Acetone hydrogenation over co-precipitated Ni/Al2O3, Co/Al2O3 and Fe/Al2O3 catalysts

Article abstract of DOI:10.1039/a708124c

Co-precipitated Ni/Al₂O₃, Co/Al₂O₃ and Fe/Al₂O₃ catalysts containing 10-50 wt.% of metal are prepared and characterized by BET surface area, pore volume and XRD. Hydrogen chemisorption and oxygen uptake measurements are used to estimate the extent of reduction and the available surface metal sites. Acetone hydrogenation under vapour phase conditions is carried out over the catalysts in the temperature region 373-523 K. Conversion and selectivity of the catalysts are compared and are correlated with their hydrogen and oxygen uptake characteristics. Nickel catalysts are very active followed by cobalt and iron. A dual function mechanism involving both metal and acidic properties decide the selectivity of the products. The selectivity for propan-2-ol depends only on the metal availability whereas the selectivity for isobutyl methyl ketone depends on both acidic and metallic character of the catalyst. A mechanism explaining the different products formed is postulated.

Full text of DOI:10.1039/a708124c

View Article Online / Journal Homepage / Table of Contents for this issue
Acetone hydrogenation over co-precipitated Ni/Al O , Co/Al O and
2 3
2
3
Fe/Al O catalysts
2
3
Sankarasubbier Narayanan*¤ and Ramachandranpillai Unnikrishnan
Catalysis Section, Indian Institute of Chemical T echnology, Hyderabad-500 007, India
Co-precipitated Ni/Al O , Co/Al O and Fe/Al O catalysts containing 10È50 wt.% of metal are prepared and characterized by
2
3
2 3
2 3
BET surface area, pore volume and XRD. Hydrogen chemisorption and oxygen uptake measurements are used to estimate the
extent of reduction and the available surface metal sites. Acetone hydrogenation under vapour phase conditions is carried out
over the catalysts in the temperature region 373È523 K. Conversion and selectivity of the catalysts are compared and are
correlated with their hydrogen and oxygen uptake characteristics. Nickel catalysts are very active followed by cobalt and iron. A
dual function mechanism involving both metal and acidic properties decide the selectivity of the products. The selectivity for
propan-2-ol depends only on the metal availability whereas the selectivity for isobutyl methyl ketone depends on both acidic and
metallic character of the catalyst. A mechanism explaining the di†erent products formed is postulated.
Catalytic vapour phase hydrogenation of acetone has gained
attention recently. In the temperature region 423È523 K, this
reaction yields products that can be used in chemical heat
pumps for waste heat recovery and up-grading.1 A good con-
version and a high selectivity to propan-2-ol (2-P) are required
for this application. Acetone hydrogenation at high tem-
perature also gives isobutyl methyl ketone (IBMK), a valuable
chemical. Vapour phase reaction involving a direct conversion
of acetone in a single process step is an alternative to the pre-
sently available commercial process2,3 requiring multi-
reaction set-ups where acetone is Ðrst converted to diacetone
alcohol (DAA; 4-hydroxy-4-methylpentan-2-one) by aldol
condensation followed by dehydration of DAA to mesityl
oxide (MSO). MSO is then selectively hydrogenated to
IBMK. IBMK is mainly used as a solvent for cellulose and
resin-based coating systems and also for vinyl, epoxy and
acrylic resins.
Ni/Al O catalysts.16 Keeping in mind the importance of
2
3
acetone hydrogenation and, as a part of the on-going studies
on hydrogenation reactions over supported metal catalysts,
we are reporting here a detailed investigation of this reaction
over co-precipitated Ni/Al O , Co/Al O and Fe/Al O cata-
2
3
2 3
2 3
lysts. These catalysts are commonly used in methanation and
FischerÈTropsch synthesis. The present study is interesting as
it involves the hydrogenation of a simple symmetric molecule,
viz., acetone. This study also compares the selectivity patterns
and hydrogenation functions of Fe, Co and Ni.
