCPDS data file [21]. The measurements were per-
formed on ground samples with the use of a rotating sample
holder.
2.3. Catalyst characterisation
The specific surface areas of the samples were ob-
tained from a multipoint Brunauer–Emmet–Teller (BET)
analysis of the nitrogen adsorption isotherm, recorded on
a Micromeritics ASAP 2400 instrument at liquid nitrogen
temperature. All samples were degassed under vacuum at
350 ◦C for 16 h before analysis. Specific surface areas and
notation of the samples are given in Table 1.
Temperature-programmed desorption (TPD) experiments
were performed with a TAP-2 reactor system with a microre-
actor connected to a UTI 100C mass spectrometer [20]. Fifty
milligrams of catalyst was loaded between inert layers of
quartz sand of the same particle size. Before the experiment
was started, the catalyst was heated for 10 min at 350 ◦C with
pure O2 flowing through the reactor (40 Nml/min), cooled
down to 50 ◦C, and evacuated. Then, a flow with 20 vol%
NH3 in Ar (58 Nml/min) was passed through the sample
for 30 min. After evacuation of the reactor, the temperature
was increased linearly by 10 ◦C/min from 50 up to 350 ◦C.
Simultaneously, the mass signals from NH3 (17 amu), H2O
(17 and 18 amu), H2 (2 amu), N2 (28 amu), N2O (28, 30
Temperature-programmed reduction (TPR) was per-
formed on a Micromeritics TPD/TPR 2900 instrument. The
temperature was increased by 10 ◦C/min from 40 to 900 ◦C.
The reducing gas was 8.5 vol% H2 in Ar, and a flow rate of
40 ml/min was used. The water produced during the reduc-
tion process was condensed in a cold trap at about −90 ◦C.
The 4-mm quartz tube reactor was loaded with ∼ 20 mg of
sample with particle sizes in the range of 250–425 µm.
High-resolution transmission electron microscopy
(HREM) and elemental analysis were performed with a JEM
3000F field-emission microscope equipped with an Oxford
energy-dispersive X-ray spectrometer (XEDS). The micro-
scope was operated at 300 kV in both conventional and
scanning mode. A Fischione high-angle annular dark-field
detector (HAADF) was used for selective imaging of plat-
inum particles on the support. Mo grids with a continuous
amorphous carbon film were used, rather than the usual Cu
grids, to eliminate spurious Cu radiation from the back-
ground. The resolution of the microscope is 0.162 nm in
conventional HREM imaging and 0.14 nm in STEM mode.
A probe size of 0.2–0.7 nm was used for imaging, and 1.0 nm
for XEDS mapping.
Table 1
Notation, composition and specific surface area of prepared Pt/CuO/Al O
2
3
catalysts and the Al O support
2
3
Sample
Pt
(wt%)
CuO
(wt%)
Specific surface
area (m /g)
2
3. Results
Al O
–
0.5
1
2
4
–
119
99
94
100
99
2
3
3.1. TPD of adsorbed ammonia
0.5 wt% Pt/CuO/Al O
1 wt% Pt/CuO/Al O
2 wt% Pt/CuO/Al O
4 wt% Pt/CuO/Al O
20
20
20
20
2
3
2
3
3
3
2
Fig. 1 displays the TPD profiles of ammonia and formed
products as obtained after ammonia is adsorbed at 50 ◦C on
2
◦
Fig. 1. TPD profiles of ammonia and formed products as recorded after adsorption of ammonia at 50 C on the preoxidised 1 wt% Pt/CuO/Al O .
2
3

4
G. Olofsson et al. / Journal of Catalysis 230 (2005) 1–13
Fig. 2. A comparison of (a) 1 wt% Pt/CuO/Al O and (b) 4 wt% Pt/CuO/Al O for the selective catalytic oxidation of ammonia to produce nitrogen. Feed:
2
3
2 3
700 ppm NH and either 0.5 or 8.0 vol% O in N . Space velocity: 3000 ml/(min g ).
cat
3
2
2
1 wt% Pt/CuO/Al2O3. The figure shows that desorption of
ammonia occurs from low temperatures up to 250 ◦C, with a
maximum at about 80 ◦C. This feature shows that ammonia
is comparatively strongly adsorbed at 50 ◦C. In parallel with
the ammonia desorption, formation and desorption of hydro-
gen were observed, indicating that upon heating some of the
adsorbed ammonia was activated and transformed into an
NHx type of species. Some of the hydrogen formed reacts
with oxygen to produce water, as was noted by the resem-
blance of the hydrogen and water profiles.
for both catalysts. At the highest oxygen content the selec-
tivity increases with the temperature, whereas at the lowest
oxygen content, the selectivity passes through a weak max-
imum. At low temperatures the selectivity generally is the
highest for the lowest oxygen content. The selectivity for
nitrogen on both catalysts is about 85–90% at complete am-
monia conversion.
3.3. Influence of the temperature and the oxygen content on
the selectivity for nitrogen at high ammonia conversion
The nitrogen profile in Fig. 1 presents a maximum at
225 ◦C. Of the nitrogen oxides, N2O gives a major peak ex-
tending from low temperatures up to 200 ◦C. Compared with
the nitrogen peak, this peak is small. Almost no formation of
NO2 and NO is observed in the experiment. The TPD pro-
files demonstrate that Pt/CuO/Al2O3 is active and selective
for nitrogen formation in a window around 225 ◦C, which
thus would be the most suitable temperature for use of the
catalyst.
