Ammoxidation of Propane on Nickel Antimonates: The Role of Vanadium as Promoter

Article abstract of DOI:10.1006/jcat.1997.1865

The catalytic behaviour of Ni-Sb mixed oxides doped by vanadium has been investigated for the ammoxidation of propane and propene to acrylonitrile. The binary nickel antimonates, with 1:1 < Ni:Sb < 1:3, were found to be active and selective in the ammoxidation of propene to acrylonitrile (selectivity >80%) but they showed no activity in propane ammoxidation till 470°C. The activity/gram and the yield in acrylonitrile (ACN)/gram presented a maximum at Ni:Sb 1:2 due to a balance between the surface area and the doping effect of antimony. With the addition of vanadium to the Ni-Sb system, the activity and productivity of the catalysts were increased markedly, both in propane and propene ammoxidation. The optimum vanadium loading in terms of ACN yield was found for NiSb₂O₆ to be V:Ni 0.2:1 atomic ratio, a compromise between activity and selectivity. It was found that sites containing vanadium are involved in the selective nitrogen insertion step in propene ammoxidation, as well as in the activation of propane. The ammoxidation of propane is a cleaner reaction than the ammoxidation of propene, as smaller amounts of hydrogen cyanide (HCN) and acetonitrile (AceN) were formed for the same yield of acrylonitrile. X-ray analysis revealed the presence of NiSb₂O₆ and free αSb₂O₄ in all samples. In the Ni-Sb vanadium doped oxides the FTIR characterisation showed that up to a V:Ni ratio of 0.2, vanadium species different from V₂O₅, and very likely interacting with the NiSb₂O₆, were formed; these species are the ones involved in propane activation. With higher loadings of vanadium, V₂O₅ species formed which are responsible for the lowering of acrylonitrile selectivity.

Full text of DOI:10.1006/jcat.1997.1865

JOURNAL OF CATALYSIS 172, 55–63 (1997)
ARTICLE NO. CA971865
Ammoxidation of Propane on Nickel Antimonates
The Role of Vanadium as Promoter
1
T. J. Cassidy, M. Pollastri, and F. Trifiro`
Dipartimento di Chimica Industriale e dei Materiali, Viale Risorgimento 4, 40136 Bologna, Italy
Received January 2, 1997; revised June 23, 1997; accepted July 28, 1997
of alkanes is gradually becoming an achievable industrial
The catalytic behaviour of Ni–Sb mixed oxides doped by vana- goal. However, the conversion of n-butane to maleic anhy-
dium has been investigated for the ammoxidation of propane dride (1) is presently the only large-scale process which uses
and propene to acrylonitrile. The binary nickel antimonates, with
a paraffin as the primary hydrocarbon reactant. In many
1
:1 < Ni :Sb < 1 : 3, were found to be active and selective in the am-
cases the use of the alkane, as opposed to the alkene, is de-
sirable from an economic viewpoint because their market
value is lower. Pilot plants already exist for propane am-
moxidation (2) and there is a considerable amount of patent
literature concerning possible catalytic formulations; see,
for example (3–7):
moxidation of propene to acrylonitrile (selectivity >80%) but they
�
showed no activity in propane ammoxidation till 470 C. The ac-
tivity/gram and the yield in acrylonitrile (ACN)/gram presented a
maximum at Ni : Sb 1 : 2 due to a balance between the surface area
and the doping effect of antimony. With the addition of vanadium
to the Ni–Sb system, the activity and productivity of the catalysts
were increased markedly, both in propane and propene ammoxida-
tion. The optimum vanadium loading in terms of ACN yield was
found for NiSb2O6 to be V : Ni 0.2 : 1 atomic ratio, a compromise
between activity and selectivity. It was found that sites containing
vanadium are involved in the selective nitrogen insertion step in
CH3–CH2–CH3 + NH3 + 2O2 ⇒ CH2 == CH–CN + 4H2O
(1H = � 632 kJ/mol).
