One-step synthesis of methyl isobutyl ketone from acetone catalysed by Pd supported on ZnII-CrIII mixed oxide

Article abstract of DOI:10.1016/j.jcat.2005.11.028

Pd metal (0.1-1 wt%) supported on Zn^II-Cr^III (1:10) mixed oxide was found to be an efficient bifunctional catalyst for the one-step synthesis of methyl isobutyl ketone (MIBK) from acetone and H₂ in gas- and liquid-phase processes. The reaction involves the acid-catalysed condensation of acetone to form mesityl oxide, followed by its hydrogenation to MIBK. The gas-phase reaction produces MIBK with a selectivity up to 78% at 40-66% acetone conversion (200-350 °C, ambient pressure). Diisobutyl ketone (DIBK) is the main byproduct, with a total MIBK + DIBK selectivity of up to 93%. The catalyst reaches a steady state in ca. 1 h and shows constant activity and selectivity for at least 50 h on stream. The liquid-phase reaction yields MIBK with up to 83% selectivity and total selectivity to MIBK + DIBK up to 9 at 56% acetone conversion (200 °C, 5 bar H₂ pressure). As a catalyst support, the amorphous Zn-Cr oxide calcined at 300 °C provides higher catalytic activity than the crystalline Zn-Cr oxide calcined at 400 °C, probably due to the lower surface area of the crystalline oxide. XRD of 0.3-1%Pd/Zn-Cr oxide showed no pattern of Pd metal, indicating a fine dispersion of Pd particles in the catalyst.

Full text of DOI:10.1016/j.jcat.2005.11.028

Journal of Catalysis 238 (2006) 286–292
www.elsevier.com/locate/jcat
One-step synthesis of methyl isobutyl ketone from acetone catalysed
II
III
by Pd supported on Zn –Cr mixed oxide
∗
Elena F. Kozhevnikova, Ivan V. Kozhevnikov
a
Department of Chemistry, University of Liverpool, Liverpool L69 7ZD, UK
Received 19 September 2005; revised 7 November 2005; accepted 21 November 2005
Available online 18 January 2006
Abstract
II
III
Pd metal (0.1–1 wt%) supported on Zn –Cr (1:10) mixed oxide was found to be an efficient bifunctional catalyst for the one-step synthesis
of methyl isobutyl ketone (MIBK) from acetone and H in gas- and liquid-phase processes. The reaction involves the acid-catalysed condensation
2
of acetone to form mesityl oxide, followed by its hydrogenation to MIBK. The gas-phase reaction produces MIBK with a selectivity up to 78%
◦
at 40–66% acetone conversion (200–350 C, ambient pressure). Diisobutyl ketone (DIBK) is the main byproduct, with a total MIBK + DIBK
selectivity of up to 93%. The catalyst reaches a steady state in ca. 1 h and shows constant activity and selectivity for at least 50 h on stream. The
◦
liquid-phase reaction yields MIBK with up to 83% selectivity and total selectivity to MIBK + DIBK up to 9 at 56% acetone conversion (200 C,
◦
5
bar H pressure). As a catalyst support, the amorphous Zn–Cr oxide calcined at 300 C provides higher catalytic activity than the crystalline
Zn–Cr oxide calcined at 400 C, probably due to the lower surface area of the crystalline oxide. XRD of 0.3–1%Pd/Zn–Cr oxide showed no
pattern of Pd metal, indicating a fine dispersion of Pd particles in the catalyst.
2
◦
©
2005 Elsevier Inc. All rights reserved.
Keywords: Acetone; Methyl isobutyl ketone; One-step process; Palladium; Zn–Cr mixed oxide; Bifunctional catalysis
1
. Introduction
acidic cation-exchangers, zeolites, or zirconium phosphate with
added platinum group metals, generally palladium. In one-step
processes, the three steps occur simultaneously to yield MIBK
in an exothermic reaction (ꢀH = −117 kJ/mol) [2]. Although
high selectivity to MIBK (80–95% at 30–50% conversion) has
been obtained, the high pressure is a disadvantage. Because the
one-step process is simpler and more economically attractive,
there is a great interest in finding new, improved catalyst sys-
tems operating at lower pressures. One-step synthesis of MIBK
Methyl isobutyl ketone (MIBK) is one of the most widely
used aliphatic ketones. Its primary uses included as a sol-
vent for paint and protective coatings, an extracting agent for
the production of antibiotics and commercial lubricating oils,
a reagent for the separation and purification of metal ions, and
an initiator for polymerisation [1]. Diisobutyl ketone (DIBK),
a consecutive product in the synthesis of MIBK, is an excep-
tionally good solvent for a wide variety of natural and syn-
thetic resins. It is also used in pharmaceutical and mining in-
dustries. Traditionally, MIBK is manufactured via a three-step
process: (1) base-catalysed aldol condensation of acetone to di-
acetone alcohol (DA), (2) acid-catalysed dehydration of DA to
mesityl oxide (MO), and (3) metal-catalysed hydrogenation of
in the gas phase is also attractive. Usually, this is carried out at
◦
1
40–340 C and ambient pressure; however, the MIBK selec-
tivity is generally lower than in the liquid-phase reaction, and
catalyst deactivation may be a problem in this process.