The catalytic performance was investigated for a feed
with 700 ppm ammonia at varying temperatures from 210 to
335 ◦C and different oxygen concentrations (0.5–20.0 vol%)
in the feed. Fig. 3 displays the selectivity dependences for
nitrogen formation on 1 wt% Pt/CuO/Al2O3. The catalyst
is highly active and the conversion is above 99% at all
temperatures except 210 ◦C. For the two lowest oxygen
concentrations, 0.5 and 1.0 vol% (Figs. 3a and 3b), the se-
lectivity for nitrogen decreases with increasing temperature,
whereas for higher oxygen concentrations (2.0–20.0 vol%)
it passes through a maximum (Figs. 3c–3f). However, as can
be seen in Fig. 2, at lower temperatures there is a selectiv-
ity maximum also for the lowest oxygen concentration. The
temperature for the maximum seems to shift towards higher
temperatures with increasing oxygen content in the gas.
A comparison of the data for 210 and 335 ◦C in Figs. 3a–
3f reveals that the selectivity for nitrogen decreases with
increasing oxygen concentration. At 210 ◦C the selectivity
drops from 88 to 77% when the oxygen concentration is in-
creased from 0.5 to 20.0 vol%. The corresponding decrease
at 335 ◦C is from 85 to 68%. In the whole range of oxy-
gen concentrations, however, at intermediate temperatures
(∼ 260 ◦C) the selectivity for nitrogen is about 85% at com-
plete ammonia conversion.
3.2. Catalyst ignition profiles
The oxidation of ammonia was studied over 1 and 4 wt%
Pt/CuO/Al2O3 at temperatures from 150 ◦C up to around
230 ◦C. A feed with 700 ppm ammonia and either 0.5 or
8.0 vol% oxygen in nitrogen was passed over the catalyst
at 3000 ml/(min g) in space velocity. The results are plotted
in Fig. 2. The graphs show that the conversion of ammonia
increases with the temperature. Compared with the conver-
sions that are obtained with 8.0 vol% oxygen in the feed,
both catalysts generally give higher conversions for the feed
with 0.5 vol% oxygen. Moreover, an increase in the Pt con-
tent in the catalyst from 1 to 4 wt% has a modest influence
on the activity. For the feed with 8.0 vol% oxygen, com-
plete conversion of ammonia over 1 wt% Pt/CuO/Al2O3 and
4 wt% Pt/CuO/Al2O3 is obtained at 230 and 220 ◦C, respec-
tively, compared with 210 and 195 ◦C, respectively, for the
feed with 0.5 vol% oxygen. The selectivity for nitrogen for-
mation shows the same type of behaviour with temperature
The by-products formed over Pt/CuO/Al2O3 are NO,
NO2 (NOx), and N2O. The formation of the respective
product depends on the temperature and the oxygen concen-
tration in the gas, as illustrated in Fig. 4. With low oxygen

G. Olofsson et al. / Journal of Catalysis 230 (2005) 1–13
5
Fig. 3. The influence of the temperature and the oxygen concentration on the selectivity to nitrogen formation in ammonia oxidation on 1 wt% Pt/CuO/Al O .
2
3
Space velocity: 3000 ml/(min g ). Feed: 700 ppm NH and (a) 0.5, (b) 1.0, (c) 2.0, (d) 4.0, (e) 8.0, and (f) 20.0 vol% O in N .
cat
3
2
2
content in the gas (Fig. 4a), N2O and NO are the major by-
products. In the interval 210–335 ◦C the selectivity for N2O
is almost constant at ∼ 9%. Increase in the temperature leads
to more NOx being formed, with more NO than NO2.
For feeds with high oxygen contents, the maximum in
the selectivity for nitrogen (Fig. 3) is concurrent with a
minimum in the amount of N2O being produced (Fig. 4b).
Compared with the data in Fig. 4a, higher selectivities for
N2O are observed for the oxygen-rich feed (Fig. 4b). More-
over, with increasing temperature the selectivity for NOx
(NO + NO2) increases up to 14% at 335 ◦C, compared with
8% at the lower oxygen pressure. However, in contrast to the
data at low oxygen content, more NO2 than NO is formed.
3.4. Influence of water and CO on the performance of the
catalysts for ammonia oxidation to nitrogen
In Fig. 5 the performances of Pt/CuO/Al2O3 samples with
0.5–4.0 wt% Pt are compared for ammonia oxidation to ni-
trogen under dry and wet conditions, respectively. Under dry

6
G. Olofsson et al. / Journal of Catalysis 230 (2005) 1–13
Fig. 4. Formation of N O, NO and NO on 1 wt% Pt/CuO/Al O as a function of the reaction temperature. Space velocity: 3000 ml/(min g ). Feed: 700 ppm
cat
2
2
2 3
NH with (a) 0.5 and (b) 20.0 vol% O , respectively, in N .
3
2
2
Fig. 6. A comparison of the performances of Pt/CuO/Al O catalysts
2
3
Fig. 5. A comparison of the performances of Pt/CuO/Al O catalysts with
2
3
with different Pt contents for ammonia oxidation to nitrogen in dry
different Pt contents for ammonia oxidation to nitrogen in dry and wet
◦
and wet gas with CO. Reaction temperature: 235 C. Space velocity:
◦
gas, respectively. Reaction temperature: 235 C. Space velocity: 3600 ml/
3600 ml/(min g ). Dry feed: 700 ppm NH , 1.0 vol% CO and 8.0 vol%
cat
3
(min g ). Dry feed: 700 ppm NH and 8.0 vol% O in N . Wet feed:
cat
3
2
2
O
in N . Wet feed: 700 ppm NH , 1.0 vol% CO, 8.0 vol% O and 8 vol%
2
2 3 2
700 ppm NH , 8.0 vol% O , and 8 vol% H O in N .
3
2
2
2
H O in N .
