Comparison of hydrogen adsorption and aniline hydrogenation over co-precipitated Co/Al2O3 and Ni/Al2O3 catalysts

Article abstract of DOI:10.1039/a608074j

Co/Al₂O₃ catalysts with different cobalt content (10-50 wt.%) have been prepared by a co-precipitation method at pH 8. The catalysts were characterised by physical measurements (pore volume, bulk density, surface area and XRD) and by

Full text of DOI:10.1039/a608074j

View Article Online / Journal Homepage / Table of Contents for this issue
Comparison of hydrogen adsorption and aniline hydrogenation over
co-precipitated Co/Al O and Ni/Al O catalysts
2
3
2 3
Sankarasubbier Narayanan* and Ramachandran Pillai Unnikrishnan
Catalysis Section, Indian Institute of Chemical T echnology, Hyderabad - 500007, India
Co/Al O catalysts with di†erent cobalt content (10È50 wt.%) have been prepared by a co-precipitation method at pH 8. The
2
3
catalysts were characterised by physical measurements (pore volume, bulk density, surface area and XRD) and by H
and O2
2
chemisorption. The catalytic properties were evaluated using the aniline hydrogenation reaction. The conversion and product
selectivities are discussed with respect to metal content, reaction temperature and contact time. The adsorption and catalytic
properties of Co/Al O were compared with those of Ni/Al O catalysts containing the same metal content and prepared under
2
3
2 3
similar conditions. The conversion and selectivity di†erences are discussed.
Supported metal catalysts, especially precious-metal impreg-
Experimental
nated catalysts, have been widely used in the petroleum and
petrochemical industry for reforming, isomerisation and
cracking. Supported transition-metal catalysts are used in
steam reforming, hydrocracking, FischerÈTropsch synthesis
and hydrodesulfurisation. These catalysts also Ðnd application
in the manufacture of Ðne chemicals and in the synthesis of
organic chemical intermediates. The use of transition-metal
catalysts in organic transformation reactions is cheaper than
using precious-metal catalysts. It is possible to study the role
of the support in reducibility and metal dispersion1h3 and on
the conversion and selectivity owing to the di†erent oxidation
states in which these metals can exist. We have studied a
number of hydrogenation reactions such as benzene hydro-
genation,1,3h5 phenol hydrogenation,6 aniline hydro-
genation7h9 and acetone hydrogenation10 using transition-
metal supported catalysts. Very often, these catalysts
have a high metal content and are prepared by the co-
precipitation method whereas precious-metal supported cata-
lysts are usually prepared by an impregnation method since
this involves only a small amount of metal. The metal content,
the nature of the metallic species and the method of prep-
aration all contribute to the catalytic activity and selectivity.
We have discussed in detail the inÑuence of support, pretreat-
ment conditions and additive e†ect on metal support inter-
action, reducibility and dispersion of metal in supported
nickel systems prepared by an impregnation method.1h4
Recently, we have described the preparation, by a co-
precipitation method, characterisation and evaluation of sup-
ported nickel catalysts, with di†erent amounts of nickel, for
the aniline hydrogenation reaction.9 Here, we focus on a sup-
ported cobalt catalyst, an important constituent, like nickel, in
several catalyst systems used in FischerÈTropsch synthe-
sis,11h15 hydrocracking16h18 and selective hydrogenation reac-
tions.19h22 The hydrogen adsorption behaviour and catalytic
performance in the hydrogenation of aniline of Co/Al O
Catalyst preparation
A series of Co/Al O catalysts with metal loading varying
2 3
from 10 to 50 wt.% were prepared by a co-precipitation tech-
nique at a constant pH as described earlier for Ni/Al O cata-
lysts.9 A mixture of Co(NO ) É 6H O and Al(NO ) É 9H O
2 3
3 2
2
3 3
2
solutions (each 0.5 M) was added together with a solution of
sodium hydroxide (1 M) to a beaker containing 300È500 cm3
of distilled water and equipped with a stirrer and a pH elec-
trode. The precipitation was carried out at constant pH 8 by
controlled addition of nitrate and alkali solutions. The pink
coloured precipitate was aged in the mother liquor overnight
at room temperature. It was then washed, dried and sieved
and particles of ca. 1000È1400 lm were used. The catalyst was
calcined in air at 673 K for 12 h at a heating rate of 5 K
min~1 and then reduced in Ñowing hydrogen (120 cm3 min~1)
at 723 K for 24 h. The reduced sample was passivated in an
airÈnitrogen mixture at room temperature before being
stored.23
Characterisation
Apparent pore volume and bulk density of the calcined cata-
lysts (before reduction) were determined by conventional
methods. Water was used as an absorbate for the pore volume
determination. The BET area of the catalyst was measured by
a single-point method using N as an adsorbate at 77 K with
a Micromeritics Pulse Chemisorb Unit (Model 2700).
