Novel transition-metal-free heterogeneous epoxidation catalysts discovered by means of high-throughput experimentation

Article abstract of DOI:10.1002/chem.200700074

Various transition-metal-free oxides have been studied as catalysts for the epoxidation of cyclooctene with hydrogen peroxide by means of high-throughput experimentation. Different boron, aluminium, and gallium oxides were prepared according to various synthesis methods. A number of pure aluminium and gallium oxides showed very good catalytic performances, while the results obtained with boron oxides or mixed oxides were less positive. The best results were obtained with a gallium oxide catalyst, which gave an epoxide yield of 71 % and a selectivity of 99% after reaction for 4 h at 80°C. Gallium oxides had not been reported previously as active epoxidation catalysts. The use of high-throughput experimentation proved useful both for discovering new active catalysts and for identifying a number of relationships between the synthesis conditions and the catalytic properties of the transition-metal-free oxides.

Full text of DOI:10.1002/chem.200700074

DOI: 10.1002/chem.200700074
Novel Transition-Metal-Free Heterogeneous Epoxidation Catalysts
Discovered by Means of High-Throughput Experimentation
Paolo P. Pescarmona,* Kris P. F. Janssen, and Pierre A. Jacobs[a]
Abstract: Various transition-metal-free
oxides have been studied as catalysts
for the epoxidation of cyclooctene with
hydrogen peroxide by means of high-
throughput experimentation. Different
boron, aluminium, and gallium oxides
were prepared according to various
synthesis methods. A number of pure
aluminium and gallium oxides showed
very good catalytic performances, while
the results obtained with boron oxides
or mixed oxides were less positive. The
best results were obtained with a galli-
um oxide catalyst, which gave an epox-
ide yield of 71% and a selectivity of
99% after reaction for 4 h at 808C.
Gallium oxides had not been reported
previously as active epoxidation cata-
lysts. The use of high-throughput ex-
perimentation proved useful both for
discovering new active catalysts and for
identifying a number of relationships
between the synthesis conditions and
the catalytic properties of the transi-
tion-metal-free oxides.
Keywords: epoxidation · heteroge-
neous catalysis · high-throughput
experimentation
·
sustainable
chemistry
oxides
·
transition-metal-free
Introduction
Epoxides are versatile com-
pounds used in the synthesis of
many fine chemicals with rele-
vant industrial applications.[1]
Scheme 1. Epoxidation of alkenes.
The most common way to pre-
pare epoxides is by the catalyt-
ic oxidation of alkenes (Scheme 1).
A number of good homogeneous and heterogeneous cata-
lysts for the epoxidation of alkenes has been reported.[2–4]
However, none of them meets all of the requirements for ef-
ficient and sustainable chemistry. An ideal epoxidation cata-
lyst should exhibit the following features:
the reactive oxygen atom of H2O2 to the substrate should
be catalysed in an efficient way, avoiding the formation
of O2.
It should be a stable heterogeneous catalyst, to allow
easy separation from the products and recycling without
loss of activity.
*
*
*
It should not release transition metals into the reaction
medium, or preferably should not contain them at all.
It should show optimum performance in environmentally
acceptable solvents or in the absence of solvents.
*
It should be active under mild reaction conditions for a
wide range of substrates. It should also be selective, im-
plying that it should not contain too strongly acidic sites
that might cause further reaction of the formed epoxides.
*
It should be usable in conjunction with hydrogen perox-
Recently, it was discovered that partially hydroxylated
ide (H2O2)as the oxidant, since this readily available
compound yields water as a by-product. The transfer of
aluminium oxides are active and selective heterogeneous
catalysts for the epoxidation of a variety of alkenes with hy-
drogen peroxide.[5–7] The proposed catalytic cycle involves
ꢀ
the formation of surface hydroperoxide groups (Al OOH)
[a] Dr. P. P. Pescarmona, K. P. F. Janssen, Prof. Dr. P. A. Jacobs
Centre for Surface Chemistry and Catalysis, K. U. Leuven
Kasteelpark Arenberg 23, 3001 Heverlee (Belgium)
Fax : (+32)16-321-998
ꢀ
by the reaction of surface hydroxyl functions (Al OH)with
hydrogen peroxide. The activities and stabilities of the alu-
minas strongly depend on the synthesis method used to pre-
pare them.[8,9] The best catalytic performances are obtained
E-mail: paolo.pescarmona@biw.kuleuven.be
6562
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 6562 – 6572

