Selective reduction of NO by HNCO over Pt promoted Al2O3

Article abstract of DOI:10.1006/jcat.1998.2239

The activity of Pt supported on γ-Al₂O₃ towards HNCO oxidation and reduction of NO by HNCO under oxygen excess is examined under transient conditions (temperature ramps between 100 and 500°C). Formation of N₂, N₂O, NO, and NO₂ is observed. Isotope labelled ¹⁵NO is used to show the scrambling of nitrogens in N₂ (¹⁴N¹⁵N) and N₂O (¹⁴N¹⁵NO). Adsorbed species on the surface are characterised by in-situ FTIR spectroscopy in order to obtain information on reaction intermediates. Adsorbed NH_x species are discussed as important intermediates in the N₂ and N₂O formation. The absence of Pt results in a delay in HNCO ignition and only a small N₂O formation is observed.

Full text of DOI:10.1006/jcat.1998.2239

JOURNAL OF CATALYSIS 179, 528–536 (1998)
ARTICLE NO. CA982239
Selective Reduction of NO by HNCO over Pt Promoted Al2O3
,
,
,1
,
ꢀ
ꢀ
ꢀ
Filip Acke, † § Bjo¨rn Westerberg, ‡ and Magnus Skoglundh
ꢀ
Competence Centre for Catalysis, †Department of Inorganic Chemistry, ‡Department of Chemical Reaction Engineering, Chalmers University
of Technology, S-412 96 Go¨teborg, Sweden; and §Department of Inorganic Chemistry, Go¨teborg University
Received April 21, 1998; revised July 13, 1998; accepted July 14, 1998
amples are reviewed by Matyshak and Krylov (3). O2, at
high enough concentrations, has been shown to react with
these adsorbed isocyanate species (4). Isocyanate forma-
tion has also been observed for the selective reduction of
NO over supported noble metal catalysts by hydrocarbons
under lean conditions. Infra red absorption bands assigned
to adsorbed isocyanates were observed for the Rh/Al2O3
system after exposure to a C3H6/NO gas mixture (5) and
for the Pt/ꢀ -Al2O3 system after exposure to a C3H6/NO/O2
gas mixture (6). The observed absorption bands were as-
signed to -NCO adsorbed on the support, since the stability
of these species adsorbed on the noble metal is low (7–10).
It was also shown that the presence of Pt is not necessary
for -NCO formation (11).
The activity of Pt supported on ꢀ -Al2O3 towards HNCO oxida-
tion and reduction of NO by HNCO under oxygen excess is exam-
ined under transient conditions (temperature ramps between 100
and 500ꢁC). Formation of N2, N2O, NO, and NO2 is observed. Iso-
tope labelled 15NO is used to show the scrambling of nitrogens in
N2 (14N15N) and N2O (14N15NO). Adsorbed species on the surface
are characterised by in-situ FTIR spectroscopy in order to obtain
information on reaction intermediates. Adsorbed NHx species are
discussed as important intermediates in the N2 and N2O formation.
The absence of Pt results in a delay in HNCO ignition and only a
c
small N2O formation is observed.
ꢂ 1998 Academic Press
1. INTRODUCTION
The role of the adsorbed isocyanate in the reduction of
NO by CO or hydrocarbons is still the subject of consid-
erable debate. Two opposing opinions, spectators or active
intermediates, are still discussed. A role as spectator finds
support in the very short lifetime of these species on Pt (10),
in the observation that these species reside mainly on the
support (12) and in the absence of a correlation between
the intensity of an -NCO absorption band and conversion
of NO (3). Others have proposed reaction mechanisms
involving isocyanate as an active intermediate. Adsorbed
-NCO species have been suggested to participate in the for-
mation of N2, N2O, NO2 and adsorbed NH2 and NH3 (5,
13–15).
In this article the activity and selectivity of HNCO to-
wards reduction of NO under oxygen excess are exam-
ined over ꢀ -Al2O3-supported Pt at temperatures between
100 and 500ꢁC by flow reactor studies and in-situ Fourier
transform infrared (FTIR) spectroscopy. Isotope labelled
15NO is used to determine the origin of the nitrogen in the
formed reaction products. Adsorbed NHx species are dis-
cussed asimportant intermediatesin the N2 and N2O forma-
tion in accordance with the reaction mechanisms available
in the literature. The role of the noble metal is examined
by comparing the activity and selectivity of the ꢀ -Al2O3
support and Pt/ꢀ -Al2O3 sample. Finally, the role of sub-
strate adsorbed -NCO species in the reduction of NO is dis-
cussed.
Numerous combinations of catalytic materials and reduc-
ing species have been tested to reduce emissions of nitro-
gen oxides (NOx) under net-oxidising (lean) conditions. At
present, the most common technique for NOx abatement
from stationary sources is the selective catalytic reduction
by ammonia. In search for alternative techniques, a pro-
cess based on HNCO as reducing agent has been suggested:
the RAPRENOx process (1). HNCO was shown to reduce
NO in diesel exhausts at temperatures above 400ꢁC. The
HNCO was produced from the decomposition of cyanuric
acid. A reaction mechanism based on the formation of NH
and NH2 radicals was proposed; i.e., NH reacts with NO to
form N2O, whereas NH2 is addressed as the precursor of N2.