The proposed general mechanism of the one-step synthe-
+
sis of MIBK on a bifunctional catalyst Pd/{H } is shown in
◦
MO to MIBK [2]. One-step processes operating at 120–200 C
and 20–50 bar H2 pressure in liquid phase have been devel-
oped using bifunctional acid–base/redox catalysts comprising
Scheme 1 [3]. MIBK forms in three steps (route 1): (1) acetone
condensation to DA over acid sites, (2) acid-catalysed dehy-
dration of DA to MO, and (3) selective hydrogenation of the
C=C bond of MO on Pd sites to yield MIBK [4]. Concurrently,
the Pd-catalysed hydrogenation of the C=O group of acetone
yields isopropanol (IP), which can further be dehydrated on
*
Corresponding author. Fax: +44 151 794 3589.
E-mail address: kozhev@liverpool.ac.uk (I.V. Kozhevnikov).
0021-9517/$ – see front matter © 2005 Elsevier Inc. All rights reserved.
doi:10.1016/j.jcat.2005.11.028

E.F. Kozhevnikova, I.V. Kozhevnikov / Journal of Catalysis 238 (2006) 286–292
287
Scheme 1.
the acid sites to propene, followed by the formation of propane
Table 1
Catalyst characterisation
via propene hydrogenation on Pd sites (route 2). Therefore, the
selectivity to MIBK depends on the relative rates of routes 1
and 2, which can be varied by tuning the acid and hydrogena-
tion properties of the bifunctional catalyst [5]. Other products,
such as DIBK and C9+ condensation products, may also form in
further reactions of MIBK (Scheme 2) [6]. More about various
products of this reaction can be found elsewhere [6]. The over-
all selectivity depends not only on the active sites in the catalyst,
but also on the catalyst pore structure controlling the diffusion
of reactants and products to and from the active sites [4].
Numerous bifunctional catalysts have been studied in the
one-step synthesis of MIBK in both gas-phase and liquid-phase
reactions [6–20]. These include Pt supported on gallosilicate
and alumosilicate [7], NaX, and CsX zeolites [8], Pd on ZSM-
a
Entry Catalyst
S
BET
2
Pore
Pore volume H O content
2
3
(wt%)
(m /g) size (Å) (cm /g)
1
2
3
4
5
6
Zn–Cr (1:10) oxideb
169
39
79
35
43
29
67
0.16
0.17
0.15
0.16
0.11
0.14
8.3
2.0
6.0
6.4
6.0
2.2
c
Zn–Cr (1:10) oxide (cr) 87
1%Pd/Zn–Cr
159
150
135
84
0.3%Pd/Zn–Cr
0.1%Pd/Zn–Cr
1%Pd/Zn–Cr (cr)
d
a
◦
From TGA as a weight loss in the temperature range of 30–700 C.
b
◦
Amorphous oxide calcined at 300 C for 5 h under N ; 8.4 wt% Zn and
2
63.0 wt% Cr content in anhydrous oxide (from ICP).
c
Crystalline oxide obtained from the amorphous oxide (entry 1) by calcina-
◦
tion at 400 C for 5 h under N .
2
d
Crystalline Zn–Cr (1:10) oxide as a support.
5
[9], SAPO-11 and AlPO-11 [10], zirconium phosphate [17],
actions, such as hydrocarbon fluorination by HF [25,26]. We
found that the Pd/Zn–Cr catalyst exhibits high activity and se-
lectivity, as well as long durability, in the one-step synthesis of
MIBK.
niobic acid and alumina [18,19], and Pd, Ni, Cu, or Pd–Cu sup-
ported on MgO, Mg/Al hydrotalcite, and alumina [6,11–16,20].
Here we report the one-step synthesis of MIBK in gas and liquid
phase over a new catalyst: palladium supported on ZnO–Cr2O3
mixed oxide, Pd/Zn–Cr.