Several groups have recently published results in the
propene ammoxidation, as well as in the activation of propane. The open literature on the role of different catalytic systems, in-
ammoxidation of propane is a cleaner reaction than the ammoxida- cluding GaSbO (8), BiVMoO (9, 10), and SbVO (11–16),
x
x
x
tion ofpropene, assmaller amountsofhydrogen cyanide(HCN)and
acetonitrile (AceN) were formed for the same yield of acrylonitrile.
X-ray analysis revealed the presence of NiSb2O6 and free �Sb2O4 in
all samples. In the Ni–Sb vanadium doped oxides the FTIR charac-
terisation showed that up to a V : Ni ratio of 0.2, vanadium species
different from V2O5, and very likely interacting with the NiSb2O6,
were formed; these species are the ones involved in propane activa-
tion. With higher loadings of vanadium, V2O5 species formed which
MoVNbTeOx (17); SnVSbOx, as well as on mechanistic as-
pects of the reaction (18). The best catalytic performance
was achieved using a V : Sb : WOx 1 :5 :1, producing yields
in acrylonitrile between 35–40% (12). The vanadium anti-
monate, having a rutile crystalline structure, is known as a
paraffin activator (12).
In this paper the catalytic behaviour in the ammoxidation
of propane and propene of pure NiSb2O6, and doped with
vanadium, was investigated. NiSb2O6 has a trirutile struc-
ture (19) (two neighbouring antimony atoms are present),
similar to the rutile structure of SbVO4, the most investi-
gated catalyst for propane ammoxidation. The aim of this
work was to stabilise vanadium species in a rutile struc-
ture similar to VSbO4, but with different neighbours, which
may contribute to the reaction mechanism increasing activ-
ity and selectivity in acrylonitrile production.
are responsible for the lowering of acrylonitrile selectivity.
�c 1997
Academic Press
1
. INTRODUCTION
Acrylonitrile is a key raw material in the production of
acrylic fibres for clothing and high-performance thermo-
plastics such as ABS (acrylonitrile-butadiene-styrene) and
SAN (styrene-acrylonitrile) resins. It isa large-scale process
and it is predicted that the worldwide demand for acryloni-
trile in the year 2000 will be over five million tons. Today,
over 90% of acrylonitrile is produced by propene ammox-
idation but there has been growing interest during the past
decade in developing a cheaper alternative. The activation
2
. EXPERIMENTAL
2
.1. Stoichiometric and Nonstoichiometric Binary
Nickel Antimonates
A series of nickel antimonates with the following atomic
ratios Ni : Sb 1 : 1, Ni : Sb 1 : 2 (stoichiometric ratio), and
1
Corresponding author.
55
0021-9517/97 $25.00
Copyright �c 1997 by Academic Press
All rights of reproduction in any form reserved.

`
CASSIDY, POLLASTRI, AND TRIFIRO
5
6
TABLE 1
data were collected on a Philips PW1050/81 diffractome-
ter, using Cu K� radiation (� = 0.15418 nm).
Calcination Temperature and Surface Areas of the Binary
and Ternary Catalysts
The catalytic data were collected using a stainless steel,
fixed bed, down flow microreactor. The normal feed gas
ratios were 10% propane, 10% oxygen, and 5% ammonia
and 2.5% propene, 10% oxygen, and 5% ammonia, with
helium as the diluent in all cases. All gas flows were in-
dividually controlled with Brooks mass flow meters; the
Catalyst
atomic ratio)
Calcination temp.