However, an experimental and theoretical study of the ho-
mogeneous HNCO oxidation showed that a temperature
of at least 1127ꢁC was needed for a conversion of about
50% (2). The oxidation reaction proceeds mainly through
an NCO intermediate, which subsequently oxidises to NO
or reacts with this NO to form N2 and N2O.
The role of adsorbed -NCO species has also been dis-
cussed related to the reduction of NO by CO or hydrocar-
bons. Surface adsorbed -NCO has been observed by means
of infra red spectroscopy in the reduction of NO by CO
over oxide-supported noble metal catalysts. Numerous ex-
1 Corresponding author. E-mail: filip@inoc.chalmers.se.
528
0021-9517/98 $25.00
c
Copyright ꢂ 1998 by Academic Press
All rights of reproduction in any form reserved.

REDUCTION OF NO BY HNCO OVER Pt/Al2O3
2. EXPERIMENTAL
2.1. HNCO Production
529
TABLE 1
Composition of the Respective Gas Mixtures for the Flow Reactor
and the FTIR Studies
The isocyanic acid (HNCO) is produced by depolymeris-
ing the cyclic trimer of cyanuric acid (16 and references
therein). Cyanuric acid starts to sublimate around 330ꢁC.
A gaseous trimer (HNCO)3 is formed. This trimer de-
polymerises into HNCO by increasing the temperature to
700ꢁC. The process is not complete and monomer, trimer, as
well as HCN, CO, CO2, NH3, and H2O are found in the gas
flow. The HCN is oxidised to HNCO using a silver catalyst
(Ag2O) at room temperature. Both monomer and trimer
are then condensed at ꢃ80ꢁC and a distillation at ꢃ32ꢁC
results in liquid HNCO without any trimer.
Gas compositiona
HNCOb
O2
NO
15NO Gas flow
Experiment
Stabilisation
(ppm) (vol% ) (ppm) (ppm) (ml/min)
Flow
200–350
-
475
-
475
-
-
-
-
748
748
748
873
reactor HNCO oxidation 200–350
9.7
9.7
9.2
study NO reduction
15NO reduction
200–350
200–350
475
FTIR
HNCO oxidation 350–550
350–550
8
8
-
-
-
500
500
study NO reduction
1000
a All gas compositions for the flow reactor studies are balanced with
Ar; the others are balanced with N2.
b Approximate values are given for the HNCO concentration since
several parameters (temperature, flow) influence the concentration and
it was experienced to be difficult to obtain reproducible HNCO amounts.
2.2. Catalyst Preparation
The catalyst support material, ꢀ -Al2O3, was prepared
by calcining boehmite (Disperal, Condea) in air at 600ꢁC
for 8 h. The calcined alumina was sieved and the 125 to
210 ꢁm fraction was used for further preparation. This frac-
tion was dispersed in distilled water. The pH was adjusted
by ammonia (NH4OH solution, 25% , Riedel-de Hae¨n) to
about 11. A platinum hydroxide solution (Tetraammine-
platinum(II)hydroxide, Johnson Matthey) was added drop-
wise under continuous stirring to yield a Pt loading of
2 wt% . The slurry was kept under continuous stirring for
1 h. The slurry was then freeze dried and finally calcined in
air at 550ꢁC for 1 h.
were initially stabilised for 30 min in a gas mixture contain-
ing HNCO and NO in Ar at 500ꢁC. Concentrations and flow
rates are given in Table 1. Temperature ramps were done in
gas flows of HNCO, NO, and O2, as well as HNCO and O2.
The HNCO was added to the reaction mixture by flowing
small amounts of Ar over liquid HNCO in a wash bottle
which was kept in an isothermic bath at ꢃ30ꢁC. The gases
used in the experiments were 99.99(9)% Ar, 99.99(9)% O2,
5000ppm NO, and 3000ppm 15NO, the latter both in Ar. The
gas flows were controlled by mass flow controllers (Brooks,
type 5850E). The amount of HNCO in the gas mixture was
initially determined by absorption of the feed gas in an
acidic aqueous solution (100 ml 0.1 M H2SO4). HNCO is
hydrolysed by H2O and forms NH+4 ions. The amount of
ammonium ions, which is thus a measure for the HNCO
concentration, was determined by titration (17). Once the
HNCO concentration was known from the wet chemical
method, the FTIR gas analyser could be calibrated; the ab-
sorption peak at 2280 cmꢃ1 was used.
2.3. Flow Reactor Studies
The flow reactor experiments were made in a vertical
fixed bed reactor. The quartz reactor (ID 22 mm, length
500 mm) had an asymmetric construction to avoid heat-
ing of the upper metal fitting and the vacuum tight Viton
O-ring. The gases were analysed using a Fourier transform
infra red gas analyser (Bruel & Kja¨r) or a quadropole mass
spectrometer (QMG 421 from Balzers). The latter was only
used for experiments using the isotope labelled 15NO. The
temperature was measured with a K-type thermocouple,
4 mm below the sample supporting sintered quartz filter.