Experimental
Catalyst preparation
Alumina-supported nickel, cobalt and iron catalysts contain-
ing 10È50 wt.% of metal were prepared by co-precipitation, at
constant pH conditions (pH \ 8 for Ni and Co and pH \ 7
for Fe) of 0.5 M nitrate salts with 1 M NaOH.13 After precipi-
tation, the hydroxide was thoroughly washed with warm dis-
tilled water until free of nitrates. The cake was dried at 383 K
overnight in an oven, crushed and sieved to 1000È1400 lm
size. The dried sample was calcined in air at 673 K for 15 h at
a heating rate of 5 K min~1 and reduced in Ñowing hydrogen
at 723 K for 24 h. The reduced sample was cooled to room
temperature in Ñowing hydrogen and was passivated in an
Acetone hydrogenation has been studied over many sup-
ported transition metal catalysts.4h10 Nearly 100% selectivity
to 2-P has been reported over silica supported Pd and Au in
the temperature range 423È493 K,7 on Pt/Al O , Pt/SiO
2
3
2
and Pt/TiO in the region 303È363 K6,8 and on Kieselghur-
2
supported Cu, Pt, Pd and Rh at 423 K.9,10 The reported
acetone conversion is only about 10%; Gandia and co-
workers1,2,4,5 have studied this reaction over various cata-
lysts. Acetone conversion of about 35% and 2-P selectivity of
nearly 99% were reported at 473 K over silica supported Ni
catalysts.4 IBMK selectivity of 60È80% was obtained over
Ni/MgO and Ni/Al O at 473 K; acetone conversion being
airÈN mixture for 1 h.
2
Unsupported Ni, Co and Fe oxides were prepared by cal-
cination in air of the respective metal nitrates at 673 K for 15
2
3
h. The oxides were reduced in H at 723 K for 24 h and were
only 10%.2,5 The formation of some low molecular weight
2
passivated at room temperature. The alumina support was
hydrocarbons (C ÈC ) have also been observed over unsup-
1
4
prepared at pH 8 by precipitation of an aqueous aluminium
nitrate solution (0.5 M) with sodium hydroxide (1 M). The
hydroxide obtained was calcined in air at 673 K for 15 h.
ported Pt, Ni, Fe and W.6,11,12 But this is not signiÐcant
(
\2%) over supported metal catalysts.6
We have reported the preparation and characterization of
co-precipitated nickel and cobalt/alumina catalysts and their
application for hydrogenation of benzene, aniline and
acetone.13h16 Benzene hydrogenation was used as a model
reaction to monitor the metal availability on supported metal
systems.17h21 Aniline hydrogenation over co-precipitated
Ni/Al O and Co/Al O catalysts yielding cyclohexylamine,
dicyclohexylamine and N-phenylcyclohexylamine have been
discussed in detail elsewhere.13h15 In a preliminary study, we
reported the selective conversion of acetone to IBMK on
Catalyst characterization
The apparent bulk densities and pore volumes of the calcined
catalysts were determined by conventional absorption of
water at room temperature.13 The BET surface area of the
catalysts were measured with a Micromeritics Pulse Chemi-
sorb Unit (Model 2700) using N
2
3
2 3
2 as adsorbate at 77 K by
single-point method.13
XRD patterns of the catalysts were obtained using a Philips
PW 1140 X-ray di†ractometer with a Ni-Ðltered Cu-Ka radi-
ation (j \ 1.540 Ó).