Surface area
�
2
(
( C)
(m /g)
NiSb 1/1
NiSb 1/2
NiSb 1/3
NiSbV 1/2/0.1
NiSbV 1/2/0.2
NiSbV 1/2/0.3
NiSbV 1/2/0.2
NiSbV 1/2/0.2
NiSbV 1/2/0.2
700
700
700
700
700
700
300
500
800
14
40
38
56
44
28
102
86
8
3
operating pressure was atmospheric. Generally, 2 cm of
pelleted catalyst was loaded into the reactor and a con-
tact time of 2 s was always used. Plugs of glass wool were
used to keep the catalytic bed within the reactor and a ther-
mocouple recorded the bed temperature. Analysis of the
heavy products/reactants—propane, propene, acetonitrile
(
AceN), and acrylonitrile (ACN)—was done by means of
a Carlo Erba 6000 vega series 2 FID GC. A six-port valve,
situated at the outlet of the reactor, enabled postreactor
samples to be eluted onto a Porapak QS column. The light
products/reactants—oxygen, nitrogen, CO, and CO2—were
manually extracted from the postreactor gas stream (1 ml
via a PTFE septum) and injected into a Carlo Erba 4300
TCD GC fitted with a carbo-sieve column. The ammonia
concentration was calculated by adsorption (for 5 min) in an
acid solution (15 ml 0.1 M HCl) followed by back titration
with NaOH 0.1 M. Hydrogen cyanide production was also
analysed by adsorption (in 15 ml NaOH 0.1 M) and back
titration (with AgNO3 0.1 M, after the addition of 2 ml KI
Ni : Sb 1 : 3 were prepared as follows. Nickel chloride and
antimony pentachloride were dissolved in an aqueous HCl
solution. The resulting green solution was added dropwise
into a vigorously stirred ammonium acetate buffer solution
(
pH 7), upon which the metal salts coprecipitated. Dur-
ing the coprecipitation the pH was maintained at 7 by the
dropwise addition of concentrated NH3 solution. The re-
sulting pale grey/green precipitate was washed thoroughly
with distilled water three times and centrifuged after each
�
wash. The catalyst precursor was dried overnight at 120 C,
�
the temperature was then raised slowly to 300 C held for
h, before being raised to 700 C. The catalyst precursor
1
M and 6 ml of NH3 6 M), a procedure known as the Liebig–
�
2
D e´ nig e` s method. Product analysis was performed over a se-
was left overnight (16 h) at this temperature to yield the
active catalyst. All catalysts were pressed and sieved (30–
�
ries of temperatures between 190 and 500 C, ranging from
0
to 100% conversion in O2 over an active catalyst. Four
4
0 mesh) before catalytic testing. Table 1 lists the surface
analyses were made at each temperature following a 30-min
equilibration period; the average value was recorded.
areas of the catalysts.
2
.2. Vanadium-Doped Ternary Nickel Antimonates
3
. RESULTS AND DISCUSSION
A second series of nickel antimonates was doped with
increasing vanadium loadings, from 0 to 0.3 V/Ni atomic
ratio (atomic ratio being the V : Ni ratio). The preparation
procedure was identical to that described above (Section
3
.1. Surface Area Measurements
The surface area of each catalyst was measured after cal-
2
.1) with the exception that vanadyl acetylacetonate was cination. The surface areas of the catalysts depended on:
dissolved in the correct mole ratio during the initial step (i) the metal ratio, (ii) the addition of vanadium, and (iii)
when the salts were dissolved in HCl(aq). An excess of the temperature of calcination. The trends of the three sets
the ternary catalyst (Ni :V :Sb 1 :0.2 :2) was prepared for of catalysts are clear. The binary nickel antimonate cata-
studying the effect of the calcination temperature on the lyst NiSb 1/1 has a markedly lower surface area than the
physical characterisation of the catalysts. The surface area stoichiometric (NiSb2O6) and excess antimony containing
of the ternary catalysts are listed in Table 1.
catalysts. The ternary catalysts show a decrease in surface
area as the vanadium content is increased. By increasing
�
2
.3. Characterisation
the calcination temperature from 300 to 800 C, a drastic
2
decrease in the surface area, from 102 to 8 m /g, is observed.
Surface area measurements were determined using the
BET method on a Carlo Erba Sorpty 1750 apparatus by
N2 adsorption; 0.5 g of powdered catalyst was used for
each sample. FT-IR spectra were recorded using a Perkin
3
.2. Powder XRD Studies
Figure 1 shows the XRD patterns of the three catalyst
Elmer 1750 spectrometer by meansof the conventional self- series after calcination and before catalytic testing. The re-
supporting disk technique (0.5% in KBr). Powder XRD flections at 27.2, 35.1, and 53.3 2� are the three primary

THE ROLE OF THE COMPOSITION OF CATALYSTS
57
FIG. 1. Powder XRD patterns of the binary nickel antimonates, NiSb 1/1, NiSb 1/2, NiSb 1/3; the ternary nickel antimonates, NiSbV 1/2/0.1, NiSbV
/2/0.2, NiSbV 1/2/0.3; and ternary sample NiSbV 1/2/0.2 calcinated at various temperatures, NiSbV 1/2/0.2, T = 300; NiSbV 1/2/0.2, T = 500; and NiSbV
/2/0.2, T = 700.
