Applied Catalysis A: General 175 (1998) 147±157

The properties of cobalt oxide catalyst for ammonia oxidation

Krzysztof Schmidt-Szaøowski*, Krzysztof Krawczyk, Jan Petryk
Faculty of Chemistry, Warsaw University of Technology, Koszykowa 75, 00-664 Warsaw, Poland

Received 17 March 1998; received in revised form 9 June 1998; accepted 12 June 1998

Abstract

A cobalt oxide catalyst for ammonia oxidation of high and stable activity and selectivity has been obtained and tested in a
¯uidized bed laboratory reactor. Its active component is Co3O4. It has been found that much less nitrous oxide is produced in
ammonia oxidation on this catalyst than on platinum. The macrostructure of the cobalt catalysts, and especially the presence
of mesopores of diameter <0.1 mm have an essential effect on their activity and selectivity and on the stability of these
properties. In the macrostructure of selective and stable catalysts the share of the mesopores <0.1 mm is extremely low, while
the total pore volume amounts from some tens to a hundred and some tens of mm3/g. A hypothetical model has been presented
to explain the effect of the catalyst's macrostructure on its properties. # 1998 Elsevier Science B.V. All rights reserved.

Keywords: Ammonia oxidation; Cobalt catalyst; Co3O4; Reagent interaction; CoO formation; Macrostructure of granules;
Selectivity; Deactivation mechanism

1. Introduction activity [3,4]. The reducing agents may be ammonia

and/or intermediate products (hydrogen, or NH and
In the oxidation of ammonia to nitrogen oxide, the NH2 radicals). It has been found that the reduction of
cobalt oxide catalyst is the one which may substitute Co3O4 may proceed despite of the presence of the
the commonly used catalytic gauze of noble metal considerable excess of oxygen with respect to ammo-
alloys with platinum as the main component [1,2]. nia. This harmful reduction of Co3O4 under the action
Tricobalt tetraoxide Co3O4 is its basic active compo- of the reacting gas mixture was the main reason of the
nent. The attempts of using cobalt oxide as a catalyst delay in implementing the commercial use of cobalt
for ammonia oxidation failed for a long time because catalysts. It was noticed, however, that the catalyst's
its catalytic activity was not stable enough under the activity may be saved for a long time due to the fact
conditions of ammonia oxidation. The main reason of that the catalyst partially reduced to CoO is easily
this phenomenon was identi®ed: in the reacting gas reoxidized to Co3O4. Therefore, the deactivation can
mixture (ammonia with air containing about 10 mol% be successfully prevented by periodically moving the
of NH3) at temperatures >7008C, the active compo- catalyst to the part of the bed where the ammonia
nent of the catalyst Co3O4 was reduced with formation concentration is low. This observation has led to a
of a considerable amount of CoO of much lower concept of a ¯uidized bed [5]. The portion of the
catalyst partially reduced by the gas rich in ammonia
*Corresponding author. can recover its activity due to the reoxidation in the

0926-860X/98/$ ± see front matter # 1998 Elsevier Science B.V. All rights reserved.
PII: S0926-860X(98)00206-3
148 K. Schmidt-Szaøowski et al. / Applied Catalysis A: General 175 (1998) 147±157

