to acrylonitrile is considered the key aspect, using the stoi-
chiometric feed (feed 112) the highest selectivity is shown by
sample SbV5a, but using feed 121 the highest selectivity is
shown by the samples SbV2a or SbV1v. In terms of acrylo-
nitrile productivity, instead, the best results are shown by
catalyst SbV2a using feed 112, by catalyst SbV5a using feed
depends on the speciÐc feed composition and is di†erent when
selectivity or productivity to acrylonitrile is considered.
In propane ammoxidation, therefore, in contrast to propene
ammoxidation where the higher hydrocarbon reactivity and
stronger chemisorption with respect to propane considerably
limits the occurrence and role of the side oxidation of
ammonia on the surface reactivity, the feed composition plays
an important role in determining the surface reactivity, but
di†erent e†ects determine the catalytic behavior depending on
the speciÐc characteristics of the catalyst. It is thus not pos-
sible to determine a unique relationship between catalyst
characteristics and catalytic behavior without taking into con-
sideration the e†ect of the feed composition on the surface
reactivity.
122 and by catalyst SbV1v using feed 121. This indicates that
in the study of the structure, activity and selectivity relation-
ships care must be given to the problem of the inÑuence of the
feed composition on the catalytic results.
The second problem in the analysis of the data shown in
Fig. 4 is to understand the relationship between feed composi-
tion and surface catalytic behavior. It may be noted in Fig.
4(c) that the global selectivity (propene plus acrylonitrile)
increases considerably for all catalysts in going from the stoi-
chiometric feed (112) to that with an excess of ammonia and
deÐciency in oxygen (feed 121) passing through that with an
excess of ammonia but the same oxygen content as the stoichio-
metric feed (feed 122). The e†ect is particularly relevant for the
samples with the lower Sb : V ratio. On the other hand, the
excess of ammonia in the feed only in some cases promotes
the selectivity and productivity to acrylonitrile [Fig. 4(b) and
References
1
G. Centi, in Elementary Reaction Steps in Heterogeneous Cataly-
sis (NATO Symp. Ser.), ed. R. W. Joyner and R. A. Van Santen,
Kluwer Academic Press, Dordrecht, 1993, p. 93.
2
3
4
5
6
7
8
9
0
Y. Moro-oka and W. Ueda, in Catalysis, The Royal Society of
Chemistry, Cambridge, UK, 1994, vol. 11 p. 223.
G. Centi, R. K. Grasselli and F. TriÐro
` , Catal. T oday, 1992, 13,
6
61.
V. D. Sokolovskii, A. A. Davydov and O. Yu. Ovsitser, Catal.
(
d)]. It may be noted, for example, that for SbV1a the e†ect of
Rev.-Sci. Eng., 1995, 37, 425.
the feed composition is parallel in terms of increase in global
selectivity and selectivity to acrylonitrile, whereas the opposite
occurs for SbV10a. Ammonia is a coreactant for the synthesis
of acrylonitrile, but an increase in its concentration in the feed
does not cause proportional modiÐcations in the surface reac-
tivity, because various e†ects are possible. (i) Since a side reac-
tion of NH ] N oxidation is present, an increase in
A. T. Guttmann, R. K. Grasselli and J. F. Brazdil, US Pat., 4 746
641 and 4 788 317, 1988.
L. C. Glaser, J. F. Brazdil, D. D. Suresh, D. A. Orndo† and R. K.
Grasselli, US Pat., 4 767 739 and 4 788 173, 1988.
M. A. Toft, J. F. Brazdil and L. C. Glaser, US Pat., 4 784 979,
1
988.
M. A. Toft, J. J. Brazdil and L. C. Glaser, US Pat., 4 879 264,
989.
1
3
2
ammonia would favour the amount of ammonia surface
species necessary for the synthesis of acrylonitrile, but also
decrease the availability of surface oxygens consumed by this
side reaction. (ii) Since the NH ] N reaction is probably
C. S. Lynch, L. C. Glaser, J. F. Brazdil and M. A. Toft, US Pat., 5
094 898, 1992.
1
R. Catani, G. Centi, F. TriÐro
Chem. Res., 1992, 31, 107.
11 A. Andersson, S. L. T. Andersson, G. Centi, R. K. Grasselli, M.
Sanati and F. TriÐro, in New Frontiers in Catalysis, ed. L. Guczi,
F. Solymos and P. Tetenyi, Elsevier, Amsterdam 1993, p. 691.
` and R. K. Grasselli, Ind. Eng.
3
2
catalyzed by surface V5` sites,29 an increase in its rate favours
the reduction of this species with beneÐcial aspects in terms of
global selectivity, but negative aspects in terms of selectivity to
acrylonitrile as discussed above.