1
1
reflections of the NiSb2O6 tri-rutile structure (20, 24, 25). at 29.0, 30.4, and 33.8 2�, although always in small amounts.
The diffraction patterns of NiSb 1/2 and NiSb 1/3 are rel- Evidence of trace amounts of Sb6O13 can be seen in NiSb
atively broad, indicating the formation of small crystallites 1/3 (the reflection at 30.0 is clearly visible).
in agreement with the higher surface area. NiSb 1/1, on the
Upon the addition of vanadium to the nickel antimonate
other hand, has a more defined pattern. In the spectra of system, the XRD reflections show no change in position
NiSb 1/1 and 1/2, the reflections due to �-Sb2O4 can be seen (Fig. 1). Onlyan increase ofcrystallinityisobserved (SbVO4

`
CASSIDY, POLLASTRI, AND TRIFIRO
5
8
reflections are very similar to those of NiSb2O6 and would (V : Sb 0.2 : 1) and of the ternary nickel antimonates with
be difficult to identify if present in small amounts). increasing vanadium loadings. All spectra were measured
In Fig. 1 the evolution of the crystallites is followed as after calcination and before catalytic testing. In all spec-
a function of calcination temperature. It can be seen that, tra the bands of �-Sb2O4, as well as those of NiSb2O6, are
�
� 1
even at 300 C, the characteristic tri-rutile reflections are detected. The bands at � 710, 635, 600, 535, and 512 cm
present, albeit very weak and diffuse, indicating that the are attributed to NiSb2O6 (22). It should be noted that it
microcrystalline structure forms during preparation. As the is difficult to differentiate �-Sb2O4 and the NiSb2O6 from
�
temperature of calcination is raised to 500 C the reflections their IR spectra, and that NiO has no IR active bands in the
become more clearly defined and the minor nickel anti- investigated region.
monate peaks appear. Figure 1 shows that the evolution of
With the addition of vanadium, some new bands appear:
�
� 1
the �-Sb2O4 phase appears at 700 C. The catalyst calcined a band at � 995 and 840 cm , and in sample NiSbV 1/2/0.3,
�
� 1
.
at 800 C no longer has the reflections due to �-Sb2O4, but also a band at 1020 cm
�
1
rather it has the characteristics of �-Sb2O4, indicating the
phase transition of the free antimony oxide at this temper- group of free V2O5. In order to interpret the other new
The band at 1020 cm can be attributed to the V == O
ature, in agreement with the literature (21).
bands, it is useful to compare with the spectra of �-Sb2O4
doped with vanadium (V : Sb 0.2 : 1). A band present at
�
1
� 1
3
.3. FT-IR Studies
1016 cm is visible, very close to the 1020 cm band of
pure V2O5. Therefore we attribute the new bands appearing
in the vanadium-doped nickel antimonate sample to vana-
dium species interacting with the NiSb2O6 rutile. The band
Figure 2showsthe FT-IR spectra ofthe binarynickelanti-
monate NiSb2O6, of �-Sb2O4, of vanadium-doped �-Sb2O4
�
1
at 995 cm is characteristic of a weakened V == O stretch
23). The band at � 840 cm� can be attributed to a V–O–Sb
1
(
�
1
bending mode (V–O–V is at 823 cm ).
The postreaction spectra showed no changes in all cata-
lysts, apart from NiSbV 1/2/0.3, where the band associated
with free V2O5 had strongly decreased.