Zn–Cr oxide has long been used as a catalyst for various re-
actions of synthesis gas (CO + H2), for example, the synthesis
of methanol [21,22] and higher alcohols [23] and aliphatic and
aromatic hydrocarbons [24]. It has also been used for the fluo-
rination of hydrocarbons with HF [25,26], dehydrogenation of
alcohols [27], hydrogenation of carboxylic acids [28], synthesis
of quinoline from aniline and glycerol [29], and selective cat-
alytic reduction of NO with NH3 [30]. Pt/Zn–Cr has proven to
be an effective catalyst for the low-temperature dehydrogena-
tion of isobutane to isobutene [31]. To our knowledge, to date
neither Zn–Cr oxide nor Pd/Zn–Cr has been used for the synthe-
sis of MIBK. Zn–Cr oxide has been studied over a wide range of
Zn/Cr atomic ratios [21,22,24]. Here we chose a Zn–Cr (1:10)
oxide, which has a significant (Brønsted and Lewis) acidity,
as evidenced by its high catalytic activity in acid-catalysed re-
2
. Experimental
2
.1. Catalyst preparation
The Zn–Cr (1:10) mixed oxides (Zn/Cr atomic ratio of 1:10)
II
III
were prepared via coprecipitation of Zn and Cr hydroxides
using a procedure adapted from that reported elsewhere [22,
2
4,31]. The coprecipitation was carried out by adding 10 wt%
aqueous ammonium hydroxide dropwise to a stirred 0.2 M
II
III
aqueous solution of a mixture of Zn + Cr nitrates (1:10
◦
molar ratio) at 70 C until pH 7 was reached, followed by ag-
◦
ing the suspension for 3 h at 70 C. The precipitate was filtered
off, washed with deionised water until ammonia-free, and dried
in air at 100–110 C overnight. The amorphous and crystalline
◦
Zn–Cr oxides were obtained by calcination of the mixed hy-
droxides under nitrogen flow for 5 h, the amorphous oxide at
◦
◦
3
00 C and the crystalline oxide at 400 C.
The Pd-doped catalysts were prepared by stirring a Zn–Cr
oxide powder with 0.02 M solution of Pd(OAc)2 in benzene at
room temperature for 1 h, followed by slow evaporation of the
solvent in a rotary evaporator. The material was then reduced
◦
under hydrogen flow (30–40 ml/min) at 250 C for 2 h. The
catalysts were ground into a powder (particle size ꢀ180 µm
unless stated otherwise) and stored in a desiccator over P2O5.
3
Scheme 2.
The density of the catalyst powder was 1.54 g/cm . The Zn, Cr,

288
E.F. Kozhevnikova, I.V. Kozhevnikov / Journal of Catalysis 238 (2006) 286–292
and Pd content of the catalysts thus prepared were determined
by ICP. Catalyst characterisation is given in Table 1.
2
.2. Techniques
The BET surface area and porosity of Zn–Cr oxides and
Pd/Zn–Cr catalysts were measured by nitrogen physisorption
at 77 K on a Micromeritics ASAP 2000 instrument. Before the
◦
analysis, the samples were evacuated at 250 C for 4–6 h. Wa-
ter content in the catalysts was measured by thermogravimet-
ric analysis (TGA) under nitrogen flow using a Perkin-Elmer
TGA 7 instrument. The amount of coke deposited on the used
catalysts was measured by TGA under air as described else-
where [32]. Powder X-ray diffraction (XRD) spectra of cata-
lysts (phase analysis) were recorded on a Stoe Stadi-P diffrac-
tometer with a monochromatic Cu-Kα1 radiation from a ger-
Fig. 1. XRD patterns from top to bottom: crystalline Zn–Cr oxide (Table 1, en-
try 2), amorphous Zn–Cr oxide (Table 1, entry 1), fresh 1%Pd/Zn–Cr (Table 1,
entry 3).
◦
◦
manium monochromator in the angular range 20 ꢀ 2θ ꢀ 70 ,
◦
with a step width of 0.04 and count time of 5 s per step. XRD
contained 2.0 g of acetone, 0.30 g of decane (internal GC stan-
dard), and 0.20 g of catalyst. The autoclave was pressurised
with H2 and placed in an oil bath that was preheated to the re-
action temperature. Judging from the resulting pressure, most
of the acetone remained liquid under such conditions. After the
patterns were attributed using the JCPDS database. Ammonia
adsorption onto the catalysts was measured in a flow system
using a Setaram TG-DSC 111 differential scanning calorimeter
as described elsewhere [33]. ICP analysis was carried out on a
Spectro Ciros emission spectrometer.
completion of reaction (typically 2 h), the reactor was cooled to
◦
0
C, depressurised in a gas bag with a stopcock, and opened.
2
2
.3. Catalyst testing for MIBK synthesis
The reaction mixture was taken out of the autoclave and sepa-
rated from the catalyst by centrifugation. Products were iden-
tified by GC-MS and quantified by a gas chromatograph (Var-
ian Star 3400 CX) equipped with a flame ionisation detector.
A 30 m × 0.25 mm BP5 capillary column was used to analyse
gas products (collected in the gas bag), and a 30 m × 0.25 mm
HP-INNOWAX capillary column was used for liquid samples.