¤
Fax: ]91-40-7173757/7173387. E-mail: root=csiict.ren.nic.in
J. Chem. Soc., Faraday T rans., 1998, 94(8), 1123È1128
1123

View Article Online
Table 1 Physical properties of co-precipitated Ni/Al O , Co/Al O and Fe/Al O catalysts (calcined at 723 K for 15 h)
2
3
2
3
2
3
BET surface area/
bulk density/
g(cat) cm~3
pore volume/
cm3 g(cat)~1
metal
loading
m2 g(cat)~1
(
wt.%)
Fe
Co
Ni
Fe
Co
Ni
Fe
Co
Ni
1
2
3
4
5
0
0
0
0
0
247
233
187
132
114
274
239
195
183
135
290
253
250
247
221
0.717
0.955
0.970
0.995
1.01
0.571
0.599
0.845
0.860
0.950
0.205
0.225
0.283
0.288
0.597
0.888
0.633
0.628
0.620
0.609
1.25
1.16
0.864
0.806
0.622
4.36
3.71
2.65
2.00
1.10
H and O uptake measurements were carried out using a
surface areas and pore volumes. The XRD patterns of
Ni/Al O , Co/Al O and Fe/Al O catalysts are very broad
2
2
conventional constant-volume high-vacuum apparatus at
static conditions.13 The passivated catalyst sample was
reduced in hydrogen at 723 K for 2 h and was degassed at 673
K for 2È4 h to attain a pressure of 1 ] 10~6 cmHg prior to
hydrogen and oxygen uptake measurements. Hydrogen
uptake measurements on the supported metal catalysts were
carried out at 303 and 373 K.13,15,22h24 Oxygen uptake was
measured at 673 K over all the catalysts. Oxidation of Ni to
NiO,13,17,22 Co to Co O 25 and Fe to Fe O 26 were
2
3
2 3
2 3
and di†use. However, peaks corresponding to NiO, Co O
3
4
and Fe O , respectively, on the Ni/Al O , Co/Al O and
2
3
2 3
2 3
Fe/Al O catalysts could be identiÐed. No detectable peaks
2
3
indicating any compound formation between the active phase
and support could be identiÐed.
Adsorption properties
Table 2 allows a comparison of the adsorption properties of
the Ni/Al O , Co/Al O and Fe/Al O catalysts. Hydrogen
3
4
2 3
assumed for the calculation of the extent of reduction of the
oxides. The details of pretreatment conditions and experimen-
tal procedures are discussed elsewhere.13h19
2 3
2 3
2 3
uptake for the Fe/Al O and Co/Al O catalysts are very low
2 3 2 3
when compared to Ni/Al O . 10 wt.% Co/Al O and
Acidity of the catalysts were measured by step-wise tem-
perature programmed desorption (STPD) of ammonia using a
Micromeritics Pulse Chemisorb Unit (Model 2700) as
described elsewhere.27
2 3
2 3
Fe/Al O do not show any hydrogen uptake, indicating that
there is hardly any metal available on the surface. Even 50
2 3
wt.% Co/Al O and Fe/Al O show low hydrogen uptake.
2 3
2 3
With increasing metal content, especially beyond 30 wt.%,
hydrogen uptake increases for the nickel catalysts. Hydrogen
uptake capacity of the supported catalysts follows the order
Ni/Al O ? Co/Al O [ Fe/Al O .
Acetone hydrogenation
Acetone hydrogenation was carried out in a Ðxed-bed vertical
glass tubular reactor (10 mm id) containing ca. 0.5 g of cata-
lyst.16 Acetone was fed from the top at a controlled rate along
with a regulated Ñow of hydrogen at atmospheric pressure.
Hydrogenation was studied in the temperature region 373È
2 3
2 3
2 3
Oxygen uptake is indicative of reducibility and going by
this, all the catalysts seen to be reasonably reduced. All the
reduced metal need not be available on the surface and from
the hydrogen and oxygen uptakes, one can conclude that for
Co/Al O and Fe/Al O , most of the metal is in the bulk
5
23 K and at di†erent feed rates. The liquid products collected
2 3
2 3
and/or present as large crystallites (Co/Al O and Fe/Al O
are denser). It is clear from Table 2 that hydrogen chemisorp-
in an ice-trap were analysed using a gas chromatograph with
a 5% Carbowax on Chromosorb W column (1/8A ] 8@) in a
programmed mode from 323 to 373 K at 10 K min~1. The
products identiÐed were 2-P, IBMK, isobutyl methyl carbinol
IBMC), DAA and isophorone. Acetone conversion and
product selectivity were determined by GC.
2 3
2 3
tion for Co/Al O and Fe/Al O is more at 373 K than at 303
2 3
2 3
K, indicating a more activated nature of hydrogen adsorption
on these catalysts than on Ni/Al O . Nevertheless, even at
(
2 3
73 K, hydrogen uptake is greater on the nickel than on the
3
cobalt and iron conÐrming the larger availability of the metal
surface in the nickel catalysts.