3
.4. Microreactor Tests
As a blank test run, the normal feed gases were passed
through the empty reactor and the exit gases analysed. Neg-
ligible conversion of propane occurred in the temperature
range used in the catalytic runs, that is between 200 and
�
5
00 C. Likewise, the homogeneous oxidation of ammonia
to dinitrogen and water reached a maximum of only 3% at
�
4
80 C.
3
.4.1. The binary system. The three binary catalysts
showed poor activity in the ammoxidation of propane; they
showed appreciable activity only at temperatures above
�
4
70 C. For this reason, propene rather than propane was
used as the hydrocarbon feed with the binary catalysts. The
actual yields in the several products of the three catalysts
�
at 390 C are plotted in Fig. 3. Figure 4 shows the selec-
tivity of the three binary catalysts at 20% conversion in
the ammoxidation of propene. The values of conversion
�
per unit surface area at T = 460 C for the three catalysts
�
2
are (Fig. 5), in declining order, NiSb 1/1 = 3.63% gm
,
�
2
� 2
NiSb 1/2 = 2.35% gm , and NiSb 1/3 = 0.87% gm . There-
fore the specific activity decreases with the increase in the
amount of antimony oxides, however, the activity per gram
maximises for the sample NiSb 1/2 owing to an increase in
surface area. All three catalysts demonstrate remarkable
selectivity; COx production was negligible and is, there-
fore, not included in the diagram. As the antimony ra-
tio increases, there is a very slight increase in selectivity
FIG. 2. FT-IR spectra of: 1, �Sb2O4; 2, vanadium doped �Sb2O4
(
V :Sb = 0.2); 3, NiSb2O6; 4, NiSbV 1/2/0.1; 5, NiSbV 1/2/0.2.2; and 6,
NiSbV 1/2/0.3.

THE ROLE OF THE COMPOSITION OF CATALYSTS
59
�
FIG. 3. Yields of products over the binary catalysts in the ammoxidation of propene. Reaction conditions: C3 = 2.5% ; NH3 5% ; O2 10% ; T = 460 C;
= 2 s.
�
�
to acrylonitrile, together with a decrease in selectivity to isothermal conditions (390 C). As the vanadium loading in-
acetonitrile and hydrogen cyanide. The nickel antimonate creases to 0.2 atomic ratio, the catalytic activity of catalyst
NiSb 1/2 produces the highest yield of acrylonitrile per improves linearly, and then increases more rapidly. Figure 7
gram of catalyst. For this reason it was chosen as the ba- compares the yield in ACN, AceN, and COx at 20% propane
sic antimonate for further studies of the effect of vanadium conversion, with respect to the vanadium loading. In the
addition.
same figure we have reported the temperature at which 20%
of conversion was reached. A loading of 0.2 atomic ratio
3.4.2. The ternary system. As mentioned above, the vanadium leads to the best compromise between ACN yield
pure nickel antimonate NiSb 1/2 is not active in propane and COx production. The catalyst containing 0.3 atomic
�
conversion at temperatures below 470 C. With increasing ratio of vanadium is less selective to ACN because deep
amounts of vanadium, the activity is markedly improved. oxidation products are favoured. NiSbV 1/2/0.1 showed
Figure 6 shows the conversion of propane/surface area at good selectivity but required higher temperatures to obtain
FIG. 4. Selectivity to ACN, AceN, and HCN in the ammoxidation of propene over the binary catalysts at 20% C3 = conversion.

`
6
0
CASSIDY, POLLASTRI, AND TRIFIRO
FIG. 5. Conversion of propene/surface area and yield in ACN/surface area as a function of nickel content reaction conditions: C3 = 2.5% ; NH3
�
5
% ; O2 10% ; T = 390 C; � = 2 s.
the desired conversion. Figure 8 shows the distribution of ing a good selectivity to ACN, above 0.2 atomic ratio of
products over the temperature range between 390 and vanadium COx production becomes significant (see Fig. 7),
�
4
62 C, obtained on catalyst NiSbV 1/2/0.2. The selectivity this occurs concurrently with the appearance of V2O5 in the
to ACN increases to a maximum between 50 and 55% at IR spectra.