Linear temperature ramps were obtained by controlling the
2.4. FTIR Studies
The FTIR-experiments were performed using thin discs
temperature with a second K-type thermocouple in contact (approximately 15 mg/cm2) of catalyst in a reaction cham-
with the heating coil and connected to a Eurotherm tem- ber with CaF2-windows, similar to that described by
perature controller. Temperature ramps, using cooling and Basu et al. (18). The disc was fixed in between a folded
heating rates of 6 K/min, were started at 500 and 100ꢁC, tungsten grid placed in the centre of the reaction chamber.
respectively.
The temperature was measured with a thermocouple, in
The bed consisted of a mixture of 200 mg of active ma- contact with the grid and controlled via the voltage applied
terial, ꢀ -Al2O3 catalyst or ꢀ -Al2O3 support, with 1500 mg over the grid. The reaction chamber was placed in a FTIR
of quartz sand. The quartz sand was added to reduce the spectrophotometer. All spectra were measured with a reso-
pressure drop over the bed. The space velocity, defined as lution of 4 cmꢃ1. The fresh samples were reduced in 30%
the ratio between the flow rate and the void volume of the H2 in N2 (total flow rate of 100 ml/min) at 450ꢁC for 30 min,
bed, was between 71,000 and 125,000 hꢃ1, corresponding to stabilised in a gas mixture with 5% O2, 1000 ppm NO, and
flows of 500 and 873 ml/min, respectively. The bed materials 3000 ppm C3H6 in N2 (total flow rate 500 ml/min) for 30 min

530
ACKE, WESTERBERG, AND SKOGLUNDH
and finally degassed in N2 (500 ml/min) at 550ꢁC for 30 min.
The in-situ FTIR-measurements were performed with gas
mixtures and flows according to Table 1. The experiments
commenced at 450ꢁC and the temperature was then de-
creased in steps of 50ꢁC with a 5-min interval. Spectra from
an average of 50 scans were taken the last minute of each
interval. Reference spectra were taken with pure N2 in an
otherwise identical sequence.
3. RESULTS
3.1. Flow Reactor Studies
HNCO oxidation. The activities of the Pt/ꢀ -Al2O3 cata-
lyst and the ꢀ -Al2O3 support were tested towards HNCO
oxidation. Concentrations and flow rates are given in
Table 1. The reacted HNCO (HNCOreacted), the nitrogen
balance (Nbal), the NOx (NO + NO2) and NO concentra-
tions are displayed versus ramp temperature for the Pt con-
taining catalyst (Fig. 1). The temperature axis starts at the
high temperature (500ꢁC) to stress the experimental con-
ditions, i.e. a cooling ramp. The reacted HCNO is defined
as the HNCO bypass value minus the amount of unreacted
HNCO after passing the reactor. The nitrogen balance is
calculated as the sum of the concentrations of all infra red
detectable nitrogen-containing species ([NO] + [NO2] +
2 [N2O]). The quantification of the different reaction prod-
ucts, NO2, N2O, and N2, can be found in Fig. 1 as the differ-
ence between the concentration curves of NOx and NO, and
half the difference between Nbal and NOx and HNCOreacted
and Nbal, respectively. Complete HNCO conversion is ob-
served for temperatures from the starting temperature at
500ꢁC down to 204ꢁC. Initially, the major part of the HNCO
nitrogen is oxidised to form NO, while a minor part is fur-
ther oxidised to NO2. Some N2O formation is also observed.
Decreasing the temperature results in increased formation
FIG. 2. The fate of the carbon during HNCO oxidation over a 2%
Pt/ꢀ -Al2O3 catalyst in a cooling ramp, including the concentrations for
5
( ).
CO2 (4), HNCO (♦), and HNCOreacted
of NO2 and N2O, together with a continuous decrease in NO
formation. The maxima in the NO2 and N2O formations are
observed around 380 and 260ꢁC, respectively. No formation
of NO, N2O, or NO2 is observed for temperatures below
180ꢁC. However, at this temperature there is still a consid-
erable HNCO conversion. Gas analysers equipped with an
infra red cell are limited since no homonuclear molecules,
such as N2, can be detected. The difference between the
HNCOreacted and the Nbal is then likely to be assigned to for-
mation of N2. The validity of this assignment will be shown
later. It can be observed that for temperatures between 225
and 140ꢁC, the N2 formation ispredominant, whereasabove
225ꢁC, the N2O formation becomes substantial. Overlap
between N2O, NO2, and NO is observed. The ratio be-
tween the latter two is thermodynamically (at high temper-
atures) or kinetically (at low temperatures) controlled. The
HNCO and CO2 concentration as well as the HNCOreacted
are displayed versus ramp temperature in Fig. 2. A discrep-
ancy is observed between the HNCOreacted and the CO2
production for temperatures between 232 and 134ꢁC. This
discrepancy indicates incomplete oxidation of the HNCO
carbon.
The experiment was then repeated in the absence of Pt.
