Ammoxidation of 3-Picoline to Nicotinonitrile over Highly Dispersed V2O5/ZrO2 Catalysts

Article abstract of DOI:10.1039/a705267g

Vapour phase ammoxidation of 3-picoline to nicotinonitrile was carried out on highly dispersed vanadia catalysts supported on zirconia, the results suggest that vanadia/zirconia possesses high activity and selectivity for nicotinonitrile formation.

Full text of DOI:10.1039/a705267g

View Article Online / Journal Homepage / Table of Contents for this issue
3
14 J. CHEM. RESEARCH (S), 1998
J. Chem. Research (S),
Ammoxidation of 3-Picoline to Nicotinonitrile over
Highly Dispersed V O /ZrO Catalysts$
K. V. R. Chary,* G. Kishan, K. V. Narayana and T. Bhaskar
1998, 314±315$
2
5
2
Catalysis Section, Indian Institute of Chemical Technology, Hyderabad 500 007, India
Vapour phase ammoxidation of 3-picoline to nicotinonitrile was carried out on highly dispersed vanadia catalysts
supported on zirconia, the results suggest that vanadia/zirconia possesses high activity and selectivity for nicotinonitrile
formation.
Nicotinamide, a component of vitamin B, is an important
compound for metabolism in human beings and animals
and is used as a food additive. It is usually synthesized
by the ammoxidation of 3-picoline to nicotinonitrile and
1
±5
further hydrolysis of the nitrile formed.
catalysts either unsupported or supported on anatase TiO
Vanadium oxide
2
,
2
±4
are generally used for this ammoxidation.
zirconia has attracted considerable interest because of its
potential use as a catalyst support. It has been employed
in many industrially important reactions such as hydro-
In recent years
6
±16
8
±10
11,12
processing,
methanol and higher alcohols,
oxidation of alcohols, synthesis of
and methanation reac-
13±15
16
tions. Here we report the synthesis of nicotinonitrile by
the vapour phase ammoxidation reaction over highly dis-
persed vanadia catalysts supported on zirconia. We also
report an improved oxygen chemisorption method for deter-
mination of the dispersion of vanadia.
Zirconia support was prepared by ammonical hydrolysis
of zirconium oxychloride (Fluka) at pH 9, followed by
calcination in air at 773 K for 6 h. A series of V O /ZrO
2
2
5
catalysts with vanadia loadings ranging from 1.8 to 10.4
wt.% of V have been prepared by the wet impregnation
of ZrO (84 m
solution containing NH
2
O
5
2
�1
Fig. 1 Ammoxidation of 3-picoline to nicotinonitrile over
various vanadia±zirconia catalysts
2
g
) with requisite amounts of aqueous
VO . The impregnated catalysts
4
3
were dried at 383 K for 16 h followed by calcination
at 773 K for 4 h in air. The vanadium content in the
niconinonitrile at 683 K. The conversion of 3-picoline as
well as the selectivity for nicotinonitrile increased with
increasing vanadia loading in the catalyst. Beyond 5.34%
®nished catalysts was determined by atomic absorption
spectrometry. The dispersion of vanadia was determined by
oxygen chemisorption. Prior to adsorption measurements
the samples were pre-reduced in a ¯ow of hydrogen (40 ml
2 5 2
V O on ZrO the selectivity for nicotinonitrile was found
to be independent of vanadia content in the catalyst. Pure
zirconia was inactive for ammoxidation under the exper-
imental conditions. The active species for ammoxidation
of 3-picoline could be the formation of a VO2 phase, by
reduction of the catalysts with hydrogen. The ESR
spectra of hydrogen reduced catalysts further support the
�
1
min ) at 640 K for 2 h and evacuated at the same tempera-
ture for 1 h. Oxygen chemisorption uptakes were determined
as the di€erence between two successive adsorption iso-
therms measured at 640 K. A down ¯ow ®xed-bed reactor
operating at atmospheric pressure and made of Pyrex glass
was used for catalyst testing during the ammoxidation of
4‡
17
presence of V on zirconia. The spectra at low vanadia
loadings are well resolved with hyper®ne splitting (hfs)
3-picoline to nicotinonitrile. About 2 g of catalyst, diluted
51
due to V (I ˆ 7/2). The signal intensity decreases with
with an equal amount of quartz grains, were charged into
the reactor and supported on a glass-wool bed. Prior to
introducing 3-picoline with a syringe pump, the catalyst
was pre-reduced at 673 K for 2 h in puri®ed hydrogen
increasing vanadia content in the catalyst. Spin Hamiltonian
_ _
k k
parameters g =1.933, g =1.991, A =187 G, A =70 G in-
dicate the presence of V on zirconia. Andersson and co-
4‡
5,18,19
workers
vanadium oxide catalysts for ammoxidation is a non-sto-
chiometric VO or V 13 phase with an excess of oxygen
proposed that the active and selective phase in
�
1
¯
ow (40 ml min ). After reduction, the reactor was fed
with 3-picoline, ammonia, and air keeping the mole ratio
of 3-picoline±H O±NH ±air at 1:13:22:44 and contact time
2
6
O
2
3
ions and this phase is formed during the pre-reduction of
the catalysts by hydrogen.
In order to optimize the reaction conditions such as
0
.6 s. The reaction was carried out at 683 K after optimizing
the reaction temperature in the range 598±753 K. The liquid
products, mainly nicotinonitrile, were analysed by gas
chromatography using an OV-17 column Traces of carbon
oxides were also formed during the reaction.
Fig. 1 shows the dependence of the activity/selectivity
on vanadia loading during ammoxidation of 3-picoline to
2 5 2
temperature the reactivity of 5.34% V O ±ZrO was studied
in the range 598±753 K. The e€ect of reaction temperature
on the catalytic behaviour during ammoxidation is shown in
Fig. 2. The conversion of 3-picoline increases with increas-
ing temperature up to 683 K and levels o€ at higher tem-
peratures. However, the selectivity ®rst remains constant up
to 683 K and then monotonically decreases with increasing
temperature. Thus a reaction temperature of 683 K was
*To receive any correspondence.
$This is a Short Paper as de®ned in the Instructions for Authors,
selected to screen all other compositions of V O ±ZrO
2
Section 5.0 [see J. Chem. Research (S), 1998, Issue 1]; there is there-
fore no corresponding material in J. Chem. Research (M).
2
5
catalysts for ammoxidation.