2
X-Ray di†raction patterns of the catalysts were taken with
a Philips PW 1140 di†ractometer having Ni-Ðltered Cu-Ka
radiation (j \ 1.54 Ó).
A conventional high-vacuum apparatus was used to carry
out H and O adsorption measurements. The procedures
2
2
have been described previously.1 H adsorption on Co/Al O
2
2 3
was measured at 373 K whereas that on Ni/Al O was mea-
2
3
2 3
sured at room temperature. O adsorption was measured at
2
catalysts with di†erent metal contents, prepared by a co-
precipitation method, are studied. Aniline hydrogenation is
important for production of cyclohexylamine (CHA), N-
phenylcyclohexylamine (NPCHA) and dicyclohexylamine
673 K on both Co and Ni/Al O and was used to calculate
2
3
the extent of reduction. This is based on the assumption that
unreduced nickel is present as NiO and the reduced metallic
nickel would be converted to NiO at this temperature in the
presence of oxygen.24 In the case of cobalt, it is presumed that
cobalt oxide on calcination is present as Co O which is
reduced to Co metal.25 The percentage reduction, dispersion
and crystallite size of the catalysts were determined from
(
DCHA) which are useful intermediates.8,9 The conversion
and selectivity over the cobalt catalysts are discussed and are
compared with those over Ni/Al O catalysts. The physico-
chemical properties of the catalysts are used to explain the
product selectivity di†erences. The mechanism of product for-
mation is explained.
2
3
3 4
adsorption measurements as described earlier.1
J. Chem. Soc., Faraday T rans., 1997, 93(10), 2009È2013
2009

View Article Online
Catalysis
All the catalysts were evaluated for vapour-phase hydro-
genation of aniline. The reaction was carried out in a vertical
tubular glass reactor (10 mm id) containing ca. 1 g catalyst
packed in between two layers of glass wool and glass beads.
Aniline was fed over the catalyst from the top using a motor-
ised syringe at a controlled rate. H (80 cm3 min~1) was used
2
as both reductant and carrier gas. The reaction temperature
was varied from 473 to 623 K. The experiments were carried
out at di†erent weight hourly space velocities (WHSVs).
[
0.011 to 0.099 mol h~1 (g cat)~1]. The liquid products were
collected in an ice trap and analysed with a gas chromato-
graph using SE-30 column (1/8 in ] 8 in) in a programmed
mode from 353 to 473 K at a heating rate of 10 K min~1. The
products identiÐed were cyclohexane (CH), cyclohexylamine
(
CHA),
dicyclohexylamine
(DCHA)
and
N-phenyl-
cyclohexylamine (NPCHA).
The activity is deÐned as
XFd
A \
mol h~1 (g cat)~1
W M
where, X \ fractional conversion, F \ feed rate (ml h~1),
d \ density of the reactant, M \ molecular weight of the reac-
tant and W \ weight of the catalyst.