`
12 G. Centi and S. Perathoner, in Preparation of Catalysts V I, ed. G.
Poncelet, J. Mertens, B. Delmon, P. A. Jacobs and P. Grange,
Elsevier Science, Amsterdam, 1995, p. 59.
Therefore, when in a catalyst such as SbV1a the key aspect
of the surface reactivity is the high rate of consecutive oxida-
tion to carbon oxides and low rate of propene to acrylonitrile
transformation as discussed before, an increase in the
ammonia and a decrease in the oxygen concentration in the
feed is beneÐcial not only in terms of global selectivity, but
also of selectivity to acrylonitrile.
1
3
G. Centi, E. Foresti and F. Guarneri, in New Developments in
Selective Oxidation II, ed. V. Cortes Corberan and S. Vic Bellon,
Elsevier, Amsterdam 1994, p. 281.
1
4
G. Centi and P. Mazzoli, Catal. T oday, 1996, 28, 351.
15 G. Centi, F. Guarneri and S. Perathoner, Appl. Catal. A, 1997, in
press.
1
1
6
7
T. Birchall and A. E. Sleight, Inorg. Chem., 1976, 15, 868.
F. J. Berry, M. E. Brett and W. R. Patterson, J. Chem. Soc.,
Dalton T rans., 1983, 9.
In a catalyst such as SbV10a, instead, the large excess of
antimony limits the presence of surface V5` species as well as
the rate of NH ] N oxidation. The availability of ammonia
1
8
S. Hansen, K. Sta
Chem., 1993, 102, 340.
19 A. Landa-Canovas, J. Nilsson, S. Hansen, K. Sta
- hl, R. Nilsson and A. Andersson, J. Solid State
3
2
surface species for the synthesis of acrylonitrile is thus not
critical as in the case of SbV1a, but rather the rate of reoxida-
tion of reduced surface sites is more critical. Assuming in
analogy with previous studies on the surface reactivity of Sb-
based catalysts for propene ammoxidation28 that the active
sites for acrylonitrile synthesis are Sb5` ions, probably in the
-
hl and A.
Andersson, J. Solid State Chem., 1995, 116, 369.
20
21
22
C. Rocchiccioli-Deltche†, T. Dupuis, R. Frank and M. Harmelin,
J. Chim. Phys. Phys.-chim. Biol., 1970, 67, 2037.
C. A. Cody, L. DiCarlo and R. K. Darlington, Inorg. Chem., 1979,
1
8, 1572.
B. Darriet, J. Bovin and T. Galy, J. Solid State Chem., 1976, 19,
Sb-rich “VSbO Ï phase, and that the slow step is the reoxida-
205.
4
tion of Sb3` to Sb5`, it is reasonable that the presence of
23 D. B. Rogers, R. D. Shannon, A. W. Sleight and J. L. Gillson,
Inorg. Chem., 1969, 8, 841.
small amounts of surface V5` sites on the surface would
favour the reoxidation of Sb3` by a two-electron redox reac-
tion analogous to that used during the preparation of these
catalysts. The competition of ammonia for the reduction of
surface V5` sites thus would decrease the rate of catalyst
reoxidation with a consequent detrimental e†ect on the selec-
tivity to acrylonitrile and a parallel increase in the formation
of propene.
2
4
R. G. Teller, M. R. Antonio, J. F. Brazdil and R. K. Grasselli, J.
Solid State Chem., 1986, 64, 249.
2
2
5
6
J. K. Burdett, Inorg. Chem., 1985, 24, 2244.
W. H. Baur and A. A. Khan, Acta Crystallogr., Sect. B, 1971, 27,
2133.
27 A. Andersson, S. L. T. Andersson, G. Centi, R. K. Grasselli, M.
Sanati and F. TriÐro
G. Centi and F. TriÐro
references therein.
G. Centi, S. Perathoner, F. TriÐro
`
, Appl. Catal. A, 1994, 113, 43.
, Catal. Rev.-Sci. Eng., 1986, 28, 165 and
2
8
`
For intermediate Sb : V ratios between 1 and 10, these two
opposite e†ects are minimized and it is thus reasonable that
both the selectivity and productivity to acrylonitrile show a
maximum for an intermediate Sb : V ratio between 1 and 10
2
3
9
0
` , Appl. Catal. A, 1997, in press.
G. Centi, F. Marchi and S. Perathoner, J. Chem. Soc., Faraday
T rans., 1996, 92, 5141; 5151.
[Fig. 4(b) and (d)], although the position of the maximum
Paper 7/02017A; Received 24th March, 1997
3402
J. Chem. Soc., Faraday T rans., 1997, V ol. 93