2
2
conditions the conversion of ammonia is almost 100% on all
catalysts. Moreover, the selectivity for nitrogen is about 85%
and is independent of the Pt content. It is worth pointing out
that under the conditions of the experiments, the addition of
water vapour to the feed has no noticeable effect on the am-
monia conversion. The effect of water on the selectivity for
nitrogen is minor at low Pt contents, but there is a tendency
with increasing Pt content for the selectivity to increase. For
4 wt% Pt/CuO/Al2O3 the selectivity for nitrogen is 85 and
95% under dry and wet conditions, respectively.
in dry gas without CO (Fig. 5). The effect of water vapour on
the selectivity for nitrogen, however, is much enhanced when
ammonia is fed together with CO. In wet gas the conversions
of ammonia and CO apparently are unchanged, whereas the
selectivity for nitrogen is increased up to the 70–80% range,
depending on the Pt content. Under both wet and dry condi-
tions, N2O is the major by-product.
3.5. Influence of SO2 treatment of the catalyst
Since industrial waste gases may contain other com-
pounds in addition to ammonia, ammonia oxidation in the
presence of CO was investigated in dry and wet gases. Fig. 6
shows data for the selectivity for nitrogen for a feed with
8.0 vol% oxygen and at the almost complete conversion of
both ammonia and CO. In dry gas the selectivity for nitrogen
is approximately 56%, irrespective of the Pt content of the
catalyst, which is substantially lower than the 85% observed
The effect of pretreatment of the catalysts with SO2 was
investigated. Table 2 shows the performances of 1 and 4 wt%
Pt/CuO/Al2O3 at high ammonia conversion in wet and dry
feeds. Clearly, the data show that the pretreatment leads
to a considerable improvement of the selectivity in paral-
lel with a small decrease in the activity. For dry conditions
the SO2 treatment has no effect on the conversion, which re-
mains practically 100%, whereas the selectivity for nitrogen

G. Olofsson et al. / Journal of Catalysis 230 (2005) 1–13
7
Table 2
Ammonia conversion and selectivity to nitrogen at 235 C over Pt/CuO/
0.5 vol% oxygen and 1.0 vol% CO gives ammonia and CO
conversions on the order of 50%. The selectivity for nitro-
gen is 100%. When the oxygen concentration in the feed
is increased to 8.0 vol% the conversion of both ammonia
and CO is complete. In this case, however, the formation
of N2O is increased and the selectivity for nitrogen drops
to 60% (see also Fig. 6). Oxidation of the ammonia in a
gas with 0.5 vol% oxygen and 1.0 vol% hydrogen gives
17% ammonia conversion, compared with 73% conversion
of the hydrogen. Increase of the oxygen content in the feed
to 8.0 vol% results in complete conversion of both hydrogen
and ammonia, with a selectivity for nitrogen of 83%, com-
pared with 100% for the lower oxygen content. For the feeds
with ammonia and methane, no conversion of the methane is
observed, and the oxidation of ammonia is complete at both
oxygen levels. The selectivity for nitrogen seems to be unaf-
fected by the presence of methane, and the values are almost
identical to those obtained for the feeds with ammonia and
oxygen only.
◦
a
Al O catalysts in dry and wet gas before and after SO -treatment , re-
2
3
2
spectively
b
1 wt% Pt/CuO/Al O
4 wt% Pt/CuO/Al O
2 3
Type of feed
2
3
and treatment
with SO
Conversion Selectivity Conversion Selectivity
2
(%)
Dry feed, untreated 100
Dry feed, treated 100
Wet feed, untreated 100
(%)
(%)
(%)
79
98
89
98
100
100
100
100
84
98
95
97
Wet feed, treated
97
a
Pretreatment with 100 ppm SO in the gas as described in Section 2.
2
b
Dry feed: 700 ppm NH and 8.0 vol% O in N . Wet feed: 700 ppm
3
2
2
NH , 8.0 vol% O and 8 vol% H O in N . Space velocity: 3600 ml/
3
2
2
2
(min g ).
cat
Table 3
Ammonia oxidation in synthetic biogas. Conversion (Conv.) of NH , CO,
3
◦
H , CH and selectivity (Sel.) to N over 1 wt% Pt/CuO/Al O at 235 C
2
4
2
2 3
and a space velocity of 3600 ml/(min g
)
cat
Feed with 700 ppm NH and (%) NH (%)
Conversion (%)
3
3
For dry biogas with ammonia, CO, hydrogen, methane,
and 0.5 vol% oxygen, no conversion of methane is obtained,
and the conversions of ammonia, CO, and hydrogen are 18,
38, and 61%, respectively, corresponding to almost com-
plete consumption of the oxygen in the feed. The selectivity
for nitrogen formation is 87%. Increase in the oxygen con-
tent to 8.0 vol% leads to the almost complete conversion
of ammonia, CO, and hydrogen, whereas methane passes
unconverted. The selectivity for nitrogen drops to 55%.
Compared with dry conditions, both the ammonia conver-
sion and the selectivity for nitrogen are enhanced when there
is water in the feed. At both 0.5 and 8.0 vol% oxygen, full
conversion of the ammonia is obtained. The selectivity for
nitrogen decreases from 98 to 90% with increasing oxygen
concentration in the feed. Comparison of the respective data
in Table 3 for wet and dry biogas with 0.5 vol% oxygen in-
dicates that the presence of water vapour leads to increasing
conversion of CO at the expense of that of hydrogen. This
behaviour may be due to the water gas shift reaction, where
some CO reacts with water to form CO2 and hydrogen.