The activity of ꢀ -Al2O3 towards HNCO oxidation is shown
in Fig. 3. The HNCO oxidation, displayed as HNCOreacted
in Fig. 3, is slower in the absence of Pt as is shown by T50, the
temperature of 50% HNCO conversion. An increase from
about 160ꢁC for the Pt/ꢀ -Al2O3 system to approximately
320ꢁC in the absence of Pt is observed. The shift in the T50
temperature illustrates the importance of the noble metal
in the oxidation of HNCO. Almost all HNCO nitrogen is
oxidised to NO over ꢀ -Al2O3. A small amount is further
oxidised to NO2 at higher temperatures. The ratio between
NO and NO2 is kinetically controlled. A maximum in NO2
formation of 15 ppm is observed at 460ꢁC. No N2 is formed
duringthe HNCO conversion, while onlya minor part ofthe
HNCO nitrogen is oxidised to form N2O. The maximum in
FIG. 1. Product distribution of nitrogen containing species during
HNCO oxidation over a 2% Pt/ꢀ -Al2O3 catalyst in a cooling ramp, includ-
ing the concentrations for NO. (ᮀ), NOx (4), Nbal (♦), and HNCOreacted
5
(
).

REDUCTION OF NO BY HNCO OVER Pt/Al2O3
531
around 240ꢁC. A further decrease in temperature results in
an increase in NO concentration and at 140ꢁC, the bypass
value of NO is obtained. However, there is still a signifi-
cant HNCO conversion left at this temperature. The fate
of the HNCO nitrogen around this temperature becomes
clear in Fig. 4b, where the HNCOreacted, using an offset of
475 ppm, Nbal, NOx, and NO are displayed versus ramp
temperature. The offset for the HNCOreacted was chosen so
that a zero HNCO conversion coincides with the NO by-
pass value. This was done to visualise the N2 formation.
In contrast with the HNCO oxidation (see Fig. 1), N2 for-
mation is observed for the complete investigated temper-
ature interval. At a temperature of 224ꢁC, a maximum in
the NOx reduction is found, corresponding to a conversion
of about 87% of the feed NO. Comparison of the NOx sig-
nal with the Nbal signal, shows that almost all NOx reduc-
tion can be accounted for by the amount of N2O that has
been formed. This does not mean that all the N-atoms from
the NO molecule react to form a N2O molecule as will be
shown later under 15NO reduction. Addition of NO to the
reaction mixture also affects the HNCO oxidation. This is
observed as a decrease, compared to the HNCO oxidation,
of the T50 temperature with approximately 15ꢁ. In Fig. 5 the
HNCOreacted and the HNCO and CO2 concentrations are
displayed versus ramp temperature. An incomplete HNCO
carbon oxidation is initially observed.
FIG. 3. Product distribution of nitrogen-containing species during
HNCO oxidation over a ꢀ -Al2O3 substrate in a cooling ramp, including
5
the concentrations for NO (ᮀ), NO2 (4), N2O (᭺), and HNCOreacted
(
).
N2O formation, approximately 40 ppm, is reached around
335ꢁC. The N2O formation is significantly lower than in the
presence of Pt, while the temperature of the maximum N2O
formation is shifted towards higher temperatures.
NO reduction. In a second set of experiments, the activ-
ity and selectivity of HNCO as reducing agent for the reduc-
tion of NO in oxygen excess was investigated during cooling
ramps. Concentrations and flow rates are given in Table 1.
In Fig. 4a, the NO, NO2, and N2O concentrations, as well as
the HNCOreacted are shown for Pt-ꢀ Al2O3 versus ramp tem-
perature. Initially (high temperatures), the HNCO nitrogen
is oxidised to NO and NO2. In addition, parts of the feed
NO is also oxidised by the O2. The ratio between NO and
NO2 is thermodynamically controlled at these high temper-
atures. Formation of small amounts of N2O (about 20 ppm)
are observed around 500ꢁC. Decreasingthe temperature re-
sults in an increased NO2 formation, obtaining a maximum
around 370ꢁC, followed by a maximum in N2O formation,
The activity of the ꢀ -Al2O3 support for reduction of NO
by HNCO under net oxidising conditions was also tested.
The results, displayed in Fig. 6, are very similar as for the
HNCO oxidation in the absence of noble metal: a shift in
HNCO conversion to higher temperatures in the absence
of Pt, only a minor maximum in N2O formation, no N2 pro-
duction at all and almost no NO2 formation. The only major
effect observed is the oxidation of the HNCO nitrogen to
NO.
FIG. 4. Product distribution of nitrogen-containing species during reduction of NO by HNCO under net oxidising conditions over a 2% Pt/ꢀ -Al2O3
5
catalyst in a cooling ramp, including in Fig. 4a the concentrations for NO (ᮀ), NO2 (4), N2O (᭺), and HNCOreacted
( ) and in Fig. 4b for NO (ᮀ),
5
( ).