other parts of the bed. Since these transformations are  composition of the reagent mixture: 10 vol% of
reversible, the catalyst does not lose its initial activity. ammonia and 90 vol% of air;
It has also been found that the cobalt oxide catalyst can  atmospheric pressure;
be successfully applied in a stationary bed as a second  total flow rate: 400 Nl/h;
layer of a two-stage reactor. In such systems, platinum  bed temperature: 720±8008C;
gauzes are used as the ®rst stage but their number is  diameter of the ball-shaped catalyst granules: 0.8±
much smaller than in the common commercial reac- 1.8 mm;
tors. The activity and selectivity of the cobalt catalyst  height of the immobile catalyst bed: 60 mm.
may be stable for a long time because of low ammonia The conversion of ammonia to nitrogen oxide was
concentration in the bed. It should be added that due to calculated from the ammonia content in the reagent
the last years' progress, more stable cobalt catalysts gas mixture at the inlet and from the NO content in the
were obtained of a quality satisfactory for commercial outlet of the reactor both determined by the gravime-
use. trical-titration method. The content of ammonia in the
The aim of this work was to investigate the relation gas, entering the reactor was determined in gas mix-
between the macrostructure of cobalt oxide catalysts ture samples taken to a previously weighed ¯ask, in
(formed by heat treatment) and their activity and which the pressure was kept at about 100 hPa. The
selectivity in the ammonia oxidation. The composition ¯ask contained some amount of 0.1 N H2SO4. The
of these catalysts under the conditions of ammonia ¯ask was weighed again and the contents were titrated
oxidation and the process of their deactivation were with an alkali. Similarly, NO was determined in the
also studied. outlet gas, the samples being taken to a ¯ask contain-
ing H2O2 and some gaseous oxygen was let in. The
2. Studies on composition, macrostructure, and content was titrated. The analyses were repeated to
selectivity of the catalyst obtain results differing by less than 0.1%. The content
of nitrous oxide was determined chromatographically
2.1. Experimental and the Nessler method was used for spectrophoto-
metric determination of ammonia [7].
The method applied for preparation of the cobalt Average composition of the granules and the com-
catalyst consists in granulating Co3O4 with aqueous position of their outer layer were determined grav-
solution of a binder and calcination of the granulated imetrically by reduction with hydrogen to metallic
product. The catalyst exhibits high mechanical cobalt [9], and the main components were identi®ed
strength and can be used in a ¯uidized bed. Tricobalt with the use of a Siemens X-ray diffractometer type D-
tetraoxide of appropriate purity was used as the start- 500.
ing material [6] (Fe ± 3.410ÿ2, Mg ± 8.010ÿ3, Mn The measurements of the pore size were performed
± 5.010ÿ2, Ni ± 5.010ÿ2, Pb ± 1.010ÿ2, Cu ± using a mercury Porosimeter AutoPore II 9220 of
5.710ÿ3, V ± below 0.02, Na ± 2.210ÿ3, K ± 0.7 Micrometrics. The true density of the catalysts was
10ÿ3, Ca ± 2.810ÿ3 wt%). It was granulated in a measured by means of a helium pycnometer AccuPyc
rotary granulator with aqueous solution of an organic 1330 of Micrometrics, and the apparent density was
binder and dried at room temperature for 48 h. After determined by the pycnometric method. The macro-
drying, the granules (containing ca. 96% Co3O4 and structure of the catalyst granules was observed using a
4% of the binder) were calcined in a vertical tubular Zeiss DSM 942 scanning microscope.
furnace for 5 h in air at a constant temperature of
10008C to 12508C. The temperature was then 2.2. Cobalt catalyst in the ammonia oxidation
decreased to 8008C and the catalysts were heated at
this temperature for 24 h, also in air. 2.2.1. Composition of the catalyst during the
The activity and the selectivity of the catalysts ammonia oxidation
in ammonia oxidation was tested in an isothermal Four samples of the catalyst, calcined in different
¯uidized bed in a quartz-glass reactor (17 mm in temperatures (1000±12508C) and divided in two
diameter) under the following conditions [5,7,8]: groups, were chosen for investigation. Group I
 K. Schmidt-Szaøowski et al. / Applied Catalysis A: General 175 (1998) 147±157 149

Table 1
Properties of catalysts: A ± catalyst before the activity measurement in the ammonia oxidation process, B ± after the activity measurement, K ±
catalyst for pilot plant studies

Determination Group I Group II K

Cal. temp. (8C) Cal. temp. (8C) Cal. temp. (8C)

1000 1050 1200 1250 1250

Average composition of the O/Co granules (mol/mol) A 1.31 1.31 1.25 1.23 1.26
B 1.31 1.31 1.29 1.31 1.31

Composition of the external layer (about 1% of the total mass) A 1.38 ± ± ± 1.37
O/Co (mol/mol)
B 1.47 1.40 ± 1.43 1.45

Apparent density (g/cm3) A 3.28 3.81 4.90 5.06 5.09

B 3.32 3.94 5.00 5.05 5.17

Real density (g/cm3) A 6.14 6.10 ± 6.03 5.97

B 6.21 6.09 ± 6.02 5.98

Porosity (%) (from porosimetric measurements) A 45.3 36.6 17.6 15.2 15.2
B 44.1 33.5 17.5 15.3 15.3

Total porosity (mm3/g) (from porosimetric measurements) A 138.1 96.0 35.9 30.0 27.8
B 132.9 85.0 35.0 30.3 24.7