O
CO
H
CH
H O Conv. Sel. CO
H
CH
4
2
2
4
2
2
0.5
8.0
0.5
8.0
0.5
8.0
0.5
8.0
0.5
8.0
0.5
8.0
0.5
8.0
100
100
97
100
52
87
91
94
95
100
6.5
6.5
1.0
1.0
54
100
17
60 100
100
1.0
1.0
73
100
100
100
100
18
83
85
91
87
1.0
1.0
1.0
1.0
1.0
1.0
0
0
0
0
3
3
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
38
61
96
100
100
100
55 100
98 54
90 100
a
7.0
7.0
–
–
a
a
Could not be determined from the material balance for H because of the
high water content in the feed.
increases from 79–84% before to 98% after the treatment.
Under wet conditions a small decrease in the conversion can
be noticed for 1 wt% Pt/CuO/Al2O3. For both catalysts the
selectivity increases by a few percent and reaches about 98%
after the SO2 treatment.
3.7. XRD of prepared catalysts
3.6. Selective catalytic oxidation of ammonia in biogas
XRD powder diffractograms were recorded for CuO,
CuO/Al2O3, and the Pt/CuO/Al2O3 catalysts with different
Pt content. In Fig. 7 the XRD patterns for pure CuO, CuO/
Al2O3, and 4 wt% Pt/CuO/Al2O3 are shown. The CuO/
Al2O3 shows reflections from CuO (JCPDS file no. 41-254)
and γ -Al2O3 (JCPDS file no. 10-425) [21], and the 4 wt%
Pt/CuO/Al2O3 sample and the Pt/CuO/Al2O3 samples with
lower Pt content show the same reflections without any
noticeable Pt peaks. The reflections from γ -Al2O3 are con-
sistent with small crystallites. A surface area of 119 m2/g
corresponds to an average crystallite size of ∼ 14 nm. The
absence of visible Pt peaks is mainly due a Pt content in the
samples that is too small for detection by conventional pow-
der XRD.
The oxidation of ammonia was investigated on 1 wt% Pt/
CuO/Al2O3 in the presence of CO, hydrogen, methane, and
the respective absence and presence of water. For compar-
isons the oxidation of ammonia was also investigated in the
presence of each of the biogas components. The results are
shown in Table 3. For a dry feed without fuel components,
complete ammonia conversion is achieved with a selectivity
for nitrogen on the order of 90%. Addition of water to the
feed improves the selectivity for nitrogen. In the oxygen-rich
feed (8.0 vol%) the ammonia conversion is complete (see
also Fig. 5), whereas at the lower oxygen content (0.5 vol%)
the addition of water causes a small decrease in the con-
version. Oxidation of 700 ppm ammonia in the presence of

8
G. Olofsson et al. / Journal of Catalysis 230 (2005) 1–13
3.8. TPR profiles
3.9. HREM and XEDS of catalysts
A selection of recorded TPR profiles is displayed in
Fig. 8. Reduction of the pure CuO gives a single peak with
a maximum at 340 ◦C. Reduction of the CuO/Al2O3 sam-
ple gives a peak with a maximum at 230 ◦C, and a shoulder
is noticeable on the high-temperature side of the peak at
255 ◦C. For the Pt/CuO/Al2O3 samples the maximum of
the reduction peak shifts towards lower temperature with in-
creasing Pt content, from 210 ◦C for 1.0 wt% Pt/CuO/Al2O3
to 160 ◦C for 4 wt% Pt/CuO/Al2O3. Similar to the reduc-
tion profile of CuO/Al2O3, those for the Pt/CuO/Al2O3
samples seem to be composed of at least two overlapping
peaks. With increasing Pt content, the low-temperature por-
tion of the reduction profile increases at the expense of the
high-temperature part. Quantitative analysis of the hydro-
gen consumption corresponds to a CuO content of 20.4 wt%
in the CuO/Al2O3 sample. As a result of Pt deposition, it
seems that the CuO content is decreased somewhat. The re-
duction peaks for the Pt/CuO/Al2O3 samples correspond to
CuO contents in the range of 15–17 wt% for these samples.
At low magnification, the appearance of the catalysts
is quite similar (Fig. 9), whereas the Pt particles become
clearly visible at higher magnification (Fig. 10) because
of the dark contrast and the typical lattice spacings of Pt.
The size distribution of Pt particles is larger for the 4 wt%
Pt/CuO/Al2O3 sample (roughly 2–50 nm) than for the 1 wt%
Pt/CuO/Al2O3 sample (2–5 nm). The alumina support forms
typical raft-like crystals with darker patches, which we inter-
pret as areas richer in Cu, even though the Cu seems to be
otherwise well dispersed over the surface. All elements in
the samples were identified by XEDS. Elemental mapping
was performed with Al K, Cu K, and Pt L radiation. To have
clear visibility of the smallest Pt particles in STEM mode,
instrumental magnifications between 1 and 2 million × had
to be used, which limits the possibilities for X-ray maps
with good counting statistics because of drift of the sample.
Fig. 11 shows the result of HAADF imaging of the 1 wt%
Pt/CuO/Al2O3 catalyst and mapping of Al, Cu, and Pt. There
is a slight discrepancy in the maps of Al and Cu, supporting
Fig. 7. XRD patterns of CuO, CuO/Al O with 20 wt% CuO and 4 wt% Pt/
2
3
Fig. 8. TPR profiles for pure CuO, CuO/Al O and Pt/CuO/Al O with 1
CuO/Al O . The reflections from γ -Al O are marked with (*) and all
2
3
2 3
2
3
2 3
and 4 wt% Pt.
other peaks are from CuO.
Fig. 9. TEM images of (a) CuO/Al O , and (b) 4 wt% Pt/CuO/Al O . The Cu content is 20 wt% in both samples. The images show no major differences in
2
3
2 3
appearance at lower magnification.