NOx (4), Nbal (♦), and HNCOreacted

532
ACKE, WESTERBERG, AND SKOGLUNDH
TABLE 2
Correspondence between m/e Signals and Molecules as well
as Interference on These Masses by Cracking Products of Other
Molecules
m/e signal
Molecules
14N14N, CO
Cracking fragments of
28
29
30
31
44
45
46
14N14NO, CO2, HNCO
15N14NO
15N15NO, 14NO2
15NO2
15N14
N
15N15N, 14NO
15NO
14N14NO, CO2
15N14NO
15N15NO, 14NO2
FIG. 5. The fate of the carbon during reduction of NO by HNCO
under net oxidising conditions over a 2% Pt/ꢀ -Al2O3 catalyst in a cool-
The formation of 15N14N (m/e signal 29) at tempera-
tures between 400 and 150ꢁC for cooling ramps, shows a
scrambling of nitrogen during the reaction between 15NO
and H14NCO. Oxidation of the H14NCO nitrogen results
ing ramp, including the concentrations for CO2 (4), HNCO (♦), and
5
HNCOreacted
(
).
14
in NO formation (m/e signal 30) at 400ꢁC, which steadily
15NO Reduction. So far, the N2 formation has been cal-
culated from the nitrogen balance, assuming the missing
part to be N2. Homonuclear molecules, such as N2, can be
measured using a mass spectrometer (N2: m/e signal 28).
Interpretation of the m/e signal 28 is complicated since CO
and cracking products of CO2, N2O, and HNCO interfere
on this mass. Isotope labelled 15NO (m/e signal 31) can be
used to avoid this. The reduction of 15NO by HNCO under
net oxidising conditions is shown in Figs. 7a (cooling ramps,
400–100ꢁC) and b (heating ramps, 100–500ꢁC). Concentra-
tions and flows are given in Table 1. In Figs. 7a and b the
m/e signals 28, 29, 30, and 31, respective, 44, 45, and 46
are plotted versus ramp temperature. The molecules cor-
responding to the respective m/e signals and interference
of cracking products from other molecules are listed in
Table 2.
decreases under the cooling ramp and finally disappears at
220ꢁC. The absence of a peak in this signal suggests that
there is no formation of 15N15N. The small peak in the
m/e signal 28 at temperatures between 178 and 130ꢁC is
assigned, by comparison with Fig. 5, to incomplete oxida-
tion of HNCO to CO. Interference with cracking products
of CO2 is observed as the decreasing background on this
m/e signal between 400 and 120ꢁC. The scrambling of ni-
trogens is also observed in the N2O formation, i.e. 15N14NO
(m/e signal 45). No formation of 15N15NO (m/e signal 46) is
observed, and the change in m/e signal 44 can be ascribed
to the oxidation of HNCO carbon to form CO2. At 400ꢁC
the formation of 15N14NO is lower than the production of
15N14N; however, it increases faster than the 15N14N for-
mation resulting in a higher maximum value. Qualitatively,
these results correspond well to the results shown in Fig. 4b,
where the N2 concentration was calculated from a mass bal-
ance. After finishing the cooling ramp, the isotope exper-
iment was continued with a heating ramp to 500ꢁC (same
ramp speed, 6 K/min). The result is shown in Fig. 7b. It can
be observed that the overall 15NO reduction is enhanced
under heating ramps. This is due to a surface adsorption of
reactants at low temperatures. The maxima in 15N14N and
15N14NO are higher than for cooling ramps. The adsorption
effect also becomes clear when the m/e signals 28, 30, and
44 are compared for cooling respective heating ramps. Des-
orption peaks at 250ꢁC, ascribed to CO2 are observed on
mass 44 and 28 (cracking to CO). The small peak at this
temperature on mass 30 is assigned to NO formation from
the adsorbed HNCO nitrogen.
3.2. Determination of Surface-Adsorbed Species
by Infra Red Spectroscopy
FIG. 6. Product distribution of nitrogen-containing species during re-
duction of NO by HNCO under net oxidising conditions over a ꢀ -Al2O3
substrate in a cooling ramp, including the concentrations for NO (ᮀ), NO2
Infra red spectra obtained by exposing ꢀ -Al2O3 to a gas
5
mixture containingHNCO and O2 at temperaturesbetween
(4), N2O (᭺), and HNCOreacted
(
).

REDUCTION OF NO BY HNCO OVER Pt/Al2O3
533
FIG. 7. Product distribution of nitrogen- and carbon-containing species as measured by mass spectrometry during reduction of 15NO by H14NCO
under net oxidising conditions over a 2% Pt/ꢀ -Al2O3 catalyst in a cooling ramp (a) and heating ramp (b). Included are the masses 28 (᭺), 29 (ᮀ),
5
30 (4), 31 ( ), 44 (᭹), 45 (᭿), and 46 (᭡).
450 and 100ꢁC are shown in Fig. 8. Gas concentrations and Fig. 10. When Figs. 8 and 10 are compared, it is observed
flow rates are given in Table 1. Absorption bands associ- that Pt addition results in stronger surface isocyanate bands
ated with the degenerate stretching of surface isocyanate at lower temperatures. The absorbance of the surface hy-
(2263 and 2240 cmꢃ1) and with the stretching of surface hy- droxyl band (at 3550 cmꢃ1) is about the same at 450ꢁC, but
droxyl groups (at 3550 cmꢃ1) are observed at 450ꢁC. When
the temperature is decreased, bands associated with sur-
face coordinated ammonia and ammonium ions becomes
apparent. These are the bands at 1587 and 1270 cmꢃ1 as-
signed to the asymmetric and symmetric bending of sur-
face coordinated ammonia, the band at 1656 cmꢃ1 assigned
to the symmetric bending of hydrogen bonded ammonium
ions and the degenerate bands at 1500 and 1477 cmꢃ1 as-
signed to the asymmetric bending of ammonium ions. The
absorbance for all bands increases when the temperature is
decreased (Fig. 8). At 200ꢁC the asymmetric and symmetric
stretching bands of surface coordinated ammonia at 3388
and 3292 cmꢃ1 becomes apparent. No major difference is
observed when the experiment is repeated, adding NO to
the gas mixture (Fig. 9).