Average crystallites size (nm) A 76 ± ± 87 85

B 77 ± ± 92 96

Lattice constant (nm) Co3O4 A 0.8080 ± ± 0.8081 0.8081

B 0.8079 ± ± 0.8081 0.8081

X-ray analysis A Co3O4 ± ± Co3O4 ‡ CoO Co3O4 ‡ CoO

B Co3O4 ± ± Co3O4 ‡ CoO Co3O4 ‡ CoO

contained two catalysts calcined at 10008C and The catalysts were tested in a ¯uidized bed reactor
10508C. Their average composition (determined by (Section 2.1). They worked in an ammonia oxidation
reduction with hydrogen) was close to that of stoichio- gas mixture at temperature 720±8008C for at least
metric Co3O4 (O/Coˆ1.33). These samples are fea- 13 h, and after a series of measurements they were
tured by a ®ne-grained macrostructure (Fig. 1), cooled in closed reactor in an atmosphere of the gas
considerable porosity and small apparent density reactants (NOx, N2, O2). On the basis of reduction with
(Table 1). Group II comprised catalysts calcined at hydrogen it was found that the average composition of
a much higher temperature: 12008C or 12508C. These all the samples used (both from Group I and II) was
catalysts contained less oxygen with respect to stoi- almost identical, close to stoichiometric value of
chiometric Co3O4. Their macrostructure was coarse- Co3O4 (Table 1). Thus, the catalysts of Group II,
grained, their porosity was low, and their apparent which contained considerable amounts of CoO due
density high. X-ray analysis showed that in Group II to the thermal treatment, were reoxidized under the
also CoO is present besides to Co3O4. action of the reactant gas mixture. However, X-ray
150 K. Schmidt-Szaøowski et al. / Applied Catalysis A: General 175 (1998) 147±157

Fig. 1. Structure of the surface and interior of catalyst particles calcined at 10008C (Group I) and 12008C (Group II): 1,2 ± Surface layer of
catalyst calcined at 10008C before and after the activity measurements, respectively; 3,4 ± Surface layer of catalyst calcined at 12008C before
and after the activity measurements, respectively; 5,6 ± Structure of the interior of catalyst granules calcined at 10008C before and after the
activity measurements, respectively; 7,8 ± Structure of the interior of catalyst granules calcined at 12008C before and after the activity
measurements, respectively.
 K. Schmidt-Szaøowski et al. / Applied Catalysis A: General 175 (1998) 147±157 151