The relatively low percentage reduction of Co/Al O and
Results and Discussion
2 3
Fe/Al O catalysts calculated from oxygen uptake can be
2
3
explained as follows. The major oxide species likely to be
present in high metal containing Ni/Al O , Co/Al O and
Physical properties
2
3
2 3
The physical properties of calcined and unreduced co-
precipitated Fe/Al O , Co/Al O and Ni/Al O catalysts are
Fe/Al O catalysts are NiO,28 Co O 25 and Fe O ,26
2
3
3 4
2 3
respectively. Other oxides, such as CoO, Co O and FeO and
2
3
2 3
2 3
2 3
listed in Table 1. A comparison of the Co/Al O and
Fe O , have also been reported for Co/Al O ,29 and
2
3
3 4
2 3
Ni/Al O catalysts13,15 containing the same metal content
Fe/Al O 30 catalysts. In the case of supported nickel, the pre-
2 3
2
3
shows that the Fe/Al O catalysts are denser and have smaller
dominant species present is only NiO28 and it is presumed to
2
3
Table 2 Adsorption data on co-precipitated Ni/Al O , Co/Al O and Fe/Al O catalysts
2
3
2
3
2 3
H2
uptake/lmol g(cat)~1
Ni
Co
Fe
metal
loading
T /K
O2
Ni
uptake/lmol g(cat)~1
reduction (%)
Co
(
wt.%)
303
373
303
373
303
373
Co
Fe
Ni
Fe
10
20
30
40
50
24
56
135
188
286
10
nma
115
nma
240
0
9
36
38
54
0
15
42
52
67
0
5
7
7
10
0
9
12
16
20
506
977
1843
2339
3250
0
866
2066
2679
3679
339
893
1786
2321
3125
61
60
78
76
87
È
41
66
68
77
26
36
50
50
57
a nm; not measured.
124 J. Chem. Soc., Faraday T rans., 1998, V ol. 94
1

View Article Online
Table 4 E†ect of temperature on acetone conversion (%) over co-
be fairly easily reducible to the metallic state in a single step,
especially at high metal level. However, reduction of cobalt
and iron oxides to their zero valent states is believed to be
rather difficult and the metal oxides are expected to exist in
di†erent lower valent oxidation states after reduction. The lit-
erature information also suggests that a complete reduction of
cobalt and iron oxides to the metallic state is difficult and the
process of reduction of Co O to Co involves the intermediate
precipitated Ni/Al O , Co/Al O and Fe/Al O (constant pH,
2
3
2
3
2
3
calcinedÈreduced)
T /K
catalyst
373
423
473
523
1
0 wt.% Ni/Al O
54
0
0
86
31
0
96
50
0
86
11
0
79
70
0
81
76
20
81
64
13
77
65
22
74
64
40
75
62
19
75
65
36
65
60
50
2
3
3
4
10 wt.% Co/Al O
1
30 wt.% Ni/Al O
30 wt.% Co/Al O
steps29 Co O ] CoO ] Co, and the reduction of Fe O to
2 3
Fe involves the intermediate steps30 Fe O ] Fe O ]
2
3
3
3
0 wt.% Fe/Al O
3
4
2
2
3
3 4
2
FeO ] Fe. These intermediate oxides are more difficult to
reduce. Furthermore, there is a likelihood of metalÈsupport
interaction as the catalysts are prepared by co-precipitation
method31 and this may also contribute to the difficulty of
reduction. MetalÈsupport interaction results from a nonuni-
form or uneven incorporation of aluminium into the matrix of
the supported metal oxide particles.26 It has been reported
that aluminium dissolution also occurs in Ni/Al O 32 but
2
3
3
3
3
5
0 wt.% Fe/Al O
2
0 wt.% Ni/Al O
2
5
0 wt.% Co/Al O
2
3
3
5
0 wt.% Fe/Al O
2
Acetone \ 4.5 cm3 h~1, weight of the catalyst \ 0.5 g.
2
3
2
may be to a lesser extent when compared to Co/Al O and
Fe/Al O . Co/Al O and Fe/Al O catalysts can contain
been reported33 that the strength of adsorption of acetone on
these metals increases in the order Ni \ Co \ Fe. The tem-
perature at which maximum conversion obtained is also in the
same order. Also, at high temperature (P523 K), acetone can
undergo decomposition to yield gaseous products, which are
not monitored during the experiments. At 473 K, all the cata-
3
2
3
2 3
2 3
both reduced metal as well as some metal oxides in di†erent
oxidation states. It can be presumed that oxygen uptake by
reduced Co/Al O and Fe/Al O might include oxidation of
both zero-valent as well as the lower oxidic forms of Co and
Fe to Co O and Fe O , respectively. Therefore, calculation
2
3
2 3
lysts show reasonable activity in the order, Ni/Al O [
3
4
2 3
2 3
of the extent of reduction from oxygen uptake might not be
truly representative for Co/Al O and Fe/Al O catalysts.