The infra red spectra obtained by exposing the Pt/
FIG. 8. Infra red absorption spectra of surface species on ꢀ -Al2O3
exposed to HNCO and O2. The spectra are successively recorded after
5 min at 450, 400, 350, 300, 250, 200, 150, and 100ꢁC. The wavenumber
scale is split at 2000 cmꢃ1 and the baselines are separated by 0.1.
ꢀ -Al2O3 catalyst to a gas mixture containing HNCO and
O2 at temperatures between 450 and 100ꢁC are shown in

534
ACKE, WESTERBERG, AND SKOGLUNDH
FIG. 9. Infra red absorption spectra of surface species on ꢀ -Al2O3
FIG. 11. Infra red absorption spectra of surface species on 2% Pt/
ꢀ -Al2O3 exposed to HNCO, NO, and O2. The spectra are successively
recorded after 5 min at 450, 400, 350, 300, 250, 200, 150, and 100ꢁC. The
wavenumber scale is split at 2000 cmꢃ1 and the baselines are separated by
0.05.
exposed to HNCO, NO, and O2. The spectra are successively recorded
after 5 min at 450, 400, 350, 300, 250, 200, 150, and 100ꢁC. The wavenumber
scale is split at 2000 cmꢃ1 and the baselines are separated by 0.1.
increasesmore with decreasingtemperature in the presence
of Pt. At 450ꢁC the bands associated with surface coordi-
nated ammonia and ammonium ions are absent. Instead
bands at 1560, 1298, and 1247 cmꢃ1 assigned to the asym-
metric and symmetric stretch of mono and bidentate ni-
trate can be observed. The absorbance of the nitrate bands
increases in each successive spectra until the temperature
reaches 250ꢁC. At 250ꢁC bands associated with surface co-
ordinated ammonia and ammonium ions become apparent.
The absorbance of these bands increases when the tempera-
ture is decreased. The absorbance of the overlapping bands
of mono and bidentate nitrate remains at approximately the
same level. When the same experiment is repeated with NO
added to the gas mixture (Fig. 11), no major difference is
observed in the absorbance of the surface isocyanate bands
or the bands associated with surface coordinated ammonia
and ammonium ions. The surface hydroxyl bands are some-
what more intense at the lower temperatures and the nitrate
bands are stronger at all temperatures, when compared to
the same experiment without NO.
4. DISCUSSION
The main results of the oxidation of HNCO and the re-
duction of NO by HNCO under net oxidising conditions
from both flow reactor and FTIR studies are summarised
and possible intermediates for the reaction of HNCO
with NO and/or O2 over Pt/ꢀ -Al2O3 are discussed. HNCO
oxidation over ꢀ -Al2O3 supported Pt leads to the formation
of substantial amounts of N2 and N2O (Fig. 1). In the ab-
sence of Pt only a minor formation of N2O is observed, all
of the remaining HNCO nitrogen is oxidised to NO (Fig. 3).
Bands in infra red spectra assigned to adsorbed isocyanate,
surface hydroxyl groups, and surface coordinated ammonia
and ammonium ions are observed during exposure of the
ꢀ -Al2O3 substrate to a gas mixture containing HNCO and
O2. More bands, assigned to mono and bidentate nitrate, ap-
pear in the infra red spectra when the alumina is promoted
by Pt. That these species do not appear in the absence of Pt
is not surprising since ꢀ -Al2O3 has a low activity towards
the oxidation of NO to NO2 (see Fig. 3). High conversions
of NO are observed for the reduction of NO by HNCO
over Pt/ꢀ -Al2O3. Isotope studies, using 15NO, show a mix-
ing of nitrogen isotopes in N2 (15N14N) and N2O (15N14NO);
the HNCO nitrogen is coupled with the nitrogen of a NO
molecule. An increased activity for reduction of NO is ob-
served under heating ramps, compared to cooling ramps
(Fig. 7b versus 7a). The corresponding infra red spectra
clearly show adsorbed -NCO species at temperatures be-
low about 350 to 300ꢁC. Surface coordinated ammonia and
ammonium ions are also present from 300 down to 100ꢁC
(Fig. 11). The increased activity during heating ramps is
ascribed to desorption or spillover of reducing species that
have been adsorbed on the alumina surface at low tempera-
tures. A similar effect has been discussed previously for the
reduction of NO by propene under lean conditions over
Pt/ꢀ -Al2O3, where heating ramp results were compared
FIG. 10. Infra red absorption spectra of surface species on 2% Pt/
ꢀ -Al2O3 exposed to HNCO and O2. The spectra are successively recorded
after 5 min at 450, 400, 350, 300, 250, 200, 150, and 100ꢁC. The wavenumber
scale is split at 2000 cmꢃ1 and the baselines are separated by 0.05.