analysis of Group II catalysts, taken from the reactor

after measurements, still showed the presence of small
amounts of CoO, and thus the reoxidation to Co3O4
was not complete. It appears that in the catalysts
calcined at higher temperatures (1200±12508C) cobalt
oxide is so stable that it remains in the interior of
granules even after long residence (13 h) under the
conditions of ammonia oxidation. CoO does not occur,
however, in Group I catalysts calcined at 1000±
10508C, neither prior to nor after the experiment.
Moreover, it was noticed that granules of the catalyst,
taken from the bed after measurement, were not
homogeneous. This was found on studying the com-
position of the outer layer of the granules ± about 1%
of their total mass (Table 1). While the average gran-
ule composition is close to that of Co3O4 and the mean
O/Co molar ratio is 1.29±1.31, in the outer layer this
value was within 1.40±1.47, thus evidently over-stoi-
chiometric. (It is known that tricobalt tetraoxide can
occur with a considerable excess of oxygen with
respect to stoichiometric Co3O4. This fact was Fig. 2. Catalyst selectivity in ammonia oxidation. PA-100, PA-105
and PA-125: catalysts calcined at 10008C, 10508C and 12508C,
observed in former studies, e.g. [10,11], and in our
respectively. The numbers at the points correspond to the order of
experiments [7]). The high degree of oxidation in the measurements (for the catalyst calcined at 12508C PA-125).
outer layer of the catalyst, established under the
ammonia oxidation conditions, gives an evidence of
a strong interaction of the catalyst surface with the gas volume was suf®cient for complete ammonia conver-
reactants. sion to proceed. X1 and X2 were found on the basis of
In addition, a special experiment was carried out gas analysis and it could be assumed that the remain-
with a sample of catalyst granules previously calcined ing portion of ammonia was converted to N2 (X3ˆ
at 12508C. This sample was collected directly from the 100%ÿX1ÿX2). The measurements were carried out
¯uidized bed, operating under the ammonia oxidation at 720±8008C range selected on the basis of previous
conditions, and cooled in an atmosphere of nitrogen experiments. At temperatures >8008C the catalyst was
free of oxygen. In the surface layer (about 1% of total found to lose its activity because of the reduction of
mass) the O/Co molar ratio was 1.40, whereas the Co3O4 to CoO.
average ratio for the whole granule (1.23) was evi- It seems interesting to note that the conversion
dently smaller than in the stoichiometric Co3O4. Thus, indexes X1 and X2 were almost independent of tem-
it can be seen that the composition of granules perature of the catalyst bed (Fig. 2, Table 2). This may
becomes complex under conditions of the ammonia result from the selection of the experimental condi-
oxidation with the external layer being different from tions: the mass of catalyst bed and the gas ¯ow rate.
the bulk. Moreover, it should be noticed that the diffusion
resistance was a controlling factor of the catalytic
2.2.2. Selectivity of the catalyst reaction rates. In the ¯uidized bed, external diffusion
Under conditions of the experiment, the conversion of the gas reactants limits the rate of the reactions
of ammonia to NO (X1), to N2O (X2) and to N2 (X3) proceeding on the catalyst surface [7]. However, a
could characterize the selectivity of the catalyst bed, more precise analysis of these phenomena is not
taking into account that the overall ammonia conver- possible till now because there are no suf®cient data
sion was 100%. The residence time (about 1.8 on the mechanism and the kinetics of the reactions
10ÿ2 sec) of the reacting gas mixture in the bed involved in the ammonia oxidation on cobalt catalysts.
152 K. Schmidt-Szaøowski et al. / Applied Catalysis A: General 175 (1998) 147±157

Table 2
Measurements of the catalyst selectivity in ammonia oxidation

Sample and calcination Temperaturea of the Conversion of Conversion of Conversion of

temperature catalyst bed (8C) NH3 to NO, X1 (%) NH3 to N2O, X2 (%) NH3 to N2, X3 (%)

Group I, PA-100, 10008C 783 94.2 0.14 5.66

758 93.8 0.27 5.93
798b 93.9 0.18 5.92
733b 94.2 0.35 5.45
779b 94.0 0.20 5.80

Group I, PA-105, 10508C 775 92.1 0.54 7.36

759 92.6 0.71 6.69
742b 92.5 0.89 6.61
796b 91.8 0.69 7.51
774b 92.9 0.59 6.51

Group II, PA-120, 12008C 782 92.5 1.29 6.51

726 90.4 1.98 7.62
795b 89.9 1.60 8.50
774b 88.5 1.93 9.57

Group II, PA-125, 12508C 786 91.4 0.98 7.62

756 90.5 1.33 8.17
728b 88.4 1.35 10.25
787b 87.0 1.45 11.55

K, 12508C 784 95.5 1.60 2.90

771 93.9 1.44 4.66
732b 90.8 2.08 7.12
799b 94.8 1.91 3.29
788c 93.5 ± ±
751c 94.3 1.60 4.10
Mixture of reagents: 10 vol% of NH3 and 90 vol% of air.
a
Results of particular measurements are given in the order of their performance.
b
Second period of measurements (after a break in the operation).
c
Third period of measurements (after two breaks in the operation).

The measurement of ammonia conversion indexes were deactivated starting from the ®rst hours of the
X1 and X2 showed an essential difference between measurements. During a 13 h work, X1 decreased by
Group I catalysts (PA-100 and PA-105) calcined at 4%, and at the same time X2 increased considerably.
10008C and 10508C, and those from Group II (PA-120 As mentioned above, the deactivation of cobalt
and PA-125) calcined at 12008C and 12508C (Fig. 2, catalysts in the course of ammonia oxidation may
Table 2). Group I is characterised by a favourable be due to reduction of active Co3O4 to relatively
selectivity which is evidenced by the high conversion inactive CoO [3,4]. Probably this process is not
to NO (X1 ˆ92±94%), and a low conversion to N2O effected throughout the whole granule volume, but
(X2ˆ0.2±0.9%). No deactivation of these catalysts is limited to the surface layer, where most of the gas
was found during the series of measurements carried reactants participate in the reaction. However in our
out over two or three days (ca. 13 h of operating time). experiments, another phenomenon was observed. Cat-
The selectivity of Group II catalysts was much lower: alysts that lost rapidly their activity, featured a
X1ˆ87±92%; X2ˆ1±2%. Moreover, these catalysts strongly oxidized surface layer with over-stoichio-
 K. Schmidt-Szaøowski et al. / Applied Catalysis A: General 175 (1998) 147±157 153