Co/Al O [ Fe/Al O . Therefore, the conversion and selec-
2
3
2 3
tivities of the catalysts are compared at this temperature.
Acetone hydrogenation over alumina support, metal oxides
and supported metal oxides is given in Table 5. The conver-
sion of acetone on alumina is only 10% and the product
formed is mostly isophorone. Unsupported metal oxides are
reduced in hydrogen at 723 K for 24 h before reaction. Both
nickel and cobalt oxides convert acetone at only 50% to give
selectively 2-P. Reduced iron oxide, on the other hand, is only
very slightly active (only 5% conversion) for acetone hydro-
genation. This conforms with the poor reducibility of iron
oxide and negligible acidity. Of the supported metal oxides,
nickel is the most active closely followed by cobalt and iron,
falling at a distant third place. Increasing the metal content
does not show any signiÐcant increase in activity for the
nickel and cobalt catalyst. Fe/Al O shows low acetone
2
3
2 3
Nevertheless, such a calculation still gives a comparative and
qualitative picture of the extent of reduction of the catalysts.
For Ni/Al O , the calculation might be reasonably accept-
2
3
able.
Acidity of the catalysts
The acidity of the catalysts, as determined by STPD of
ammonia, is shown in Table 3. Alumina is very acidic com-
pared with the oxides of Ni, Co and Fe. Among the supported
metal catalysts, there is no signiÐcant di†erence in acidity
except that it decreases with metal loading.
Acetone hydrogenation
2
3
hydrogenation activity (10 wt.% Fe), which improves with
metal content. The selectivity for 2-P, which involves metal
functions, improves with increasing metal loading in all the
three metal systems. The metal function of the Ni/Al O cata-
Acetone hydrogenation reaction was carried out over 10, 30
and 50 wt.% metal supported on alumina at di†erent tem-
peratures (Table 4). The conversion increases with increasing
metal content. Only the Ni/Al O catalyst is active at load-
2
3
2
3
lysts is prominent at even a 10 wt.% metal loading. In agree-
ment with the hydrogen adsorption properties, the 10 wt.%
Co/Al O and Fe/Al O catalysts show very little metallic
ings of 10 wt.% and at temperatures as low as 373 K.
Co/Al O shows an improvement in activity with increasing
2
3
metal content at 373 K whereas Fe/Al O does not show any
2
3
2 3
2
3
character in acetone hydrogenation and hence there is no for-
mation of 2-P and IBMK. However, the acidic function of the
catalyst, due to alumina, predominates yielding mostly iso-
phorone and with increasing metal content the metallic char-
acter also contributes yielding 2-P and IBMK.
Fig. 1 describes the product formation and selectivity di†er-
ences for acetone hydrogenation over 50 wt.% metal on the
alumina catalysts. The selectivity patterns reveal that low tem-
peratures favour 2-P and high temperature favours IBMK
and other products.
activity. Table 4 shows that conversion increases with increas-
ing temperature up to a particular value and then decreases.