�
�
450 C, above which the total oxidation products, predom-
inantly CO2, become significant. The trend in ACN yield
To investigate the effect of vanadium on the mechanism
does not reach a maximum but continues to rise, indicating of the ammoxidation reaction, tests were carried out using
that, once produced, ACN is not converted to other prod- propene as the feed gas using the ternary catalysts. The
ucts in consecutive reactions. The selectivity to propene and feed gas ratios were 2.5% propene, 5% ammonia, and 10%
�
HCN is slightly higher at lower temperatures � 390 C than oxygen. For comparison we report in Table 2 the data of
at higher ones, indicating an increase in the rate of ACN Sb : V 1 : 0.2 prepared with the same method of NiSb based
�
production from propene as the temperature increases. The system. The experimental conditions were: T = 390 C; � =
addition of vanadium increases the activity, while maintain- 2 s, and catalyst weight = 2 g.
�
FIG. 6. Conversion of propane/surface area as a function of vanadium content. Reaction conditions: C3 10% ; NH3 5% ; O2 10% ; T = 390 C; � = 2 s.

THE ROLE OF THE COMPOSITION OF CATALYSTS
61
FIG. 7. Yields of ACN, AceN, and HCN, at 20% propane conversion, over the ternary catalysts. Reaction conditions: C3 10% ; NH3 5% ; O2 10% ;
T = reported in figure; � = 2 s.
4
. GENERAL DISCUSSION AND CONCLUSIONS
Sb : Ni (2 : 1), after which it remains essentially constant.
The most active catalyst in terms of yield in ACN per unit
surface area was the NiSb 1/1 catalyst (Fig. 5). The yield
4
.1. The Key Properties of Ni–Sb Mixed Oxide
The Ni–Sb mixed oxides were inactive in propane amm- per unit area decreases strongly with increasing antimony
�
oxidation at temperatures below 470 C, but showed good content. This in turn implies that the antimony atoms, in
activityin propene ammoxidation. With propene asthe feed excess of a 1 : 2 ratio, dilute the active sites and, in doing so,
hydrocarbon, the selectivities to ACN are all high (85–95% increase the selectivity to ACN slightly (also known as site
at 20% conversion, Fig. 4), rising slightly as the antimony isolation (26)). The poor yield per gram ofcatalyst, obtained
content increases. The activity per gram presented a maxi- with NiSb 1/1, is due to its low surface area, whereas with
mum for NiSb 1/2 (Fig. 3). The surface area increases with NiSb 1/3, the decrease in yield with respect to NiSb 1/2 is
the addition of antimony up to the stoichiometric ratio of due to the increase in antimony, resulting in a poisoning
FIG. 8. Distribution of products over catalyst NiSbV 1/2/0.2 as a function of temperature. Reaction conditions: C3 10% ; NH3 5% ; O2 10% ; � = 2 s.

`
CASSIDY, POLLASTRI, AND TRIFIRO
6
2
TABLE 2
sion rises linearly with vanadium addition up to 0.2 atomic
ratio, after which there is a further dramatic increase in
catalytic activity (Fig. 6). With this jump in activity there
is an increase in carbon oxide production and a decrease
in the selectivity to acrylonitrile. From FT-IR spectra, one
A Comparison of Catalytic Data for Three Different Catalytic
Systems in the Ammoxidation of Propene
Conversion of
Selectivity to
ACN (% )
Yield in
=
Catalyst
Atomic ratio
C
(% )
ACN (% ) can relate this jump in activity, seen with 0.3 atomic ratio
3
of vanadium, to the presence of V == O groups of the V2O5
phase, responsible for deep oxidation products. Two hy-
NiSb
NiVSb
SbV
1 : 2
1 : 0.2 : 2
1 : 0.2
20
98
55
87
76
51
17.4
74.5
28
potheses are put forward to explain the FT-IR data; that
�
1
the vanadate stretches at 840 and 990 cm are due to (i) an
amorphous SbVO4, or (ii) due to vanadium species formed
by the incorporation of vanadium atoms within the nickel
antimonate tri-rutile structure. In other wordsa newternary
system has been formed.
effect (the surface area remains essentially equal). The
higher yield in ACN per gram of catalyst with NiSb 1/2
is due to the compromise between catalytic activity and se-
lectivity to ACN. The fact that the selectivity of all three
catalysts at equal conversion remains practically the same
must be in favour of the hypothesis that excess antimony
ions play essentially the role of diluting the active sites.