metric contents of oxygen. The fact is not easy for oxidation conditions resulting in a considerable grain
interpretation because the true composition of the growth.
catalyst in the working bed is not known (especially The results of porosimetric study of catalysts from
in the bottom zone, where the ammonia concentration the two groups (Figs. 3 and 4) also showed some
is maximum). It should be taken into account that the essential differences. Porosimetric measurements of
rate of the catalyst's conversion in the surface layer of other cobalt catalysts were also performed: catalyst K,
the granules may be very high. Thus it cannot be produced from the same material as Groups I and II
excluded that in the bottom zone of the bed the surface (for semi-industrial experiment), and two industrial
layer of the granules was really reduced, whereas the catalysts named here K1 and K2. All the results are
samples used for the analysis were taken from the discussed in Section 3.
upper zone, where a reoxidation proceeded due to the
low ammonia concentration.
3. Discussion
2.2.3. Macrostructure of the catalyst
The difference between the macrostructures of cat- 3.1. Relationships between the macrostructure and
alysts from Groups I and II was evident (Fig. 1). It is the selectivity of the catalyst
worth noting that the ®ne-grained macrostructure of
Group I catalysts did not undergo essential transfor- The newly prepared catalysts in both Groups I and
mation within the time of their work (overall ca. 13 h) II were different in composition and macrostructure of
under conditions of ammonia oxidation. Group II the granules. However, the composition of the gran-
catalysts behaved differently. Their initial coarse- ules, attained under the conditions of ammonia oxida-
grained macrostructure formed as a result of calcina- tion, was similar in both groups. On the other hand,
tion, underwent transformation under the ammonia considerable differences were noted in the macro-

Fig. 3. Pore volumes of catalysts after activity measurements. The overall porosity and distribution of pore volumes in the Group I and II
catalysts calcined at 1000±12508C. Pore diameters are shown in the diagram.
154 K. Schmidt-Szaøowski et al. / Applied Catalysis A: General 175 (1998) 147±157

Fig. 4. Pore volumes of catalysts after activity measurements. Group I (PB-100, PB-105) and II (PB-120, PB-125) catalysts calcined at 1000±
12508C, and three other catalysts characterized by high selectivity: K, K1 and K2. Data are presented in Table 1.

structure of the granules, both in new catalysts and in KII (Section 3 and Figs. 3 and 4). They can be sum-
that taken from the used bed. It might be concluded, marised as follows:
therefore, that the main and primary reason of differ- 1. The total volume of pores of selective catalysts of
ences in catalytic properties of these groups was Group I is much higher than that of Group II
connected with their macrostructure. The ®ne-grained catalysts.
macrostructure of Group I catalysts was stable under 2. A peculiar feature of Group I (of high and stable
ammonia oxidation conditions and their selectivity selectivity) consists in the absence of mesopores of
was also stable. The observed drop in selectivity in diameter <0.1 mm.
Group II was associated with transformation of their 3. With rising calcination temperature the volume of
macrostructure. It should be noted, however, that this the macropores >0.1 mm decreases, and at the same
is not a suf®cient proof of a causal-result connection time the volume of the mesopores <0.1 mm
between these phenomena. For example, the studies of increases. Therefore, in catalysts from Group II
Morozow et al. [12] on heat treatment of the catalyst the share of mesopores is significant. That group is
from pure Co3O4 at ca. 10008C showed that the characterised by a low selectivity and it undergoes
recrystallization occurring under these conditions deactivation under the conditions of ammonia oxi-
led to an essential decrease in its speci®c surface, dation.
i.e. a change in macrostructure of the granules took 4. Another cobalt catalyst K (produced from the same
place. No effect, however, was found of this phenom- material as Group I and II catalysts) exhibits a high
enon on the catalytic activity of thus treated sample. selectivity, as in Group I, despite of the small total
Some conclusions concerning the relation between pore volume which is close to that of low-selec-
the catalysts' macrostructure and their selectivity tivity Group II catalysts. However, like in Group I,
result from porosimetric studies of Groups I and II, nearly all this volume consists of macropores
including also some other catalysts named K, KI and (>0.1 mm) with a negligible share of mesopores.
 K. Schmidt-Szaøowski et al. / Applied Catalysis A: General 175 (1998) 147±157 155