.3.1. Gas-phase process
The catalytic tests were performed under atmospheric pres-
sure in a Pyrex glass fixed-bed microreactor (9 mm i.d.) with
on-line GC analysis (Varian Star 3400 CX gas chromatograph
equipped with a 30 m × 0.25 mm BP5 capillary column and
flame ionisation detector). The reactor was placed in a vertical
tubular furnace. The catalyst bed containing 0.1–0.5 g of a cat-
3
3. Results and discussion
alyst powder (ꢀ180 µm particle size; 1.54 g/cm density) was
placed in the reactor between two layers of Pyrex glass wool.
The temperature in the reactor was controlled with a Eurotherm
controller using a thermocouple placed at the top of the catalyst
bed. The gas feed was fed into the reactor from the top. Ace-
tone was supplied to the gas flow by bubbling a flow of H2 or an
H2/N2 mixture (typically 10–20 ml/min, controlled by Brooks
mass flow controllers) through a stainless steel saturator con-
taining acetone, maintained at a certain temperature to maintain
the chosen acetone vapour pressure. All gas lines were made of
stainless steel. The downstream lines and sampling valves were
3
.1. Catalyst characterisation
The bifunctional acid–base/redox catalyst under study com-
prised Zn–Cr (1:10) mixed oxide as an acid–base compo-
nent and Pd metal as a redox (hydrogenation) component. De-
pending on the calcination temperature, Zn–Cr oxide can be
amorphous (ꢀ300 C) or crystalline (>350 C), with the lat-
ter including, in general, ZnO, Cr2O3, and ZnCr2O4 crystalline
phases [22,24]. We used mostly the amorphous hydrous Zn–Cr
◦
◦
◦
◦
heated at 150 C to prevent product condensation. Before reac-
(1:10) oxide calcined at 300 C. It showed no crystallinity in its
2
tion, the catalysts were pretreated with hydrogen (10 ml/min)
XRD spectrum (Fig. 1) and had a surface area of 169 m /g, an
◦
at 300 C for 1 h. At regular time intervals, the downstream
average pore size of 39 Å, and a water loss of 8.3 wt% on heat-
ing to 700 C (Table 1, entry 1). The corresponding crystalline
◦
gas flow was analysed by on-line GC. The liquid products were
collected in an ice trap and also analysed by off-line GC using
a 30 m × 0.25 mm HP-INNOWAX capillary column (Agilent
Technologies). The products were identified by GC and GC-MS
using authentic samples.
◦
Zn–Cr (1:10) oxide obtained by calcination at 400 C exhibited
clear patterns of ZnCr2O4 spinel and Cr2O3 crystalline phases
(Fig. 1), in agreement with the literature [22,24]. It had a sur-
2
face area of 87 m /g, an average pore size of 79 Å, and a water
content of 2.0 wt% (Table 1, entry 2).
2
.3.2. Liquid-phase process
The liquid-phase synthesis of MIBK was carried out in a 45-
The acidity of the amorphous Zn–Cr oxide was characterised
◦
by NH3 adsorption at 100 C using TG-DSC (Fig. 2). The ox-
◦
ml Parr 4714 stainless steel autoclave equipped with a pressure
gauge and a magnetic stirrer. Typically, the reaction mixture
ide was pretreated at 300 C for 2 h under N2 (35 ml/min),
◦
followed by stabilisation at 100 C for 1 h. Then it was brought

E.F. Kozhevnikova, I.V. Kozhevnikov / Journal of Catalysis 238 (2006) 286–292
289
metal [35]. A wide range of methods is used for supporting Pd
nanoparticles, and the Pd particles obtained thusly vary consid-
erably in size [35]. Here we used the impregnation of Zn–Cr
oxide by a dilute (0.02 M) benzene solution of Pd(OAc) , fol-
2
II
0
lowed by evaporation of the solvent and reduction of Pd to Pd
◦
by a H2 flow under relatively mild conditions (250 C). Note
that the size of the deposited Pd particles increased with in-
II
creasing temperature of Pd reduction [36]. To estimate the dis-
persion of the Pd phase on Zn–Cr oxide, we used XRD, which
can measure the average size of Pd particles [35,36]; however,
XRD of 0.3–1%Pd/Zn–Cr, both fresh and used, revealed no pat-
terns of Pd metal. Thus the most intense 111 diffraction peak
Fig. 2. TG-DSC for ammonia adsorption on the amorphous Zn–Cr oxide (Ta-
ble 1, entry 1) at 100 C.