selectivity is deÐned as
1
00 ] conversion to product i (%)
si \
total conversion (%)
Fig. 1 Variation in surface area (S ), pore volume (V ) and bulk
BET
p
Results and Discussion
density (d ) of Co/Al O Nand Ni/Al O = catalysts
B
2
3
2
3
Physical characterisation
reduction. At low metal concentration, CoAl O species for-
The Co/Al O catalysts, on precipitation and drying were
2 4
mation (bulk/surface) is likely, as in the case of Ni/Al O .1
2 3
2
3
pink in colour. On calcination and reduction, the colour
changed to blue in the case of 10 wt.% Co on Al O and to
2
3
Adsorption data
grey and black for Co P 20 wt.%. The bulk density, pore
volume and surface area of Co/Al O catalysts containing dif-
ferent weight percentages of Co are shown in Fig. 1. The bulk
density of cobalt increases with metal loading. The pore
volume of the catalysts is relatively small and decreases slight-
H and O adsorption data on Co/Al O catalysts are given
2
3
2
2
2 3
in Table 1. It has been reported that H adsorption on cobalt
2
is an activated process26 and hence H adsorption measure-
2
ments are usually carried out at 373 or 423 K. We carried out
ly with metal content. The surface area, measured by N
adsorption, also decreases with metal content. In Fig. 1, the
physical properties of Co/Al O are also compared with those
the H uptake measurements at 373 K. H uptake on
2
2
2
Co/Al O is relatively low compared to an Ni/Al O .9 This is
2
3
2 3
reÑected in the metal surface area of the catalysts. Even
though the reducibility is reasonably good (between 40 and
77%) in the case of cobalt catalysts containing 20 to 50 wt.%
metal, the dispersion of metal is very small. The high O2
uptake and high reducibility, with low dispersion and low
surface area, may indicate that cobalt is present as large
2
3
of Ni/Al O catalysts having the same metal content. Note
2
3
that the bulk density of cobalt catalysts is much higher and
the pore volume much lower than for the corresponding
nickel catalysts. The surface area of Co/Al O catalysts is also
2
3
lower than those of Ni/Al O catalysts. The fall in surface
2
3
area with metal content is less for Ni/Al O than for
crystallites and hence fewer cobalt atoms are exposed to H .
2
3
2
Co/Al O . This is due to the larger pore volume of Ni/Al O
catalysts, which facilitates the accommodation of a greater
amount of metal with minimum loss of surface area.
10 wt.% Co/Al O catalyst, however, shows negligible H
2
3
2 3
2 3
2
uptake. The catalyst is difficult to reduce, probably because of
the metalÈsupport interaction forming cobalt aluminate
surface and bulk species. The blue colour of the catalyst, even
after reduction, lends support to the presence of cobalt alu-
minate species. The O /H ratio also suggests that cobalt is
XRD patterns of cobalt catalysts show evidence of cobalt
oxide species present as large crystallites. There is no indica-
tion of any compound formation except perhaps in the case of
2 2
mostly present in the bulk rather than on the surface. One of
1
0 wt.% Co/Al O3
which is rather blue in colour, even after
2
Table 1 Adsorption data on co-precipitated Co/Al O catalysts
2
3
cobalt
loading
metal
surface area
/m2 (g cat)~1
crystallite
size
/nm
H
uptake
O
uptake
reduction
(%)
dispersion
(%)
2
2
(
wt.%)
/lmol (g cat)~1
/lmol (g cat)~1
O /H
2
2
1
2
3
4
5
0
0
0
0
0
È
15
42
52
67
È
866
2066
2679
3679
È
1
4
4
6
È
41
66
68
77
È
2
3
2
2
È
47
34
47
47
È
57
50
52
55
2010
J. Chem. Soc., Faraday T rans., 1997, V ol. 93

View Article Online
the reasons for the high O uptake is that cobalt can exist in
2
oxidation states higher than two.27
Catalysis
Aniline hydrogenation over cobalt catalysts was carried out at
di†erent temperatures and feed rates.
E†ect of metal content. Aniline conversion and product
selectivity at 473 K and WHSV 0.041 mol h~1 (g cat)~1 [3.75
cm3 h~1 (g cat)~1] over di†erent Co/Al O catalysts are
2
3
shown in Fig. 2. The activity increases with metal loading
beyond 20 wt.%. 10% Co/Al O shows no activity, in agree-
2
3
ment with its adsorption data. It is hardly likely that any CoO
species is present on this catalyst surface to e†ect any hydro-
genation activity. The activity of the catalysts is generally low;
even 50% Co/Al O shows a conversion of only 23%. This is
Fig. 3 Aniline hydrogenation over 50 wt.% Co/Al O . E†ect of tem-
2
3
perature of reaction on conversion (>) and selectivity to (@) CHA,
(
…) NPCHA and (=) CH. WHSV \ 0.041 mol h~1 (g cat)~1.
2
3
because of the low metal area of cobalt catalysts as compared
with Ni/Al O catalysts with the same metal content.9 CHA
2
3
and NPCHA are the two main products formed on cobalt
catalysts. The selectivity for CHA is very high (ca. 95%) on
2
0 wt.% cobalt catalyst at a very low conversion of 5%. At
formation. The selectivity is similar to that observed on
higher metal content ([30 wt.%), the cobalt catalysts show a
conversion of ca. 20% and selectivity for CHA and NPCHA
of 60 and 40%, respectively. The 50 wt.% Co/Al O catalyst
was selected for studies involving temperature and feed rate
e†ects.