G. Olofsson et al. / Journal of Catalysis 230 (2005) 1–13
9
the idea of a patchwise occurrence of CuOx. The 2–5-nm
Pt particles can be distinguished in the Pt map, which has a
tendency to follow the Cu map rather than the Al map.
To further verify the occurrence of Pt in conjunction with
Cu, selective imaging of Pt was performed with large de-
tection angles (116–309 mrad) for the HAADF imaging,
followed by imaging at lower angles (57–152 mrad), where
there is an increased Rutherford scattering contribution from
Cu. Keeping above 50 mrad efficiently removes Bragg scat-
tering from the image formation process. The result is shown
in Fig. 12. The profiles show an increase in background from
primarily the Cu content for the lower angle.
4. Discussion
4.1. Catalyst structure
HREM imaging of Pt/CuO/Al2O3 shows the patchwise
occurrence of copper oxide on the alumina surface (Fig. 10).
The presence of crystalline CuO in the catalysts is veri-
fied by the X-ray diffraction patterns in Fig. 7. Moreover,
elemental mapping of Cu by HAADF imaging shows that
Fig. 10. HREM image of 4 wt% Pt/CuO/Al O . The darker Pt-particles are
2
3
identified by the d
fringes, which are measured as 0.218 nm in fair agree-
111
ment with 0.226 nm in the literature. Crystallites of the alumina support are
distinguishable. Darker patches (one marked C) presumably having a higher
CuO content are frequent, and the Pt particles are primarily attached to
x
these areas.
Fig. 11. Scanning transmission image and X-ray emission maps of 1 wt% Pt/CuO/Al O . (a) STEM image using a high-angle annular dark-field detector. The
2
3
bright Pt particles are typically 2–5 nm in diameter. The boxed area was selected for mapping. (b) Mapping of Al detecting Al Kα1 radiation. (c) The map of
Pt verifies the Pt particles in (a) and a distribution following the occurrence of the Cu shown in (d). Cu is well dispersed over the support material giving rather
low counting statistics, being limited by the substantial drift of the specimen stage. Mo grids were used instead of the traditional Cu grids in order to reduce
the stray Cu signal.

10
G. Olofsson et al. / Journal of Catalysis 230 (2005) 1–13
Fig. 12. STEM bright-field and dark-field images of 1 wt% Pt/CuO/Al O . (a) Bright field image including Bragg scattering. (b) HAADF image with a
2
3
collection angle of 116–309 mrad, giving selective imaging of high-Z elements (e.g., Pt) through incoherent Rutherford scattering. (c) At lower collection
angles (57–152 mrad) a larger contribution to image intensity is due to primarily the Cu content with a smaller contribution from Al. Bragg diffraction contrast
is still excluded from the image. The intensity profile along the line trace X–Y is shown in (d). The profiles are normalised to the same particle, thus showing
quantitatively the increased scattering contribution at this longer camera length from CuO and Al O in areas surrounding the Pt particle.
x
2
3
copper is well dispersed on the alumina support and cov-
ers a large part of the surface (Fig. 11). The existence of
well-dispersed copper oxide species is in agreement with
the TPR profiles in Fig. 8, showing that the reduction peak
for CuO on alumina is shifted towards lower temperature
(230 ◦C) compared with that for bulk CuO (340 ◦C). Quan-
tification of the reduction peak for CuO/Al2O3 verified that
Cu is present as Cu(II). Thus, our results clearly demon-
strate that copper on the alumina is present mainly in the
form of CuO. No indications from XRD or HREM were ob-
tained for the formation of a CuAl2O4 phase, which is in
agreement with the report that formation of bulk CuAl2O4
requires heating above 700 ◦C [22]. However, on the basis
of the Cu map in Fig. 11 we cannot exclude the possibility
that a minor part of the Cu is present in the form of a surface
copper aluminate phase. The formation of such a structure
has been reported in the literature. Wolberg and Roth [22],
using X-ray absorption spectroscopy, reported that for low
surface concentrations of Cu on alumina, a surface phase
is formed that resembles the structure of CuAl2O4. It was
found that this phase is undetectable by XRD. Friedman
and co-workers [23], in a comprehensive investigation com-
bining various spectroscopic techniques, reported that in
the copper aluminate surface structure the cupric ions are
predominantly in a tetragonally distorted octahedral envi-
ronment. In bulk CuAl2O4 60% of the copper ions are
in tetrahedral coordination and 40% are in octahedral co-
ordination. The threshold value between the formation of
the copper aluminate surface phase and the appearance of
crystalline CuO is around 4 wt% Cu (∼ 5 wt% CuO) per
100 m2/g of γ -Al2O3 [23,24]. Thus, considering that the
CuO content in our catalysts is about 20 wt%, the appearance
of strong XRD lines from CuO (Fig. 7) does not exclude the
possibility that some of the Cu is in the form of a surface
aluminate phase. The appearance of a main reduction peak
at 230 ◦C with a shoulder at 255 ◦C in the TPR profile of
CuO/Al2O3 (Fig. 8) gives some support for the existence of
two types of copper species. Indeed, quantitative XRD mea-
surements [25] have indicated that for copper concentrations
above the threshold value, both the copper surface phase and
CuO are present on alumina. In this regard it should be men-
tioned that in the literature there is some doubt about the
formation of a copper aluminate surface phase. In an investi-
gation of Ag/CuO/Al2O3 [15] the formation of such a phase
was disregarded because of LEIS signal intensities.