The fall in conversion is attributed only to a temperature
e†ect. The maximum conversion obtained varies with the
nature of catalyst (Table 4). This could be due to the di†erence
in the strength of adsorption of acetone on these metals. It has
Table 3 Acidity of the catalyst, constant pH, calcinedÈreduced (NH
desorption between 353 and 673 K)
3
The inÑuence of contact time on acetone conversion and
product selectivity is shown in Fig. 2. The reaction was carried
out over 50 wt.% metal catalysts at 523 K. The contact time
was varied by varying the acetone feed and keeping the weight
of the catalyst at 0.5 g. Generally, increasing contact time
increases conversion. The nickel catalyst seems to give selec-
tively more IBMK than 2-P at low contact time compared
with Co/Al O and Fe/Al O catalysts and the two products
acidity/mmol NH g(cat)~1
3
metal loading (wt.%)
catalyst
Al O
oxide
10
30
50
0.63
0.0
0.06
0.06
È
È
È
È
È
È
È
È
È
È
È
È
2
3
Ni oxide
Co oxide
Fe oxide
Ni/Al O
2
3
2 3
È
È
È
become nearly equal at high contact time. The Co/Al O
0.72
0.86
0.74
0.70
0.80
0.61
0.62
0.73
0.42
2 3
2
Co/Al O
2
3
catalyst gives the simple hydrogenated product, namely 2-P,
3
more than IBMK (cf. Ni/Al O ). With increasing contact time
Fe/Al O
2
3
2
3
the formation of 2-P and IBMK becomes nearly equal to that
J. Chem. Soc., Faraday T rans., 1998, V ol. 94 1125

View Article Online
Table 5 Acetone hydrogenation over support, metal oxides and supported metal oxides
selectivity (%)
conversion
catalyst
Al O
0 wt.% Ni/Al O
0 wt.% Co/Al O
(%)
2-P
IBMK
IBMC
DAA
isophorone
10
81
64
13
77
65
22
74
64
40
50
50
5
0
40
0
0
35
4
0
13
2
30
6
7
66
0
0
25
0
0
42
9
0
5
16
0
0
9
0
0
4
91
0
81
50
0
0
0
0
0
0
0
2
3
1
1
1
3
3
3
5
5
2
3
2
3
3
3
0 wt.% Fe/Al O
0
4
2
0 wt.% Ni/Al O
2
64
69
4
73
80
36
100
100
8
25
24
17
20
19
34
0
0 wt.% Co/Al O
2
3
3
3
0 wt.% Fe/Al O
2
0 wt.% Ni/Al O
2
0 wt.% Co/Al O
2
3
3
5
0 wt.% Fe/Al O
2
NiOa
0
0
0
Co O
a
0
50
0
0
3
4 a
Fe O
2
3
Temperature \ 473 K, weight of the catalyst \ 0.5 g, acetone \ 4.5 cm3 h~1. a Reduced in hydrogen at 723 K for 24 h.
of Ni/Al O . However, Fe/Al O gives all the products in the
2 3
conversion level varies from 35 to 70% for Ni/Al O , from 47
2 3
2
3
feed-rate region studied.
to 65% for Co/Al O , and from 23 to 50% for Fe/Al O as
2
3
2 3
The selectivity results discussed above and shown in Fig. 2
for di†erent contact times over the three catalysts are not
based on constant conversion of acetone. For example, the
the contact time increases from 3.1 to 8.2 h g(cat) mol~1. It is
rather difficult to compare the catalyst efficiency or behaviour
towards the selectivity at di†erent contact times and tem-
Fig. 1 E†ect of temperature on acetone conversion and selectivity
Fig. 2 E†ect of contact time of acetone on conversion and product
over 50 wt.% Ni/Al O , Co/Al O and Fe/Al O catalyst.
selectivity over 50 wt.% Ni/Al O , Co/Al O and Fe/Al O catalysts.
2
3
2
3
2
3
2
3
2
3
2 3
Acetone \ 4.5 cm3 h~1. 1, Conversion; 5, 2-P; 3, IBMK; C,
Catalyst \ 0.5 g, temperature \ 523 K. 1, Conversion; 5, 2-P; 3,
IBMC; 4, DAA.
IBMK; C, IBMC; 4, DAA.
1126
J. Chem. Soc., Faraday T rans., 1998, V ol. 94

View Article Online
perature conditions because the conversion levels vary widely.
If one wants to get a general picture of selectivity pattern, one
has to compare the catalysts at nearly constant conversion
levels thereby avoiding the inÑuence of drastic di†erences in
conversion on selectivity.
In order to get a real picture of the selectivity patterns, 50
wt.% Ni/Al O , Co/Al O and Fe/Al O are compared at a
2
3
2 3
2 3
conversion level of nearly 50% (Fig. 3). The reaction was
carried out at 523 K with 0.5 g of catalyst. The constant con-
version was achieved by varying feed rates. Ni/Al O is found
2
3
to be more selective for producing IBMK than 2-P whereas
Co/Al O is more selective for producing 2-P than IBMK.