View Article Online
J. CHEM. RESEARCH (S), 1998 315
Table 2 Results of ammoxidation of 3-picoline to
nicotinonitrile over supported vanadium oxide catalysts
Catalyst
% Conversion
% Selectivity
5
6
7
.34% V
2
O
5
±ZrO
±TiO
±Al
2
64
97
55
81
92
86
% V
% V
2
O
5
2
2
O
5
2 3
O
4
73 K led to the formation of V . Thus the formation of
‡
7
3‡
V
depends on the nature of the support employed and
also the pre-reduction temperature. In the present investi-
gation oxygen chemisorption was carried out on the samples
pre-reduced at low reduction temperatures (640 K) to ensure
only surface reduction of vanadia supported on zirconia.
The reactivity of the present V O ±ZrO catalysts is com-
2
5
2
pared with those of V
catalysts during ammoxidation of 3-picoline in Table 2.
The results show that the V ±ZrO catalysts are more
active than V O ±Al O and less active than V O ±TiO
2
2 5 2 2 5 2 3
O ±TiO (anatase) and V O ±Al O
2
O
5
2
Fig. 2 Dependence of reaction temperature on the conversion
or selectivity for 5.34% V ±ZrO catalyst
2
O
5
2
2
5
2
3
2
5
18
(
anatase). The high activity of vanadia catalysts supported
on zirconia is attributed to the higher acidity of the latter
and also its stronger interaction with vanadium oxide.
However, we did not ®nd a direct correlation between the
An improved method of oxygen chemisorption suggested
20
by Oyama et al. has been employed for determination of
the dispersion of vanadia supported on zirconia. The results
of oxygen uptake measured at 640 K and other information
such as oxygen atom site density, dispersion, etc., derived
therefrom are given in Table 1. At low vanadia loadings the
dispersion of vanadia (O/V) is very high and nearly equal
to the stoichiometry of one oxygen atom per vanadium
atom. X-Ray di€raction shows the presence of a crystalline
vanadia phase only at high vanadia loadings in addition to
the lines due to the monoclinic and tetragonal phases of
2 5 2
oxygen chemisorption capacity of the V O ±ZrO catalysts
and activity. Thus zirconia-supported vanadia catalysts were
highly active and selective for the conversion of 3-picoline
into nicotinonitrile, and oxygen chemisorption at 640 K
(Treduction=Tadsorption=640 K) is found to be a facile
method to measure the dispersion of vanadia supported on
zirconia.
17
G.K. and K.V.N. thank CSIR, New Delhi and T.B.
thanks UGC, New Delhi for the award of a Senior
Research Fellowship.
zirconia. The decrease in the dispersion with increasing
vanadia loading might be due to the presence of a crystal-
line vanadia phase at higher loadings. For the same compo-
sition of V
2 5
O , the dispersion of vanadia supported on
Received, 22nd July 1997; Accepted, 9th February 1998
Paper E/7/05267G
zirconia was found to be much higher than that of vanadia
supported on g-Al O . For example, the dispersion of
21
2
3
vanadia for a catalyst having 3.38 wt.% V O supported on
2
5
References
2 5
zirconia was 76% (Table 1). However, for the same V O
21