From XRD and FTIR analysisthe major phase in allthree
binary catalysts is the tri-rutile, NiSb2O6. No evidence for
NiO was found, which is especially surprising for the NiSb
From Table 2, the effect of vanadium in modifying the
activity of NiSb2O6, is further demonstrated. The yield to
acrylonitrile, in propene ammoxidation, is raised from 18
�
to 75% under isothermal conditions at 390 C, proving that
vanadium is involved not only in the activation of the paraf-
fin but also in the ammoxidation of the olefin. These data
indicate that the binary nickel antimonate is very selective
to ACN but shows poor activity. With the vanadium-doped
nickel antimonate, both the conversion and the selectiv-
ity to ACN remain high, resulting in a good yield in the
1
/1 catalyst.
4
.2. The Role of Vanadium
With the addition of vanadium to NiSb2O6, not only the desired product. Besides that, the ternary compound not
activity in propene ammoxidation increases, but the cata- only is more active and selective than Ni–Sb, but also than
lysts also become active in the ammoxidation of propane. V–Sb. These data seem to support the hypothesis that a new
At the same time, the selectivity to ACN is reduced due to ternary phase containing nickel, vanadium, and antimony
an increase in total combustion products. The best compro- has been formed, rather than a nickel antimonate with an
mise in terms of ACN yield was NiSbV 1/2/0.2. New FT-IR amorphous layer of SbVO4.
bands appear which are assigned to a vanadium species in-
Figure 9 compares the sum of the yields in by-pro-
teracting with the bulk tri-rutile structure. Propane conver- ducts (acetonitrile and hydrogen cyanide) in propene and
FIG. 9. Comparison of the sum of by-products (AceN and HCN) with equal ACN yields in the ammoxidation of propane and propene over NiSbV
1
/2/0.2.

THE ROLE OF THE COMPOSITION OF CATALYSTS
63
propane ammoxidation at equal yields in acrylonitrile. This
figure demonstrates that propane ammoxidation is a much
cleaner reaction than propene ammoxidation. The smaller
amounts of hydrogen cyanide and acetonitrile formed in
propane ammoxidation, notwithstanding the fact that the
mechanistic pathway is through propene, may be explained
by the low surface concentration of propene. The propene
formed in the oxidation of propane is, in fact, very quickly
transformed to acrylonitrile over nickel vanadium antimo-
nate catalysts.
6. Guttmann, A. T., Grasselli, R. K., and Brazdil, J. F., U.S. Patent
,746,641 (1988).
. Bradzil, J. F., Little, I. B., and Hazen, J. B., U.S. Patent 5,214,016
1993). [Assigned to The Standard Oil Co.]
. (a) Osipova, Z. G., and Sokolovskii, V. D., Kinet. & Katal. 20, 510,512,
1979); (b) Osipova, Z. G., Sokolovskii, V. D., and Burylin, S. Yu.,
4
7
8
(
(
Kinet. & Katal. 24, 639 (1983); (c) Osipova, Z. G., and Sokolovskii,
V. D., et al., Kinet. & Katal. 30, 494, 1251 (1983).
9. (a) Moro-oka, Y., et al., “Catal. Sci. and Technology,” Vol. 1, p. 439
(
S. Yoshoda, N. Takezawa, and T. Ono, Eds.), Kodansha, Tokyo/VHC,
Weinhein, 1990. (b) Moro-oka, Y., “New Developments in Selective
Oxidation” (G. Centi and F. Trifir o` , Eds.), p. 491, Elsevier, Amsterdam,
The coprecipitation procedure employed and a calci-
nation temperature of 700 C overnight, ensured catalysts
1990.
�
¨
1
0. (a) Minow, G., Schnabel, K. H., and Ohlmann, G., React. Kinet. Catal.