◦
◦
of the Pd phase at 2Θ = 40.1 was not seen (Fig. 1), indicat-
ing a very fine dispersion of Pd in the Pd/Zn–Cr catalysts. This
may be one reason for the high efficiency of these catalysts (see
below). Loading Pd (0.1–1 wt%) on Zn–Cr oxides slightly de-
creased the surface area and porosity of the catalysts (Table 1),
as expected. Note that we were unable to accurately measure
the NH3 adsorption on Pd/Zn–Cr, because TGA showed an ad-
sorption peak rather than a plateau, which could be caused by
the base-catalysed loss of water from the oxide [33].
in contact with a NH3 + N2 (1:1) gas flow (35 ml/min),
which caused a sharp weight gain and heat release. After 1
h, the gas flow was switched to pure N . The initial NH ad-
2
3
sorption under NH3 flow on Brønsted and Lewis acid sites
was 1.95 mmol/gcat, and the enthalpy of adsorption was
−
117 kJ/molNH . After the gas flow was switched to N2,
.52 mmol/gcat ammonia remained adsorbed on the oxide, re-
3
1
flecting the number of stronger acid sites in the amorphous
Zn–Cr oxide. Similarly, for the crystalline Zn–Cr oxide, the ini-
3
.2. Gas-phase synthesis of MIBK
tial adsorption of NH was 1.21 mmol/gcat, and the enthalpy of
3
The gas-phase reaction was studied in a fixed-bed flow reac-
adsorption was −127 kJ/molNH . This demonstrates that both
◦
3
tor under the following conditions: 200–350 C, ambient pres-
oxides had acid sites of comparable strength, close to that of
HY zeolite (115–130 kJ/mol) [34]. As expected, the crystalline
oxide had a smaller density of acid sites due to its lower surface
area.
−1
sure, GHSV of 6500–18,000 h , 9–50 vol% acetone, and
2
3–70 vol% H2 in the gas feed (N2 balance). The Pd load-
ing on the amorphous Zn–Cr oxide varied from 0.1 to 1 wt%.
◦
Representative results at 300 C are given in Table 2. MIBK
The deposition of Pd on Zn–Cr oxide provides the catalyst
with hydrogenation activity. This activity strongly depends on
the dispersion of the Pd phase, because the size of Pd particles
is a key parameter of the catalytic efficiency of the supported
was the major reaction product, and a significant amount of
DIBK also formed. Byproducts included C9 acetone con-
+
densation products (mainly the tetramer 2,6,8-trimethylnonane-
4-one) and small amounts (ꢀ1% selectivity) of propene and
Table 2
a
Gas-phase synthesis of methyl isobutyl ketone over Pd/Zn–Cr catalysts
b
Entry
Catalyst
GHSV
Conversion
(%)
Selectivity (%)
−
1
(h
)
MIBK
DIBK
C
3
MO
Others
1
2
3
4
5
6
7
8
9
Zn–Cr oxide
1%Pd/Zn–Cr
1%Pd/Zn–Cr
1%Pd/Zn–Cr
1%Pd/Zn–Cr
1%Pd/Zn–Cr(cr)
0.3%Pd/Zn–Cr
0.3%Pd/Zn–Cr
12,900
12,900
6500
6500
6500
6500
6500
6500
6500
16.7
40.7
57.3
42.0
66.4
48.7
41.6
36.9
36.1
32.1
52.5
37.9
15.3
69.4
64.1
77.6
43.8
55.8
60.9
65.9
61.3
44.3
40.5
41.9
9.4
22.9
25.4
14.9
33.7
30.1
17.6
23.2
20.6
20.6
25.6
14.9
0.2
0.8
0.3
0.8
1.1
0.1
0.6
0.3
0.3
0.3
1.4
0.7
63.2
0.1
0.1
0.2
0.1
0.3
1.0
0.3
2.3
1.0
0.1
1.4
12.0
6.8
10.2
6.5
21.4
13.7
19.8
10.3
15.5
33.8
32.4
41.0
c
d
c
0.1%Pd/Zn–Cr
e
1
1
1
0
1
2
1%Pd/Zn–Cr
18,000
10,200
10,200
f
1%Pd/Zn–Cr
f,g
1%Pd/Zn–Cr
a
◦
3
Reaction conditions: 300 C, acetone/H = 30:70 (vol.), 0.1–0.2 g catalyst powder of ꢀ180 µm particle size and 1.54 g/cm density, 2 h time on stream.
2
b
C
3
is propene and propane, MO mesityl oxide, others are mainly C acetone condensation products (mostly 2,6,8-trimethylnonane-4-one) and ca. 1% iso-
9+
propanol.
c
Gas feed diluted with N ; acetone/H /N = 30:35:35 (vol.).
2
2
2
d
Catalyst particle size ꢀ90 µm.
e
f
Acetone/H = 50:50 (vol.).
2
Acetone/H = 9:91 (vol.).
2
g
Gas feed diluted with N ; acetone/H /N = 9:23:68 (vol.).