Ni/Al O catalysts9 and the mechanism involved may be the
2 3
same for both Co and Ni/Al O systems. The only di†erence
2 3
2
3
between the two catalysts is with respect to aniline conversion.
Increasing temperature increases aniline conversion over
Co/Al O but over Ni/Al O , increase in temperature from
2 3 2 3
73 to 673 K reduced the conversion from 85% to a steady
4
InÑuence of temperature. The e†ect of temperature on con-
version and selectivity for aniline hydrogenation is shown in
Fig. 3. The conversion increases from 20 to 35% over the tem-
perature range 473È623 K. This is in contrast to the activity
observed on Ni/Al O .9 Furthermore, hydrogen adsorption
value of ca. 20È25%.9
InÑuence of contact time. The inÑuence of contact time on
aniline conversion and product selectivity is shown in Fig. 4.
Increase in contact time increases conversion. At 1/WHSV of
2
3
on Co/Al O is an activated process26 and increases with
2
3
temperature [e.g. H adsorption on 50% Co/Al O at 298 K
2 3
was 53 lmol (g cat)~1 whereas at 373 K, it was 67 lmol
20.1 mol~1 h (g cat), the conversion is ca. 10%. On increasing
2
the contact time of aniline to 91 mol~1 h (g cat), the conver-
sion increased to ca. 70%. CHA and NPCHA are the two
major products. CHA formation is slightly favoured at low
contact time and NPCHA is formed in roughly similar
amounts at all contact times. There is hardly any selectivity
di†erence between CHA and NPCHA with respect to feed
rate. However, at high contact time, DCHA and CH are also
formed, contributing to high conversion. The mechanism of
aniline hydrogenation may be envisaged as follows,
(
g cat)~1]. In the case of Ni/Al O , hydrogen adsorption
2
3
decreases with increase in temperature [e.g. on 50%
Ni/Al O , H adsorption was 286 lmol (g cat)~1 at 298 K
2
3
2
and 240 lmol (g cat)~1 at 373 K]. This suggests that, on
Co/Al O , both hydrogen adsorption and aniline hydro-
2
3
genation are activated processes. The apparent activation
energy for the aniline hydrogenation reaction over Co/Al O
is calculated to be 36 kJ mol~1, compared with 29 kJ mol~1
over Ni/Al O .
The temperature e†ect on product selectivity assumes sig-
niÐcance only in the temperature region 473È523 K. The
selectivity for CH increases and reaches almost 100% whereas
that for CHA decreases, from 73% at 473 K to almost zero at
2
3
2
3
573 K. The NPCHA goes through a maximum at 523 K.
Above 523 K hydrodeamination is favoured, leading to CH
Scheme 1
CHA is probably the primary product of aniline hydro-
genation over both Co and Ni/Al O catalysts. The high
2
3
activity of Ni/Al O catalyst enables the CHA formed to
2
3
undergo further reaction on the catalyst surface, producing
secondary products such as DCHA and NPCHA. Secondary
reactions are not pronounced over the less active Co/Al O
catalysts, thereby producing CHA selectively. The e†ect of
contact time on product selectivity also points towards this.
At high contact time [90 mol~1 h (g cat)], the conversion or
2
3
Fig. 2 E†ect of metal loading on aniline conversion and product
selectivity over Co/Al O catalysts at 473 K and WHSV of 0.041 mol
2
3
h~1 (g cat)~1. (>) Conversion, (…) CHA, (=) NPCHA.
J. Chem. Soc., Faraday T rans., 1997, V ol. 93
2011

View Article Online
tivity for NPCHA. The selectivity pattern for Co follows the
order CHA [ NPCHA [ DCHA A CH; while that for Ni is
NPCHA [ DCHA \ CHA A CH. Since CHA is the primary
product of aniline hydrogenation,9 the selectivity di†erence
over Co and Ni/Al O can be tentatively explained as being
2
3
due to its high activity and the slow rate of desorption of
CHA from the nickel surface, giving it ample time to undergo
further reaction with aniline or with itself, producing NPCHA
and DCHA, respectively. Over the less active Co/Al O , the
2
3
CHA formed is desorbed quickly from the catalyst surface
hence giving high selectivity.