Moreover, other products like IBMC are also formed over
Co/Al O , at a more signiÐcant level than over Ni/Al O .
2
3
2
3
2 3
Over Fe/Al O , there is no selectivity preference for any of
2
3
Fig. 3 Comparison of acetone hydrogenation product selectivity
the products and all of them are formed in nearly equal
amounts. The pattern obtained is almost the same even at dif-
ferent acetone contact times over the catalysts (see Fig. 2).
This experiment was repeated with other metal wt.% catalysts,
i.e., 30 and 40 wt.%, and the product selectivity pattern
obtained was the same as that shown in Fig. 3.
over 50 wt.% Ni/Al O , Co/Al O and Fe/Al O catalysts at a con-
2
3
2
3
2 3
stant conversion level of nearly 50%. Catalyst \ 0.5 g,
temperature \ 523 K. 5, 2-P; 3, IBMK; C, IBMC; 4, DAA.
Unsupported nickel and cobalt oxides, which are fairly
easily reducible to the metallic state, produce only 2-P. 10
wt.% Co/Al O and Fe/Al O are poorly reduced (Table 2)
The conversion and selectivity studies of acetone hydro-
genation reaction over the Ni/Al O , Co/Al O and
Fe/Al O catalysts emphasize the metallic and acidic func-
2
3
2 3
2 3
2 3
and hence the contribution of metal to the reaction is low. In
other words, the acidity of these catalysts is more instrumental
in determining the product selectivity and hence isophorone is
formed preferentially (Table 5). For higher metal containing
catalysts, both the metallic and acidic nature contribute in
deciding the selectivity depending upon their extent of contri-
bution. As the metal loading increases, 2-P selectivity
increases (Table 5). Thus, the product selectivity is governed
by a combination of the metallic and acidic properties of the
catalyst and the extent of their involvement.
2
3
tionalities of the Ni/Al O more than the Co/Al O or
2 3
Fe/Al O catalysts, as evidenced by the selective formation of
2
3
2 3
-P and IBMK. The less reducible Fe/Al O catalyst gives
2
2
3
other hydrogenation and condensation products in addition
to 2-P. The catalytic properties observed are in accordance
with the hydrogen adsorption properties as well as the
reducibility of the catalysts. In the acetone hydrogenation
reaction using the supported metal catalysts, both the metallic
as well as the acidic functions of the catalysts are involved.
The product formation depends on the extent of involvement
of these two sites and their availability. Based on the dual
function characteristics of these catalysts, the formation of dif-
ferent products can be postulated following the mechanism
proposed by Gandia and Montes.5 In fact, three routes are
possible.
Conclusions
From the above studies, the following conclusions can be
made. The metallic character promotes the conversion of
acetone to 2-P or IBMK depending on the extent of involve-
ment of the metallic and acidic functions of the catalyst.
Alumina gives only isophorone. Low metal loading and the
involvement of the acidity of the catalyst favour IBMK forma-
tion. An increase in metal content increases 2-P production.
The overall acetone hydrogenation activity follows the order
Ni/Al O [ Co/Al O [ Fe/Al O . However, the product
Route (i) is a single-step direct hydrogenation of acetone to
2-P over a metallic site. Route (ii) is a bi-functional, multi-step
route involving both the metallic phase as well as the acidÈ
base nature of the catalyst giving a variety of products. Route
(
iii) is a condensation reaction of three molecules of acetone to
isophorone involving the acidic function of the catalyst.
2
3
2 3
2 3
selectivity distribution depends on the metal content, the type
as well as reaction conditions. Hydrogenation activity of the
catalysts are in accordance with their hydrogen adsorption
properties. The mechanism of acetone conversion, involving a
dual function of supported metal catalyst, is discussed. A
correlation of catalyst characteristics with catalysis and the
product selectivity is attempted with the available informa-
tion. An undisputed mechanistic explanation of di†erent
product formation will require more detailed studies.
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.