1 Br. Pat., 1 317 064, 1973.
2 3
content supported on g-Al O the dispersion was 63%.
2
3
4
Ger. Pat., 2 435 134, 1975; 2 438 263, 1975; 2 427 204, 1974.
A. Andersson and S. T. Lundin, J. Catal., 1979, 58, 383.
A. Andersson and S. T. Lundin, J. Catal., 1980, 65, 9.
The oxygen atom site density values reported in Table 1
were less than the average number density of V1O groups
1
8
�2 20
on the low index planes of V
2
O
5
(5.0Â 10
m ) and this
5 A. Andersson, J.-O. Bovin and P. Walter, J. Catal., 1986, 98, 204.
6 T. Yamaguchi, Catal. Today, 1994, 20, 199.
suggests that the oxygen chemisorption method described
here samples the surface but not the bulk.
The reduction behaviour of supported vanadia systems
7
P. D. L. Mercera, Ph.D thesis, Twente Institute of Technology,
991.
S. W. Oliver, T. D. Smith, J. R. Pilbrow, K. C. Prat and
1
8
2
0±22
20
has been of interest in recent years.
Oyama et al.
V. Christov, J. Catal., 1988, 111, 1988.
9 B. M. Reddy, K. V. R. Chary, B. Rama Rao, V. S.
Subrahmanyam, C. S. Sunandana and N. K. Nag, Polyhedron,
1986, 5, 191.
showed that, to achieve complete reduction of the vanadia
surface without reducing the bulk, the reduction must be
22
carried out at 640 K. Nag and Massoth found using a
gravimetric method that reduction with hydrogen at 773 K
1
1
1
0 M. Vrinat, D. Hamom, M. Breysse, B. Durand and T. des
Courieres, Catal. Today, 1994, 20, 273.
1 K. V. R. Chary, B. Rama Rao and V. S. Subrahmanyam, Appl.
Catal., 1991, 74, 1.
2 J. G. van Ommen, P. J. Gellings and J. R. H. Ross, in Methane
Conversion, ed. D. M. Bibby, C. D. Chang, R. F. Howe and
S. Yurchak, Elsevier, Amsterdam, 1988, p. 213.
3 Y. Amenomiya, Appl. Catal., 1987, 30, 57.
for 6 h resulted in 70% reduction of silica-supported V
2 5
O
2
O
3
2
O
5
2 3
to V . However, reduction of V ±Al O samples at
Table 1 Oxygen uptake, dispersion, oxygen atom site density
and surface area of various V
O
2 5
±ZrO
2
catalysts
1
Catalyst
composition/
Oxygen
atom site
Uptake / density/ Dispersion
14 T. Iizuka, M. Kojima and K. Tanabe, J. Chem. Soc., Chem.
Commun., 1983, 638.
15 M. Ichikawa, Bull. Chem. Soc. Jpn., 1978, 51, 2268.
16 R. A. Dalla Betta, A. G. Piken and M. Shelef, J. Catal., 1975,
a
wt.% V
on ZrO
2
O
5
Surface
area/m g
O
2
2
�1
�1
18 �2
b
2
mmol g
10
m
(O/V)
4
0, 173.
1
3
5
6
8
.83
.38
.34
.53
.59
84
84
79
78
77
73
96.4
141.1
210.5
256.6
316.0
356.2
1.38
2.02
3.21
3.96
4.94
5.88
0.96
0.76
0.72
0.71
0.67
0.62
1
1
7 K. V. R. Chary, G. Kishan and T. Bhaskar, unpublished results.
8 A. Andersson and J.-O. Bovin, Naturwissenschaften, 1985, 72,
S.209.
9 A. Andersson, J. Catal., 1982, 76, 144.
0 S. T. Oyama, G. T. Went, K. B. Lewis, A. T. Bell and G. A.
Somorjai, J. Phys. Chem., 1989, 93, 6786.
1
2
1
0.38
a
b
T
reduction ˆ Tadsorption ˆ 640 K. Dispersion ˆ fraction of
21 K. V. R. Chary and G. Kishan, J. Phys. Chem., 1995, 99, 14 424.
22 N. K. Nag and F. E. Massoth, J. Catal., 1990, 124, 127.
vanadium atoms at the surface assuming Oads/Vsurf ˆ 1.