The HREM and STEM results in Figs. 10–12 for Pt/CuO/
Al2O3 show that the Pt is in the form of particles that are
preferentially localised on areas with copper, which can ex-

G. Olofsson et al. / Journal of Catalysis 230 (2005) 1–13
11
plain the synergy between Pt and CuO on alumina that has
been observed in the SCO of ammonia. In previous TPD ex-
periments with adsorbed ammonia [26] we observed that the
nitrogen formation peaks for CuO/Al2O3 and Pt/Al2O3 ap-
pear at higher temperatures than the 225 ◦C that is shown in
Fig. 1 for Pt/CuO/Al2O3. Further support for the Pt parti-
cles being in direct contact with the copper species is given
by the TPR profiles in Fig. 8 for 1 wt% Pt/CuO/Al2O3
and 4 wt% Pt/CuO/Al2O3. Compared with the reduction
peak for CuO/Al2O3, the reduction peak for Pt/CuO/Al2O3
shifts towards lower temperature with increasing Pt con-
tent of the sample, suggesting spillover of hydrogen atoms
from Pt to the copper species. The reduction peak for the
Pt/CuO/Al2O3 samples is skewed, which can be explained
by the reduction first of copper species in close contact with
the Pt particles, followed by the reduction of more distant
species. This explanation is in agreement with the obser-
vation that the low-temperature part of the reduction peak
increases with increasing Pt content of the sample.
For an oxygen content in the feed of 2.0 vol% and above,
the plots in Fig. 3 show that at complete ammonia conver-
sion the selectivity for nitrogen formation passes through a
maximum when the temperature is increased. According to
Fig. 4b the maximum is a result of the formation of N2O
passing through a minimum and the formation of NOx si-
multaneously increasing with increasing temperature. The
minimum in the N2O formation agrees with the N2O profile
in Fig. 1, showing two peaks at 80 and 250 ◦C, respec-
tively. Obviously, there are two different mechanisms for
the formation of N2O. The mechanism at low temperature
may involve adsorbed O2 species, whereas the formation at
high temperatures is likely to involve monoatomic oxygen
species. At low oxygen pressure the number of adsorbed O2
species is scarce, explaining that the selectivity for nitrogen
formation shows no maximum in Figs. 3a and 3b for low
oxygen concentrations in the feed.
Considering the results in Fig. 5 and Tables 2 and 3 for
high conversion conditions, it is evident that the addition of
water vapour to the feed has almost no effect on the conver-
sion of ammonia. Depending on the reaction conditions, the
conversion either remains unchanged or shows a slight de-
crease with the addition of water vapour. On the other hand,
the data also show that the selectivity for nitrogen in most
cases is strongly improved by the addition of water. These
findings are in general agreement with literature reports
on several catalyst systems, for example, CuO/Al2O3 [10],
Pt/Al2O3 [2], Pd-ZSM-5 [2], PdO/Al2O3 [2], Pd–V2O5–
WO3/TiO2–SiO2 [3], MoO3/SiO2 [7], and Fe-ZSM-5 [13].
The observed improvement of the selectivity for nitrogen in
the presence of water can be seen as a consequence of the
increase in the number of –OH groups at the surface rel-
ative to the number of oxygen species. For the successive
stripping of hydrogen from adsorbed ammonia, only –OH
groups are needed, whereas the formation of nitrogen ox-
ides also requires oxygen species. This explanation agrees
with our previous results [26], showing that the major route
on Pt/CuO/Al2O3 for nitrogen formation proceeds via direct
oxidation of ammonia involving two NHx species without
any participation of formed NOx. In accordance with pre-
vious reports [7,10], the slightly reduced activity induced
by water may be associated with competitive adsorption be-
tween ammonia and water. That the effect is minor can be
explained by the stronger interaction of ammonia with the
acidic adsorption sites when compared with the less basic
water molecule. The present results are for high conversion
conditions, and, therefore, the effect of water on the ac-
tivity may be enhanced at lower conversion. However, the
improved selectivity for nitrogen caused by water is not an
effect of operating the catalyst below its capacity, because
at low temperature nitrogen is an inert product that conse-
quently does not react to form nitrogen oxides. Operation
of a catalyst below full capacity is common practice in in-
dustrial applications, where it is a means of prolonging the
lifetime of the catalyst load by increasing its tolerance to-
wards poisoning.
4.2. Catalytic performance of Pt/CuO/Al2O3
The TPD profiles in Fig. 1 for adsorbed ammonia and
products indicate that Pt/CuO/Al2O3 is active and selective
for nitrogen formation in the interval 200–250 ◦C. The data
in Fig. 2 from flow experiments under stationary conditions
confirm that almost complete conversion of ammonia is ob-
tained at 200–230 ◦C, with a selectivity for nitrogen of about
85%. A comparison of the conversion curves in Fig. 2 for
1 wt% Pt/CuO/Al2O3 and 4 wt% Pt/CuO/Al2O3 reveals that
the latter catalyst is only slightly more active than the former
sample. This fact is in agreement with the observations in
HREM that the size distribution of the Pt particles becomes
wider with increasing Pt content on the support. When we
consider the data in Fig. 2, we see that Pt/CuO/Al2O3 is
more active for ammonia oxidation (700 ppm) at 0.5 vol%
than at 8.0 vol% oxygen, indicating competitive adsorp-
tion of O2 and NH3. Since adsorbed O2 is an electron-
withdrawing species and NH3 has a lone pair of electrons,
competitive adsorption between the species implies that the
adsorption site is a reduced species. In a previous mecha-
nistic investigation [26] proofs were presented that the main
route for the SCO of ammonia on Pt/CuO/Al2O3 involves re-
action at the phase boundary between NHx species adsorbed
on Cu sites and oxygen adsorbed on Pt. Thus, poisoning of
the ammonia adsorption site by oxygen suggests that this
site is a reduced copper site (Cu+). It has been verified by
FTIR that ammonia is preferentially adsorbed as NH3(ad)
on CuO/Al2O3 and that NH4+ species are much less abun-
dant [1,16]. Under flow conditions with ammonia in the
stream, bands from NH3(ad) were observed up to 350 ◦C [16]
and up to 200 ◦C [1] under static conditions with adsorption
followed by evacuation. These results are in line with the
TPD profile of ammonia in Fig. 1, showing desorption of
unreacted ammonia up to 250–300 ◦C.