From the adsorption measurements, it may be inferred that
Co catalysts have a low hydrogen adsorption capacity com-
pared with the corresponding Ni catalysts prepared by the
same method. This is reÑected in the behaviour of the cata-
lysts for benzene hydrogenation. This reaction is carried out
to verify the adsorption data as the activity of benzene hydro-
genation is directly related to the metal surface area of the
catalyst, since the reaction gives only a single product viz. CH.
The low hydrogen adsorption and hydrogenation properties
of Co catalysts may also be ascribed to the formation of
several sub-oxides during the reduction process,27 not all of
which are reduced to metallic species. Furthermore, Co cata-
lysts seem to form large crystallites more readily than Ni cata-
lysts, burying the metal in the bulk. The relatively high density
and low surface area of the catalysts also contribute to the low
activity of Co, conÐrming the above observation of physical
properties.
Aniline hydrogenation was chosen as a reaction to study
the di†erences in the hydrogenating properties of Ni and Co
catalysts since it gives more than one reaction product and
hence the selectivity patterns can be compared. The major
products of the reaction are all important chemical interme-
diates and is possible to selectively synthesise a particular
product by choosing appropriate catalysts and experimental
conditions.
Fig. 4 Aniline hydrogenation over 50 wt.% Co/Al O catalyst at
2
3
4
73 K. E†ect of contact time of aniline on conversion 1 and selec-
tivity to 4 CHA, C NPCHA, 5 DCHA, D CH.
the activity is high and hence the formation of secondary pro-
ducts increases.
Comparison with Ni/Al O . Nickel and cobalt catalysts
2
3
containing 50 wt.% metal are compared for aniline hydro-
genation at 473 K and a WHSV of 0.041 mol h~1 (g cat)~1 in
Fig. 5. Ni/Al O shows nearly 85% conversion compared with
2 3
0% conversion over Co/Al O . The Co/Al O catalyst gives
2
2
3
2 3
only CHA and NPCHA; the selectivity towards CHA is
nearly twice that of NPCHA. Over Ni/Al O , CHA, NPCHA
2
3
and DCHA are formed in roughly equal amounts with a small
amount of CH.
To eliminate the inÑuence of conversion on the selectivity
patterns, both Ni and Co catalysts containing 50 wt.% metal
are compared at a constant conversion of 57% in Fig. 6. The
conversion value is chosen at 473 K, irrespective of the feed
rate or contact time of the feed. The Co catalyst shows high
selectivity for CHA whereas the Ni catalyst shows high selec-
From this study, it is clear that Co/Al O is not as active as
2
3
Ni/Al O for the hydrogenation of aniline. Whilst oxidised
2
3
nickel exists mostly in the ]2 state, cobalt can exist in di†er-
ent oxidation states. The metal area and percentage reduction
of Co/Al O are lower than those of Ni/Al O catalysts
2
3
2 3
having the same metal loading. The dispersion of both cata-
lysts is, however, very low (\10%). The activity of Co/Al O
2
3
catalysts for aniline hydrogenation increases with temperature
whereas that of Ni/Al O decreases with temperature. This is
2
3
attributed to the activated nature of the adsorption of the
reactant molecules over Co/Al O . With respect to selectivity,
2
3
Co favours CHA over NPCHA and DCHA whereas Ni
favours NPCHA. At low conversion levels also (20È25%,
T \ 473 K), Co shows high selectivity (65%) for CHA
whereas Ni shows a high selectivity for NPCHA (90%).
However, at high temperature, both catalysts give CH by
hydrodeamination.
Fig. 5 Comparison of aniline conversion and product selectivity of
Co/Al O 3 and Ni/Al O D containing 50 wt.% metal. T \ 473 K,
2
3
2 3
WHSV \ 0.041 mol h~1 (g cat)~1.
Conclusion
Co/Al O and Ni/Al O prepared by a co-precipitation
2
3
2 3
method are well characterised by hydrogen and oxygen
chemisorption and evaluated by aniline hydrogenation yield-
ing valuable products viz. CHA, DCHA and NPCHA. The
co-precipitated Ni and Co/Al O catalysts can be e†ectively
2
3
used for synthesising CHA or NPCHA selectively. Ni/Al O ,
although more active than Co/Al O with the same metal
content, leads to secondary products DCHA and NPCHA, in
preference to CHA. Co/Al O catalysts can therefore be judi-
ciously used to synthesise CHA selectively.