REDUCTION OF NO BY HNCO OVER Pt/Al2O3
535
with steady state results. It was observed that the tempera-
ture intervalofNO reduction coincided with that ofsurface-
bound isocyanate species (6). In the absence of noble
metal, HNCO showed no activity towards the reduction of
NO.
The reaction path for the formation of N2 and N2O during
reduction of NO includes coupling of unpaired nitrogens. It
has been suggested that this coupling is enhanced if the re-
ducing agent already contains an unpaired nitrogen atom,
as is the case for ammonia, urea, and cyanuric acid (19,20).
Otto et al. (19) studied the surface reaction between NO and
ammonia over Pt/Al2O3 using nitrogen-15 isotope labelling
at temperatures between 200 and 250ꢁC. N2 was shown to
be mainly formed from the interaction of a 15NH3 molecule
FIG. 12. Relative size (by deconvolution) of the bands at 1247 (♦),
1270 (ᮀ), and 1298 (4) cmꢃ1 for Pt/ꢀ -Al2O3 exposed to HNCO and O2.
with a 14NO molecule resulting in a mixing of isotopes, i.e. HNCO and support for this is found in the FTIR results.
15N14N, whereas N2O was suggested to be predominantly Deconvolution of the bands in the 1200–1300 cmꢃ1 region
formed from the interaction of two 14NO molecules in the for the experiment with ꢀ -Al2O3 exposed to HNCO and O2
presence of chemisorbed hydrogen (14N14NO). Adsorbed shows an increase of surface bound nitrates down to 150ꢁC
NH2 species were pointed out as the intermediate lead- (Fig. 12), indicating the formation of NO. The major differ-
ing to N2 formation, whereas adsorbed HNO species were ence between the HNCO oxidation and the reduction of
suggested to give N2O. Similar results were obtained for NO by HNCO in oxygen excess is the overlap in N2 and
Cu(II)Y Zeolites at temperatures below 110ꢁC (21). The N2O formation for the latter (Figs. 1 and 4b). This overlap
experiments with isotope labelled 15NO and HNCO pre- is also observed in the experiments with isotope labelled
sented in this investigation show a similar effect for the 15NO (Fig. 7) and must be related to the presence of NO in
formation of N2, i.e. a coupling of a nitrogen from a HNCO the feed gas.
molecule with one from NO. A similar mechanism as for the
The debate on the role of support-adsorbed -NCO on the
reduction of NO by NH3 could be suggested, since surface reduction of NO has been mentioned in the Introduction.
coordinated ammonia and ammonium ions are observed in Comparison of the reduction of NO by HNCO under cool-
the in-situ infra red spectra (Figs. 8 and 9). These species ing and heating ramp conditions (Figs. 7a and b) shows an
also form in the absence of Pt, but the noble metal is neces- increased reduction of NO with a corresponding increased
sary in a further reaction to N2 and N2O. The N2O formation formation of N2 and N2O during heating ramps. This in-
appears to be controlled by a similar mechanism as the one creased reduction of NO is ascribed to the presence of ad-
for the formation of N2. No 15N15NO is observed, as could sorbed -NCO, adsorbed at the end of the cooling, as well as
be expected from the NH3 results as presented by Otto et al. at the beginning of the heating ramp. Support for the pres-
(19). On the contrary, a mixing of isotopes is formed with ence of adsorbed -NCO species is found in the overshoot
one nitrogen from the HNCO molecule and one from the in mass 44 in Fig. 7b when compared to Fig. 7a. Compar-
NO. This is in agreement with the reaction scheme as pro- ison with data from the FTIR gas analyser allowed us to
posed by Perry and Siebers (1) for the reduction of NO by ascribe this overshoot on mass 44 to the formation of CO2.
HNCO under lean conditions. NH species were suggested No reduction of NO by HNCO was observed in the ab-
to be the active intermediates forming N2O. For 15NO and sence of noble metal. This was also seen for the reduction
H14NCO, such a reaction scheme will lead to the formation of NO by ammonia over Al2O3 (19). The presence of in-
of an N2O with a mixing of isotopes. The intermediate that fra red absorption bands assigned to adsorbed -NCO in the
was suggested by Perry and Siebers to form N2 is the same absence of noble metal (11) shows that the noble metal is
as when NH3 is used as reducing agent and, indeed, similar not needed for the formation of adsorbed -NCO, but for
results are found for isotope studies with HNCO (1) and the further reaction to N2 and N2O.
ammonia (19).