12
G. Olofsson et al. / Journal of Catalysis 230 (2005) 1–13
Pretreatment of Pt/CuO/Al2O3 with SO2 results in some
–OH groups are expected to form at the surface when no
water is fed. These species are active for the abstraction of
hydrogen from ammonia and the formation of nitrogen, but
they are not involved in the formation of CO2, although they
may affect to some extent the oxidation of hydrogen. As ex-
plained earlier, lowering of the OH/O surface ratio is in line
with increasing the (N2O + NOx)/N2 product ratio.
reduction of the activity and considerable improvement of
the selectivity for nitrogen, as the data in Table 2 show. Sim-
ilar findings have been reported in previous investigations.
A comparison of CuO/TiO2 catalysts produced from CuSO4
and Cu(NO3)2, respectively, showed that the catalysts had
similar activity, but the former was more selective [8]. In a
study of CuO/Al2O3 [10] it was found that introducing SO2
into the feed gives improved selectivity and reduced activ-
ity. Moreover, data for Pt/CuO/SiO2 demonstrate a strong
boost of the selectivity for nitrogen when SO2 is added to
the feed [19]. For Fe2O3/TiO2 similar trends were observed,
and this whether the sample was prepared from sulphate
or from nitrate and in the latter case subsequent treatment
with SO2- and O2-containing gas [12]. These findings point
to sulphate formation as the origin of the observed boost
in the selectivity for nitrogen (Table 2). It is well known
that CuO and Al2O3 interact with SO2 to form sulphates
[27,28]. Under our pretreatment conditions (see Section 2.2),
assuming all SO2 forms sulphate, the amount of sulphate
corresponds to ∼ 5 µmol/m2 of surface area, that is, less
than a monolayer. Therefore, most of the sulphate species
in the catalysts can be expected to be in the form of surface
species. These species seem relatively stable, as no wors-
ening of the catalyst performance was observed during the
time of the experiment (8 h). Sulphate surface species are
electron withdrawing, resulting in an increase in the Lewis
acid strength of the surrounding cation [12]. Thereby the
surface concentration of adsorbed ammonia species is in-
creased, facilitating the interaction between two adjacent
nitrogen species forming N2, in agreement with the exper-
imental results in Table 2. The minor decrease in the activity
that is observed after the pretreatment of the catalyst with
SO2 is in agreement with stronger interaction between ni-
trogen and the Cu ion, which would affect the desorption
rate. The effect, however, is not strongly noticeable under
the present reaction conditions with operation of the cata-
lyst at high conversion. Moreover, replacement of some of
the active oxygen sites with sulphate species may contribute
to the observed changes in the catalyst performance. In par-
ticular, it would limit the minor secondary route to nitrogen
formation on Pt/CuO/Al2O3, which comprises a reaction be-
tween an NHx species and formed NO2 [26].
The data in Table 3 show complete conversion of the am-
monia in a dry biogas feed with excess oxygen (8.0 vol%)
and without added water. However, compared with the data
for the corresponding feed with 0.5 vol% oxygen, a re-
markable lower selectivity for nitrogen can be noticed (55%
compared with 87%). The decrease seems to be related to
the presence of CO, since a similar change with the oxy-
gen content is obtained for the feeds with ammonia and CO
only (see also Fig. 6). Also in that case the selectivity drops
when the oxygen content is increased from 0.5 to 8.0 vol%,
whereas no corresponding decrease is observed with a feed
of ammonia and oxygen only. In the latter case complete
conversion of ammonia is obtained at both oxygen levels,
and in both cases the selectivity for nitrogen is about 90%.
The low selectivity for nitrogen formation in the presence of
CO and excess oxygen is surprising but is possibly related to
formation of carbonate species [29–31] directing the reac-
tion towards formation of nitrogen oxides, with N2O as the
main component. Carbonate species may also form without
excess oxygen in the feed (0.5 vol%). In that case, however,
the substantially higher selectivity for nitrogen may be due
to the reaction of the formed nitrogen oxides with CO to
give nitrogen and CO2. Thus, under oxygen-rich conditions
it seems that CO preferentially does not react with NO2 but
reacts with oxygen.
5. Conclusions
The present study has shown that Pt/CuO/Al2O3 is active
and selective for the oxidation of ammonia to form nitro-
gen. At a space velocity on the order of 3600 ml/(min g),
corresponding to 240 000 h−1, full conversion of ammonia
is obtained at 220–260 ◦C, with selectivities for nitrogen in
the range of 85–90%. Adding water vapour to the feed gives
improved selectivity for nitrogen (90–95%) and decreases
the activity, although the effect on the conversion is small
at high conversion levels. The fact that the catalyst is ac-
tive below 300 ◦C and at high space velocity makes it very
competitive with other catalyst systems [2,3,6–17]. Com-
pared with Pt/Al2O3 [2], Pt/CuO/Al2O3 is more selective,
and it is also much more active than CuO/Al2O3 [9,10], in-
dicating a cooperative action between Pt and copper oxide.
HREM imaging combined with elemental analysis of the
catalyst shows intimate contact between platinum particles
and copper-rich areas, facilitating cooperation between the
phases.
Concerning the oxidation of ammonia in the presence of
water and other biogas components (CO, H2, and CH4), the
data in Table 3 show that almost complete conversion of am-
monia is achieved, and the selectivity for nitrogen is 98 and
90% for feeds with 0.5 and 8.0 vol% oxygen, respectively.