2
3
2
3
2
3
The authors thank the Council of ScientiÐc and Industrial
Research (CSIR), New Delhi, for the award of a Senior
Research Fellowship to R.U.K.
Fig. 6 Comparison of aniline hydrogenation product selectivity of
5
0 wt.% Co/Al O 3 and Ni/Al O D catalysts at a constant con-
2 3 2 3
version of 57% at 473 K
2012 J. Chem. Soc., Faraday T rans., 1997, V ol. 93

View Article Online
References
15 B. Sarup and B. W. Wojciechowski, Can. J. Chem. Eng., 1989, 67,
6
2.
1
2
3
4
5
6
7
8
9
S. Narayanan and K. Uma, J. Chem. Soc., Faraday T rans. 1,
16
17
18
19
20
C. Wivel, R. Candia, B. S. Clausen, S. Mrup and H. Topsoe, J.
1
985, 81, 2733.
Catal., 1981, 68, 453.
S. Narayanan, K. Uma and V. Venkat Rao, Proc. II. Indo-Soviet.
P. A. Arnoldy, M. C. Franken, N. Sche†er and J. A. Moulijn, J.
Catal., 1985, 96, 381.
Seminar. Catal., 1986, LVII, 517.
S. Narayanan and G. Sreekanth, J. Chem. Soc., Faraday T rans. 1,
D. G. Castler, P. R. Watson and I. Y. Chan, J. Phys. Chem., 1989,
1
989, 85, 3785.
93, 3188.
S. Narayanan and K. Uma, J. Chem. Soc., Faraday T rans. 1,
K. M. Winterbottom, in Catalysis and Chemical Processes, ed. R.
Pearce and W. R. Patterson, Wiley, New York, 1981, p. 304.
P. N. Rylander, in Catalysis: Science and T echnology, ed. J. R.
Anderson and M. Boudart, Springer, New York, 1983, vol. 4,
p. 1.
1
987, 83, 733.
V. Vishwanathan, M. S. Rajashekar, G. Sreekanth and S. Naray-
anan, J. Chem. Soc., Faraday T rans., 1991, 87, 3449.
S. Narayanan and G. Sreekanth, Indian. J. T echnol., 1993, 31,
5
07.
21
22
23
B. Sen and M. A. Vannice, J. Catal., 1988, 113, 52.
V. Vishwanathan and S. Narayanan, J. Chem. Soc., Chem.
Commun., 1990, 78.
V. Vishwanathan, S. M. Sajjad and S. Narayanan, Indian. J.
Chem. A, 1991, 30, 679.
S. Narayanan, R. Unnikrishnan and V. Vishwanathan, Appl.
Catal., 1995, 129, 9.
S. Narayanan and R. Unnikrishnan, Appl. Catal., 1996, 145, 231.
R. B. Anderson, T he FischerÈT ropsch Synthesis, Academic Press,
New York, 1984, p. 123.
L. M. Gandia, A. Diaz and M. Montes, J. Catal., 1995, 157, 461.
A. Gil, A. Diaz and M. Montes, J. Chem. Soc., Faraday T rans.,
1
991, 87, 791.
2
2
2
4
5
6
C. H. Bartholomew and R. J. Farrauto, J. Catal., 1976, 45, 41.
R. C. Reuel and C. H. Bartholomew, J. Catal., 1984, 85, 63.
J. M. Zowtiak, G. D. Weatherbee and C. H. Bartholomew, J.
Catal., 1983, 82, 230.
1
1
0
1
2
7
P. Arnoldy and J. A. Moulijn, J. Catal., 1985, 93, 38.
1
2
A. A. Adesina, R. R. Hudgins and P. L. Silvenston, Can. J. Chem.
Eng., 1986, 64, 447.
1
1
3
4
D. K. Lee, J. H. Lee and S. K. Ihm, Appl. Catal., 1988, 36, 199.
W. H. Lee and C. H. Bartholomew, J. Catal., 1989, 120, 256.
Paper 6/08074J; Received 28th November, 1996
J. Chem. Soc., Faraday T rans., 1997, V ol. 93
2013