Some important characteristics for a NOx reduction
Resemblance is also observed between the oxidation of mechanism over Pt/ꢀ -Al2O3 under oxygen excess can be
HNCO and ammonia, since in both cases nitrogen from the proposed by combination of results presented here and
reducing species has to be coupled. The oxidation of am- those available in the literature. The main feature for every
monia has been studied in detail and the involvement of an reaction mechanism concerning lean NOx reduction is the
adsorbed NO molecule, formed by the oxidation of NH3, removal of adsorbed oxygen, poisoning the catalytic sites.
was suggested (22–24). This adsorbed NO reacts then fur- The heating ramps with HNCO show initially no oxidation
ther with the available ammonia, as described previously. of the reducing agent. Increasing the temperature results in
A similar mechanism can be assumed for the oxidation of a partial oxidation of the HNCO, first to -NH2 and then to

536
ACKE, WESTERBERG, AND SKOGLUNDH
-NH, where the -NH2/-NH ratio is determined by the tem-
perature and the presence of oxygen on the surface. The
former has been pointed out as the precursor for N2, while
the latter was suggested to be the precursor for N2O. Finally,
at high temperatures, the reducing agent is fully oxidised to
H2O, CO2, and NO/NO2 and the Pt surface is covered by
oxygen.
REFERENCES
1. Perry, R. A., and Siebers, D. L., Nature 324, 657 (1986).
2. Glaborg, P., Kristensen, P. G., Jensen, S. H., and Dam-Johansen, K.,
Combust. Flame 98, 241 (1994).
3. Matyshak, V. A., and Krylov, O. V., Catal. Today 25, 1 (1995).
4. Chang, C. C., and Hegedus, L. L., J. Catal. 57, 361 (1979).
5. Bamwenda, G. R., Obuchi, A., Ogata, A., and Mizuno, K., Chem. Lett.,
2109 (1994).
6. Acke, F., Westerberg, B., Eriksson, L., Johansson, S., Skoglundh, M.,
Fridell, E., and Smedler, G., in “Studies in Surface Science and Cata-
lysis; Catalysis and Automotive Pollution Control IV” (N. Kruse,
A. Frennet, and J.-M. Bastin, Eds.), Vol. 116, p. 285. Elsevier,
Amsterdam, 1998.
5. CONCLUSIONS
Reduction of NO by HNCO under oxygen excess over
Pt/ꢀ -Al2O3 was shown to result in the formation of N2 and
N2O. The use of isotope labelled 15NO showed a scram-
bling of nitrogens between NO and HNCO, i.e. formation
of 14N15N and 14N15NO. A reaction mechanism based on
NHx species is proposed. Support for such a reaction mech-
anism is found in the presence of adsorbed ammonia and
ammonium ions as observed by in-situ FTIR spectroscopy.
The absence of Pt results in a delay in HNCO ignition.
Only a small formation of N2O is observed due to the ther-
modynamical instability of the reactive intermediates, i.e.
NH and NH2, at increased temperatures.
7. Unland, M., J. Catal. 31, 459 (1973).
8. Solymosi, F., Vo¨lgyesi, L., and Rasko´, J., Z. Phys. Chem. 120, 79 (1980).
9. Gorte, R. J., Schmidt, L. D., and Sexton, B. A., J. Catal. 67, 387 (1981).
10. Rasko´, J., and Solymosi, F., J. Catal. 71, 219 (1981).
11. Westerberg, B., unpublished results.
12. Dalla Betta, R. A., and Shelef, M., J. Mol. Catal. 1, 431 (1975/1976).
13. Voorhoeve, R. J. H., Patel, C. K. N., Trimble, L. E., Kerl, R. J., and
Gallagher, P. K., J. Catal. 45, 297 (1976).
14. Voorhoeve, R. J. H., Patel, C. K. N., Trimble, L. E., and Kerl, R. J.,
J. Catal. 54, 102 (1978).
15. Voorhoeve, R. J. H., and Trimble, L. E., J. Catal. 54, 269 (1978).
16. Belson, D. J., and Strachan, A. N., Chem. Soc. Rev. 11, 41 (1982).
17. Ru¨u¨telmann, M., Report OOK 93:03 (1993).
18. Basu, P., Ballinger, T. H., and Yates, J. T., Rev. Sci. Instrum. 59, 1321
(1988).
ACKNOWLEDGMENTS
19. Otto, K., Shelef, M., and Kummer, J. T., J. Phys. Chem. 74, 2690 (1970).
20. Shelef, M., Chem. Rev. 95, 209 (1995).
We thank Dr. M. Abul-Mihl at the department of Inorganic Chemistry
(Go¨teborg University) for synthesising the HNCO. This work has been
performed within the Competence Centre for Catalysis, which is financed
21. Williamson, W. B., and Lunsford, J. H., J. Phys. Chem. 80, 2664 (1976).
22. Fogel, Y. M., Nadykto, B. T., Rybalko, V. F., Shvachko, V. I., and
Korobchanskaya, I. E., Kinet. Katal. 5, 431 (1964).
by NUTEK—The Swedish Board for Industrial and Technical Develop- 23. Ostermaier, J., Katzer, J. R., and Manogue, W. H., J. Catal. 33, 457
ment, Chalmers University of Technology, AB Volvo, Saab Automobile
AB, Johnson Matthey-CSD, ABB Fla¨kt Industri AB, Perstorp AB, and
AB Svensk Bilprovning.
(1974).
24. Pusateri, R. J., Katzer, J. R., and Manogue, W. H., AIChE. J. 20, 219
(1974).