References
1
2
3
4
5
6
7
L. M. Gandia and M. Montes, Int. J. Energy Res., 1992, 16, 851.
L. M. Gandia and M. Montes, Appl. Catal. A, 1993, 101, L1.
Y. Q. Zhengyiwang and Ren Wang, Appl. Catal. A, 1989, 53, 63.
L. M. Gandia, A. Diaz and M. Montes, J. Catal., 1995, 157, 157.
L. M. Gandia and M. Montes, J. Mol. Catal., 1994, 94, 347.
B. Sen and M. A. Vannice, J. Catal., 1988, 113, 52.
M. Nakamura and H. Wise, Proc. 6th Int. Congr. Catal. L ondon,
1
976, ed. G. C. Bond, P. B. Wells and F. C. Tompkins, The
Chemical Society, London, 1977, p. 881.
8
F. Rositani, S. Galvagno, Z. Poltarzews, P. Staiti and P. L.
Antonucci, J. Chem. T echnol. Biotechnol., 1985, 35A, 234.
J. Chem. Soc., Faraday T rans., 1998, V ol. 94
1127

View Article Online
9
0
J. Simonikova, A. Ralkova and K. KochloeÑ, J. Catal., 1973, 29,
12.
J. Simonikova, L. Hillaire, J. Panck and K. KochloeÑ, J. Phys.
Chem., 1973, 83, 287.
A. Farkas and L. Farkas, J. Am. Chem. Soc., 1939, 61, 1336.
C. T. H. Stoddart and C. Kemball, J. Colloid Sci., 1956, 11, 532.
S. Narayanan, R. Unnikrishnan and V. Vishwanathan, Appl.
Catal. A, 1995, 129, 9.
R. Unnikrishnan, V. Vishwanathan and S. Narayanan, Catalysis:
Modern T rends, ed. N. M. Gupta and D. K. Chakrabarty,
Narosa, New Delhi, 1996, p. 333.
21 V. Vishwanathan, M. Rajashekar, G. Sreekanth and S. Naraya-
4
nan, J. Chem. Soc., Faraday T rans. 1, 1991, 87, 3449.
22 C. H. Bartholomew and R. J. Farrauto, J. Catal., 1976, 45, 41.
23 J. M. Zowtiak, G. D. Weatherbee and C. H. Bartholomew, J.
Catal., 1983, 82, 230.
24 G. D. Weatherbee, J. L. Rankin and C. H. Bartholomew, Appl.
Catal. A, 1984, 11, 73.
25 R. L. Chin and D. M. Hercules, J. Phys. Chem., 1982, 86, 360.
26 H. Jung and W. J. Thomson, J. Catal., 1991, 128, 218.
27 S. Narayanan and K. Deshpande, Appl. Catal. A, 1996, 135, 125.
28 J. Zielinski, J. Catal., 1982, 76, 157.
29 P. Arnoldy and J. A. Moulijn, J. Catal., 1985, 93, 38.
30 A. J. H. M. Kock, H. M. Fortuin and J. W. Geus, J. Catal., 1985,
96, 261.
31 A. Sannomiya, K. Ichimura, M. Yano and Y. Harano, Memoirs of
the Faculty of Engineering, Osaka City Univ., 1983, 24, 161.
32 P. H. M. de Korte, E. B. M. Doesburg, C. P. J. de Winter and L.
L. van Reijen, Solid State Ionics, 1985, 16, 73.
33 G. I. Golodets, N. V. Pavlenko, A. I. Tripolskii and G. M. Telbiz,
Kinet. Katal., 1985, 26, 557.
1
1
1
1
1
2
3
1
4
1
1
1
1
1
2
5
6
7
8
9
0
S. Narayanan and R. Unnikrishnan, J. Chem. Soc., Faraday
T rans., 1997, 93, 2009.
S. Narayanan and R. Unnikrishnan, Appl. Catal. A, 1996, 145,
2
31.
S. Narayanan and K. Uma, J. Chem. Soc., Faraday T rans. 1,
985, 81, 2733.
S. Narayanan and K. Uma, J. Chem. Soc., Faraday T rans. 1,
987, 83, 733.
S. Narayanan and K. Uma, J. Chem. Soc., Faraday T rans. 1,
988, 84, 521.
S. Narayanan and G. Sreekanth, J. Chem. Soc., Faraday T rans. 1,
989, 85, 3785.
1
1
1
1
Paper 7/08124C; Received 11th November, 1997
1128
J. Chem. Soc., Faraday T rans., 1998, V ol. 94