2
2
2

290
E.F. Kozhevnikova, I.V. Kozhevnikov / Journal of Catalysis 238 (2006) 286–292
−
1
Fig. 3. Acetone conversion and product selectivities vs. time on stream (0.2 g 1%Pd/Zn–Cr, 10 ml/min H flow, acetone/H = 30:70 (vol.), GHSV = 6500 h
,
2
2
◦
00 C).
3
propane, mesityl oxide, and isopropanol. MIBK was obtained
with up to 78% selectivity and 93% total selectivity to MIBK
+
DIBK at 42% acetone conversion (39% yield) (Table 2, en-
try 4). This compares well with the best results reported to date
[
9,15,18,20]. The highest MIBK + DIBK yield of 51% was
obtained at 64% MIBK selectivity and 57.3% acetone conver-
sion (entry 3). These results were obtained at 300 C, GHSV of
6
◦
−1
500 h , 30 vol% acetone in the gas feed, and an acetone/H2
volume ratio of 0.5–1. At higher or lower acetone/H2 ratios,
lower conversions and/or MIBK selectivities were observed
(
entries 10–12).
The reaction clearly requires both acid–base and redox catal-
Fig. 4. Acetone conversion and selectivities to MIBK and DIBK vs. reaction
ysis, which are provided by Zn–Cr oxide and Pd, respectively.
When Zn–Cr oxide was used in the absence of Pd, mesityl oxide
temperature (0.2 g 1%Pd/Zn–Cr, 10 ml/min H2 flow, acetone/H2 = 30:70
−
1
(
vol.), GHSV = 6500 h , 2 h time on stream).
(
the product of acid-catalysed condensation of acetone) was the
major product, whereas the selectivity to MIBK was only 15%
entry 1). This shows that the Zn–Cr oxide itself acts mainly as
(
an acid catalyst, with weak hydrogenation activity. The conver-
sion of acetone increased with increasing Pd loading from 0.1%
to 1% (cf. entries 3, 7, and 9). This finding may suggest that the
Pd-catalysed hydrogenation of mesityl oxide is a slow step in
acetone conversion (Scheme 1); however, data on Pd dispersion
are needed to confirm this.
Using the crystalline Zn–Cr oxide (Table 1, entry 2) as a sup-
port resulted in a lower catalytic activity compared with that of
the amorphous Zn–Cr oxide (cf. entries 3 and 6, Table 2). This
may be explained by the lower surface area of the crystalline
oxide, which can affect both the acidity and the Pd dispersion
of the catalyst.
The Pd/Zn–Cr catalyst exhibited very good stability (Fig. 3).
No catalyst deactivation was observed during at least 10 h of
continuous operation and 50 h of discontinuous operation with
overnight breaks. The catalyst reached a steady state in ca. 1 h
and after that performed with constant activity and selectivity.
Carbonaceous deposits lay on the catalyst during the reaction
Fig. 5. Acetone conversion and selectivities to MIBK and DIBK vs. contact
time (0.1–0.5 g 1%Pd/Zn–Cr, 10–40 ml/min H2 flow, acetone/H2 = 30:70
◦
(
vol.), 300 C, 2 h time on stream).
◦
at 350 C (Fig. 4). With increasing reaction temperature, MIBK
selectivity increased slightly and DIBK selectivity decreased
considerably. Fig. 5 shows the effect of contact time (the reverse
of GHSV) on acetone conversion and selectivity to MIBK and
DIBK. Note that the reactor was operating under nondifferential
conditions. The conversion increased with increasing contact
◦
(
4.6 wt% in 10 h at 300 C) but did not affect catalyst perfor-
mance.
The reaction temperature caused only a slight effect on ace-
◦
tone conversion, which increased from 29% at 200 C to 48%

E.F. Kozhevnikova, I.V. Kozhevnikov / Journal of Catalysis 238 (2006) 286–292
291
Table 3
Liquid-phase synthesis of methyl isobutyl ketone over Pd/Zn–Cr catalysts
a
b
Entry
Catalyst
T
p (H )
(bar)
Conversion
(%)
Selectivity (%)
2
◦
(
C)
MIBK
DIBK
C
IP
MO
C
3
9+
1
2
3
4
5
6
7
8
9
1%Pd/Zn–Cr
1%Pd/Zn–Cr
1%Pd/Zn–Cr
1%Pd/Zn–Cr(cr)
1%Pd/Zn–Cr
1%Pd/Zn–Cr
200
200
200
200
160
160
160
200
200
160
200
200
20
7.5
5
7.5
20
7.5
5
7.5
5
5
83.0
53.6
55.6
52.8
48.5
28.5
26.7
52.4
49.5
33.3
53.8
46.3
22.3
70.6
83.1
57.2
26.8
80.4
78.7
77.1
76.0
72.1
76.5
79.0
2.9
13.4
10.2
7.2
1.0
4.6
1.0
0.7
1.3
1.1
1.4
1.1
2.6
0
0.4
0.6
0.7
0.6
64.8
9.1
1.1
27.0
68.5
11.2
13.1
2.5
8.1
1.6
0
3.0
1.9
0.4
0
0.4
1.8
0.9
0.2
0.4
0.9
4.6
4.3
4.4
0.4
2.3
2.2
4.5
7.7
2.1
5.7
3.7
c
1%Pd/Zn–Cr
3.4
0.3%Pd/Zn–Cr
0.3%Pd/Zn–Cr
0.3%Pd/Zn–Cr
0.1%Pd/Zn–Cr
0.1%Pd/Zn–Cr
15.5
13.5
4.5
15.4
13.1
0.6
19.8
1.4
1
1
1
0
1
2
7.5
5
3.2
a
b
c
Reactions in 45 ml-autoclave, 2.0 g acetone, 0.20 g catalyst powder of ꢀ180 µm particle size, 2 h.