Lett. 22, 399 (1983); (b) Minow, G., and Schnabel, K. H., Z. Phys.
Chem. (Leipzig) 265, 145 (1984).
with high surface areas and with a high degree of homo-
geneity. This resulted in an intimate interaction between
vanadium, antimony, and nickel and may favour the for-
mation of surface ternary compounds that are active and
selective in the ammoxidation of propane.
Furthermore, while propene was always found as a prod-
uct over Sb–V based systems, with our system very little was
detected. This demonstrates the multifunctional properties
of the vanadium-doped nickel antimonate catalysts; that is,
not only do they activate propane, they also manage to be
very active and selective in propene ammoxidation.
1
1
1. Trifir o` , F., Catal. Today 16, 91 (1993).
2. Andersson, A., Andersson, S. L. T., Centi, G., Grasselli, R. K., Sanati,
M., and Trifir o` , F., in “New Frontiers in Catalysis” (L. Guczi et al.,
Eds.), p. 691, Elsevier Science, Amsterdam, 1993.
3. Centi, G., Grasselli, R. K., Patane, E., and Trifir o` , F., in “New Devel-
opments in Selective Oxidation” (G. Centi and F. Trifir o` , Eds.), p. 515,
Elsevier Science, Amsterdam, 1990.
1
1
4. Catani, R., Centi, G., and Trifir o` , F., Indus. Eng. Chem. Rev. 31, 107
(1992).
1
1
1
1
5. Centi, G., Grasselli, R. K., and Trifir o` , F., Catal. Today 13, 661 (1992).
6. Cavani, F., and Trifir o` , F., Appl. Catal. A 88, 115 (1992).
7. Moro-oka, Y., and Ueda, W., Catalysis [R. Soc. Chem.] 223, 11 (1994).
8. Albonetti, S., Blanchard, G., Burattin, P., Cavani, F., and Trifir o` , F.,
French Patent 94-07982, 1994. [Assigned to Rhone Poulenc Chimie.]
9. Wells, A. F., “Structural Inorganic Chemistry,” Clarendon, Oxford,
ACKNOWLEDGMENTS
1
T.J.C. is grateful to The Royal Society of Great Britain for providing a
European Scholarship. This work was sponsored by M.U.R.S.T. (Ministero
dell’Universit a` e della Ricerca Scientifica e Tecnologica).
1962.
2
2
2
0. Sala, F., and Trifir o` F., J. Catal. 41, 1 (1976).
1. Berry, F. J., Adv. Catal. 30, 97 (1981).
2. Husson, E., Repelin, Y., and Brusset, H., Spectrochim. Acta A 35, 1177
(1979).
REFERENCES
23. Bordoni, S., Castellani, F., Cavani, F., and Trifir o` , F., J. Chem. Soc.
1
2
3
. Centi, G., and Trifir o` , F., Catal. Today 3, 151 (1988).
. Seshan, K., Appl. Catal. 67, N5 (1990).
. Oshima, K., Kayo, A., Umezawa, T., Kiyono, K., and Sawaki, I., Europ.
Patent 529,853 A2, 1992. [Assigned to Mitsubishi Kasei Co.]
. Kahney, R. H., and McMinn, T. D., U.S. Patent 4,000,178, 1976. [As-
signed to the Monsanto Co.]
Faraday Trans. 90(19), 2981 (1994).
24. Ju a` rez L o` pez, R., Godjayeva, N. S., Cort e` s Corber a` n, V., Fierro,
J. L. G., and Mamedov, E. A., Appl. Catal. A 124, 281 (1995).
25. Powder Diffraction File—Inorganic Phases, International. Centre for
Diffraction Data, PA, USA (1991).
26. Grasselli, R. K., in “Surface Properties and Catalysis by Non-Metals”
(J. P. Bonnelle, B. Delmon, and B. Derouane, Eds.), p. 273, Kluwer
Academic, Amsterdam, 1982.
4
5
. Honda, T., and Terada, K., Europ. Patent 428,413 A1, 1990. [Assigned
to Mitsui Toatsu Chem. Inc.]