5. The study of other cobalt catalysts (K1 and K2),

highly selective and stable under the conditions of
ammonia oxidation, gave a characteristic result.
Both catalysts have a quite large total pore volume,
but their peculiar feature is the absence of meso-
pores (diameter <0.1 mm). All the pore volume lies
in the macropore range.
The examples presented con®rm the thesis that the
pore structure of the cobalt catalysts has an essential
effect on their activity, selectivity, as well as on the
stability of these properties under the conditions of
ammonia oxidation. Attention is drawn to the fact that
in the catalysts of high and stable selectivity, a very
small volume of mesopores (of diameter <0.1 mm)
was found. In these catalysts the macropores are the
most abundant and their overall volume is from tens to
a hundred and tens mm3/g.

3.2. A model of the catalyst's macrostructure Fig. 5. Scheme of the model macrostructures I and II: (- - - -)
active surface of the catalyst grains; (!) diffusion of the reactants.

A hypothetical model of the catalyst's macrostruc-

ture was considered in order to explain the relation macropores between the grains and with a weakly
between the macrostructure of the cobalt catalysts and developed external surface. In both types of the cat-
their activity and selectivity in ammonia oxidation. It alysts mesopores can also occur inside the grains
was based on the following, simplifying assumptions: (variant I-A and II).
1. NO, N2 and H2O are the only products of As the catalysts of macrostructure I have a well-
ammonia oxidation (the N2O formation is dis- developed external surface, the rate of ammonia oxi-
regarded). dation related to a unit of the granule geometrical
2. N2 is generated mainly from reactions between NO surface is high. Taking into account the external
and NH3 proceeding on the catalyst surface or in diffusion resistance it can be assumed that ammonia
the gas volume. concentration at the solid surface is much lower than
3. Two kinds of pores are distinguished in the cata- in the bulk stream of gases on this level of the bed
lyst: (a) macropores situated mainly between the (Fig. 5). Ammonia is converted completely within the
grains in the catalyst macrostructure and (b) meso- thin external layer of granules on the surface of
pores present in the interior of particular grains macropores between the grains. Inside the mesopores,
(Fig. 4). and also in the macropores located deeper in the
4. The catalytic process proceeds under a strong granule volume, there is no ammonia but only the
influence of external diffusion resistance, however, reaction products (NO, N2 and H2O) and the excess
the reactants penetrate also into the pores, mainly oxygen. In such a case, irrespective of the ammonia
in the external layer of the catalyst granules. conversion in the bulk of the gas stream, even the most
The model macrostructure of the types I and II external parts of the catalyst surface remain in contact
considered below corresponds approximately to the with gases of low ammonia concentration (if any), and
previously described Groups I and II of the catalysts Co3O4 is a stable component throughout the whole
(Fig. 5). The macrostructure I comprises ®ne, loosely granule volume due to the excess of oxygen.
linked grains with a quite large volume of macropores In the second kind of catalysts (macrostructure II)
(at the interstices) and has a developed external sur- one should consider two cases differing in the degree
face of the granules. The macrostructure II is coarse- of ammonia conversion in the gas stream. If the
grained (e.g. due to sintering), with small volume of ammonia conversion is small, then at the solid surface
156 K. Schmidt-Szaøowski et al. / Applied Catalysis A: General 175 (1998) 147±157