C is propene and propane, IP isopropanol, MO mesityl oxide, C higher acetone condensation products.
3
9+
Reuse of this run gave 17.2% conversion, 58.1% MIBK, 1.2% DIBK, 40.1% IP and 0.6% C
.
9+
Similar to the gas-phase reaction, Pd supported on the crys-
talline Zn–Cr oxide was less active than the catalyst supported
on the amorphous oxide (cf. entries 2 and 4). Pd/Zn–Cr(cr) gave
less MIBK and DIBK but significantly more isopropanol.
The Pd/Zn–Cr catalyst could be reused simply after wash-
ing it with acetone, although with a reduced activity (entry 7).
The second run gave less MIBK and DIBK and much more
isopropanol, as compared to the first run. This shows that the
recovered catalyst partly lost its acidity but retained its hydro-
genation activity. Therefore, the regeneration of catalyst acidity
is required to achieve better catalyst performance on reuse.
Scheme 3.
time, as expected. The selectivity to MIBK and selectivity to
DIBK were practically independent of the contact time; only at
long contact times did MIBK selectivity decrease in favour of
DIBK, as expected. The conversion did not depend on the gas
flow rate at constant GHSV, but increased as the catalyst parti-
cle size decreased (cf. entries 3 and 5, Table 2). Unexpectedly,
the DIBK/MIBK ratio increased with decreasing catalyst par-
ticle size. This result, together with the effect of contact time
4
. Conclusion
The present study has shown that palladium metal (0.1–
(
Fig. 5), suggest that DIBK may be formed mainly in a paral-
lel reaction from MO and acetone (Scheme 3) rather than from
MIBK (Scheme 2).
1
wt%) supported on amorphous Zn–Cr (1:10) mixed oxide
is an efficient bifunctional catalyst for the one-step synthe-
sis of MIBK from acetone in both gas-phase and liquid-phase
processes. The catalyst is easy to prepare by coprecipitation of
Zn and Cr hydroxides, followed by calcination at 300 C and
palladium salt impregnation and reduction. Both the continu-
II
III
◦
3
.3. Liquid-phase synthesis of MIBK
◦
ous gas-phase process (300 C, ambient pressure) and the batch
liquid-phase process (200 C, 5 bar H pressure) yield MIBK
The reaction in the liquid phase was carried out at 160–
00 C and 5–20 bar H pressure in the presence of Pd/Zn–Cr
catalyst with 0.1–1 wt%Pd loading; the results are given in
Table 3. The same products were generally observed in the
liquid-phase reaction as in the gas-phase reaction, except with
◦
◦
2
2
2
together with DIBK with a total selectivity of 93% at 56–57%
acetone conversion, on a par with the best results reported to
date. In the gas-phase synthesis, no catalyst deactivation is ob-
served for at least 50 h on stream. In the liquid-phase reaction,
the catalyst performs well at a very low H2 pressure of 5 bar. It
can be reused, although regeneration of its acidity is required.
Further improvement may be achieved by catalyst and process
optimisation over a broader range of Zn/Cr atomic ratios, as
well as by better characterisation of the acid and metal sites in
the catalyst.
more isopropanol and less C9 products formed in the liquid
+
◦
phase. The reaction was the most efficient at 200 C and hydro-
gen pressure of 5–7.5 bar, almost independent of the Pd loading
entries 3, 8, 9, 11, and 12). The best run (entry 3) gave 83%
(
MIBK selectivity and 93% total MIBK + DIBK selectivity at
◦
5
6% acetone conversion (200 C, 5 bar H pressure). This re-
2
sult is on a par with the best results reported to date [10,18,19],
and the H2 pressure is unprecedentedly low (typically about
0 bar or higher [10,18,19]), which may be explained by the
References
2
high dispersion of Pd in the catalyst. Increasing the H pressure
to 20 bar led to the predominant formation of isopropanol by
hydrogenation of the C=O group of acetone (entries 1 and 5).
2
[
[
1] W.F. Hoelderich, Stud. Surf. Sci. Catal. 41 (1988) 83.
2] K. Weissermel, H.-J. Arpe, Industrial Organic Chemistry, 4h ed., Wiley–
VCH, 2003, p. 282.