For the dry biogas with low oxygen content (0.5 vol%),
compared with the corresponding wet gas, the conversion
of ammonia is much lower (18%) and the selectivity for ni-
trogen is decreased from 98 to 87%. The decrease in the
activity is a consequence of the oxygen content in the gas be-
ing almost completely consumed by the oxidation of CO and
hydrogen to form CO2 and water, respectively. Concerning
the observed decrease in the selectivity for nitrogen, fewer
Sulphating the catalyst by treatment with SO2/O2 makes
it more selective for nitrogen formation and slightly less ac-

G. Olofsson et al. / Journal of Catalysis 230 (2005) 1–13
13
tive. A selectivity of about 97–98% is obtained at complete
ammonia conversion.
[8] N.N. Sazanova, A.V. Simakov, T.A. Nikoro, G.B. Barannik, V.F. Ly-
akhova, V.I. Zheivot, Z.R. Ismagilov, H. Veringa, React. Kinet. Catal.
Lett. 57 (1996) 71.
Concerning application of Pt/CuO/Al2O3 for the SCO of
ammonia in biogas, our investigation shows that under wet
conditions the catalyst efficiently converts the ammonia to
nitrogen with good selectivity (90–98%, depending on the
oxygen content). Under dry conditions, however, the cata-
lyst has either low activity for ammonia oxidation or low
selectivity for nitrogen formation, depending on whether the
oxygen content in the feed is low or high.
[9] L. Gang, J. van Grondelle, B.G. Anderson, R.A. van Santen, J. Ca-
tal. 186 (1999) 100.
[10] T. Curtin, F. O’Regan, C. Deconinck, N. Knüttle, B.K. Hodnett, Catal.
Today 55 (2000) 189.
[11] M. Amblard, R. Burch, B.W.L. Southward, Appl. Catal. B 22 (1999)
L159.
[12] R.Q. Long, R.T. Yang, J. Catal. 207 (2002) 158.
[13] R.Q. Long, R.T. Yang, Chem. Commun. (2000) 1651.
[14] R.Q. Long, R.T. Yang, J. Catal. 201 (2001) 145.
[15] L. Gang, B.G. Anderson, J. van Grondelle, R.A. van Santen, W.J.H.
van Gennip, J.W. Niemantsverdriet, P.J. Kooyman, A. Knoester, H.H.
Brongersma, J. Catal. 206 (2002) 60.
[16] M. Yang, C. Wu, C. Zhang, H. He, Catal. Today 90 (2004) 263.
[17] S.D. Lin, A.C. Gluhoi, B.E. Nieuwenhuys, Catal. Today 90 (2004) 3.
[18] R. Burch, B.W.L. Southward, J. Catal. 195 (2000) 217.
[19] F. Dannevang, US Patent 5 587 134 (1996), to Haldor Topsøe A/S,
Denmark.
Acknowledgments
The Swedish Research Council is acknowledged for fi-
nancial support. LRW acknowledges support from the Knut
and Alice Wallenberg Foundation.
[20] J.T. Gleaves, G.S. Yablonskii, P. Phanawadee, Y. Schuurman, Appl.
Catal. A 160 (1997) 55.
[21] JCPDS International Centre for Diffraction Data, Powder Diffraction
File, Swarthmore, PA, 1991.
[22] A. Wolberg, J.F. Roth, J. Catal. 15 (1969) 250.
[23] R.M. Friedman, J.J. Freeman, F.W. Lytle, J. Catal. 55 (1978) 10.
[24] B.R. Strohmeier, D.E. Leyden, R.S. Field, D.M. Hercules, J. Catal. 94
(1985) 514.
References
[1] L.I. Darvell, K. Heiskanen, J.M. Jones, A.B. Ross, P. Simell, A. Will-
iams, Catal. Today 81 (2003) 681.
[2] Y. Li, J.N. Armor, Appl. Catal. B 13 (1997) 131.
[3] M. Ueshima, K. Sano, M. Ikeda, K. Yoshino, J. Okamura, Res. Chem.
Intermed. 24 (1998) 133.
[4] C.N. Satterfield, Heterogeneous Catalysis in Industrial Practice, sec-
ond ed., McGraw-Hill, New York, 1991, pp. 312–320.
[5] J.M. Bradley, A. Hopkinson, D.A. King, J. Phys. Chem. 99 (1995)
17032.
[6] J.J.P. Biermann, F.J.J.G. Janssen, M. de Boer, A.J. van Dillen, J.W.
Geus, E.T.C. Vogt, J. Mol. Catal. 60 (1990) 229.
[7] L. Lietti, G. Ramis, G. Busca, F. Bregani, P. Forzatti, Catal. Today 61
(2000) 187.
[25] P.W. Park, J.S. Ledford, Appl. Catal. B 15 (1998) 221.
[26] G. Olofsson, A. Hinz, A. Andersson, Chem. Eng. Sci. 59 (2004) 4113.
[27] G. Centi, A. Riva, N. Passarini, G. Brambilla, B.K. Hodnett, B. Del-
mon, M. Ruwet, Chem. Eng. Sci. 45 (1990) 2679.
[28] S.M. Jeong, S.H. Jung, K.S. Yoo, S.D. Kim, Ind. Eng. Chem. Res. 38
(1999) 2210.
[29] F. Acke, B. Westerberg, L. Eriksson, S. Johansson, M. Skoglundh,
E. Fridell, G. Smedler, Stud. Surf. Sci. Catal. 116 (1998) 285.
[30] A. Manasilp, E. Gulari, Appl. Catal. B 37 (2002) 17.
[31] J.W. Park, J.H. Jeong, W.L. Yoon, Y.W. Rhee, J. Power Sources 132
(2004) 18.