its concentration will be higher than in the case of content is higher than in stoichiometric Co3O4. The
macrostructure I. Since the external surface of these model presented here does not touch this phenom-
granules is not well developed, the rate of ammonia enon.
oxidation related to a unit of their geometrical surface
is lower than that in the macrostructure I, and despite
of the diffusion resistance the ammonia concentration 4. Conclusions
can increase at the solid surface in the external layer of
the granule. Under the action of such a reagent mixture 1. A catalyst of high and stable activity and
Co3O4 may be reduced, initially in the external layer, selectivity in ammonia oxidation to NO, generat-
and the process of the catalyst's deactivation begins. ing minimum quantity of N2O, was prepared from
The catalytic reaction rate decreases and the ammonia pure cobalt oxide Co3O4. It may be used in a
concentration increases still more. Thus the reduction ¯uidized bed reactor.
and deactivation zone may expand into the bulk of the 2. The macrostructure of cobalt catalysts (especially
granule. the presence of mesopores <0.1 mm) has an essen-
Moreover, it should be noticed that irrespective of tial effect on their activity and selectivity, as well as
the Co3O4 reduction, the macrostructure II can exhibit on the stability of these properties. It has been
a smaller selectivity of ammonia oxidation to NO than found that the share of such mesopores in the
the macrostructure I. If high ammonia concentrations structure of stable selective catalysts is extremely
occur at the surface of granules, then NH3 can easily small while the total pore volume amounts from
penetrate inside the mesopores. In such a case, parasite some tens to more than a hundred mm3/g.
reactions will proceed in the mesopores, e.g. between 3. One of the possible reasons for the deactivation of
ammonia and nitrogen oxide to form nitrogen. The the cobalt catalyst (containing pure Co3O4 in the
catalytic action of the developed surface of mesopores initial state) under the conditions of ammonia
can favour those reactions, resulting in a decrease of oxidation is the reduction of Co3O4 which results
the ammonia conversion to NO. in the formation of CoO in the external layer of the
When the conversion of ammonia in the bulk of the catalyst. However, some of our experiments sug-
gas stream is high, its concentration at the catalyst gest that another deactivation mechanism is also
surface is low, even at the most external parts. Under possible causing a decrease of the catalyst activity
such conditions, the reduction of Co3O4 does not and selectivity, when the oxygen content in the
proceed and the catalyst is not deactivated. Thus, external layer of cobalt oxide is not smaller (and
the catalyst of macrostructure II could be used for sometimes even much higher) than in stoichio-
industrial purposes, but only as the second stage of the metric Co3O4. The study of the latter effect will
ammonia oxidation reactor. be continued.
To end the discussion it is worth noting that if the
model adopted describes correctly at least the main References
phenomena occurring during the ammonia oxidation,
then the most favourable catalyst may be such as that [1] S.P.S. Andrew, G.C. Chinchen, Stud. Surf. Sci. Catal. 6
presented by variant I-B (Fig. 5).This is a catalyst with (1980) 141.
developed external surface due to the presence of [2] L.E. James, R.L. Crescentini, W.B. Fisher, US Patent
4 389 339 (1983).
macropores between the grains, but the single grains
[3] T. Popowa, D. Stawrakiewa, I. Zrncziew, Chimia and
of the macrostructure are monolithic and do not con- Industria 56 (1984) 164.
tain mesopores. It should be noted, however, that the [4] M.C. Goromonzi, S.K. Raman, B.D. Padalia et al., React.
model presented is of a hypothetical character and Kinet. Catal. Lett. 15 (1980) 131.
requires further studies for con®rmation. One of the [5] J. Petryk, K. Krawczyk, Przem. Chem. 67 (1988) 272.
[6] J. Petryk, K. Krawczyk, PL Patent 149619 (1990).
questions still left unanswered is that of the mechan-
[7] K. Krawczyk, Ph.D. Thesis, Warsaw University of Technol-
ism of catalyst deactivation under conditions, when its ogy, Warsaw, 1995.
surface layer is not reduced, but on the contrary ± it [8] K. Krawczyk, J. Petryk, K. Schmidt-Szaøowski, Stud. Surf.
becomes additionally oxidised so that the oxygen Sci. Catal. 91 (1995) 683.
 K. Schmidt-Szaøowski et al. / Applied Catalysis A: General 175 (1998) 147±157 157

[9] K. Schmidt-Szaøowski, J. Petryk, K. Krawczyk, Przem. [11] è.A. Isupowa, W.J. Alieksandrow, W.W Popowski et al., Izw.
Chem. 73 (1994) 423. Sibirskowo Otdielienia AN SSSR 1 (1989) 39.
[10] I.D. Bieøowa, J.E. Roginska, J.N. Wieniewciew, ZÇurn. Nieorg. [12] N.N. Morozow, è.I. èukianowa, M.I. Tiemkin, Kinet. Katal. 7
Chimii 28 (1983) 3009. (1966) 172.