292
E.F. Kozhevnikova, I.V. Kozhevnikov / Journal of Catalysis 238 (2006) 286–292
[
[
[
[
[
3] L. Melo, P. Magnoux, G. Giannetto, F. Alvarez, M. Guisnet, J. Mol. Catal.
A 124 (1997) 155.
4] R. Ruiz, C. Pesquera, F. González, C. Blanco, Appl. Catal. A 257 (2004)
[21] G.W. Bridger, M.S. Spencer, in: M.V. Twigg (Ed.), Catalyst Handbook,
2nd ed., Wolfe, London, England, 1989, p. 455.
[22] M.C.J. Bradford, M.V. Konduru, D.X. Fuentes, Fuel Process. Technol. 83
165.
(2003) 11.
5] L. Melo, A. Llanos, L. García, P. Magnoux, F. Alvarez, M. Guisnet, G.
Giannetto, Catal. Lett. 51 (1998) 207.
6] M. Martínez-Ortiz, D. Tichit, P. Gonzalez, B. Cog, J. Mol. Catal. A 201
[23] W.S. Epling, G.B. Hoflund, D.M. Minahan, Appl. Catal. A 183 (1999)
335.
[
[
[
24] F. Simard, U.A. Sedran, J. Sepúlveda, N.S. Fígoli, H.I. de Lasa, Appl.
Catal. A 125 (1995) 81.
(2003) 199.
25] D.W. Bonniface, J.R. Fryer, P. Landon, J.D. Scott, W.D.S. Scott, M.J. Wat-
son, G. Webb, J.M. Winfield, Green Chem. (1999) 9.
26] J.D. Scott, M.J. Watson, N.P. Ewing, WO 95/27688 (1995), assigned to
ICI.
7] Y. Diaz, L. Melo, M. Mediavilla, A. Albornoz, J.L. Brito, J. Mol. Catal.
A 227 (2005) 7.
[
[
8] L.V. Mattos, F.B. Noronha, J.L.F. Monteiro, J. Catal. 209 (2002) 166.
9] P.Y. Chen, S.J. Chu, C.C. Chen, N.S. Chang, W.C. Lin, T.K. Chuang, US
Patent 5059724 (1991).
[27] D. Gulkova, M. Kraus, J. Mol. Catal. 87 (1994) 47.
[28] T. Yokoyama, N. Yamagata, Appl. Catal. A 221 (2001) 227.
[29] B.M. Reddy, I. Ganesh, J. Mol. Catal. A 151 (2000) 289.
[30] J. Sloczynski, J. Janas, T. Machej, J. Rynkowski, J. Stoch, Appl. Catal.
B 24 (2000) 45.
[
[
[
[
10] S.-M. Yang, Y.M. Wu, Appl. Catal. A 192 (2000) 211.
11] N.N. Das, S.C. Srivastava, Bull. Mater. Sci. 25 (2002) 283.
12] Y.Z. Chen, C.M. Hwang, C.W. Liaw, Appl. Catal. A 169 (1998) 207.
13] N. Das, D. Tichit, R. Durand, P. Graffin, B. Coq, Catal. Lett. 71 (2001)
[
[
31] M. Ohta, Y. Ikeda, A. Igarashi, Appl. Catal. A 258 (2004) 153.
32] I.V. Kozhevnikov, S. Holmes, M.R.H. Siddiqui, Appl. Catal. A 214 (2001)
181.
[
[
[
[
[
[
[
14] L.M. Gandía, M. Montes, Appl. Catal. A 101 (1993) L1.
15] V. Chikán, Á. Molnár, K. Baláznik, J. Catal. 184 (1999) 134.
16] K.H. Lin, A.N. Ko, Appl. Catal. A 147 (1996) L259.
47.
[
[
33] E.F. Kozhevnikova, I.V. Kozhevnikov, J. Catal. 224 (2004) 164.
34] M. Brändle, J. Sauer, J. Am. Chem. Soc. 120 (1998) 1556.
17] Y. Watanabe, Y. Matsumura, Y. Izumi, Y. Mizutani, J. Catal. 40 (1975) 76.
18] Y. Higashio, T. Nakayama, Catal. Today 28 (1996) 127.
19] Y.Z. Chen, B.J. Liaw, H.R. Tan, K.L. Shen, Appl. Catal. A 205 (2001) 61.
20] S. Narayanan, R. Unnikrishnan, Appl. Catal. A 145 (1996) 231.
[35] V. Ponec, G.C. Bond, in: B. Delmon, J.T. Yates (Eds.), Catalysis by Metals
and Alloys, Stud. Surf. Sci. Catal., vol. 95, Elsevier, Amsterdam, 1995.
[36] C.A. Müller, M. Maciejewski, R.A. Koeppel, A. Baiker, J. Catal. 166
(1997) 36.