FULL PAPER
with aluminium oxides presenting a high number of acid
sites combined with a moderate hydrophilicity, ensuring a
good interaction with both the polar H2O2 and the apolar
alkene.[10] Due to their reasonable stability in H2O, these
catalysts are active in epoxidation reactions with concentrat-
ed aqueous hydrogen peroxide, which represents the most
suitable oxidant on environmental and economic grounds.
Despite the aforementioned advantages, the catalytic pa-
rameters of these transition-metal-free oxides do not signifi-
cantly exceed those of the known heterogeneous catalyst
systems containing transition metals.[2]
tively with the large number of results that will be produced.
The choice of the HTE methodology depends on the fea-
tures of the system under study, on the kind of information
that is sought, and on the experimental equipment available.
The first step in an HTE experiment is the definition of
the parameter space to be studied. This implies the use of
personal knowledge and intuition, literature data, and com-
putational modelling to identify the parameters that can be
expected to have an influence on the generation of a system
with the desired properties. The next step is to find the opti-
mal strategy to study the selected parameter space. In a full
factorial approach, all possible combinations of the parame-
ters are studied. This strategy can be used if the number of
experiments needed to investigate the entire selected pa-
rameter space is not too large compared with the number of
experiments that may be performed with the available
equipment. If the knowledge of the system under study is
limited or if the system is very complex, the parameter
space to be explored becomes too large for most HTE
equipment. To overcome this problem, a variety of strat-
egies for HTE experimental design have been developed in
recent years.[16–18] Mathematical and statistical concepts are
used to build search algorithms that allow a reduction in the
number of experiments required to gain the desired infor-
mation about the chosen parameter space. The selection of
the experiments to be performed and the handling of the re-
sults become crucial and complex processes, which require
the use of computational tools. Generally, the reliability of a
search algorithm should be tested on a known system before
applying it to a new one.[19]
The full factorial is the most suitable and flexible ap-
proach for focussed studies: in addition to the discovery of
leads, it provides a clear correlation between parameters
and properties, allowing the identification of trends and be-
haviours. The possibility of observing trends also minimises
the risk of missing a lead in the event of experimental errors
or technical problems. Compared to search algorithms, the
use of a full factorial approach is more straightforward and
allows intuitive and rational design of the experiments as
well as analysis of the results. On the other hand, search al-
gorithms permit the investigation of much larger parameter
spaces and prevent the risk of incorrect or biased assump-
tions from the researcher by using computational techniques
to select the experiments to be performed. Both a full facto-
rial approach and search algorithms have found application
in HTE studies of the catalytic epoxidation of alkenes.[20,21]
Herein, we present a study of a new set of transition-
metal-free catalysts for the epoxidation of alkenes with
H2O2. It is anticipated that oxides of Group 13 elements
ꢀ
with the capacity to form surface peroxide groups (B OOH,
ꢀ
ꢀ
Al OOH, Ga OOH)would be capable of serving as epoxi-
dation catalysts. The catalytic performance of these oxides
will be affected by the stability of the surface peroxide
groups formed, which, in turn, depends on the nature of the
metal1 at the oxide surface and that of their close neigh-
bours. The stability of the surface peroxide groups should
not be too low, so as to avoid decomposition with O2 gener-
ation, nor too high, which would be unfavourable for the
transfer of the oxygen to the alkene. Finding the optimum
balance is not trivial. On the basis of these considerations, it
was decided to screen the composition triangle of three-
component oxide mixtures (B, Al, and Ga oxides)with the
aim of discovering improved transition-metal-free epoxida-
tion catalysts. These materials can be prepared according to
different procedures in order to tune their porosity and
structure, as well as the strength and distribution of acid
sites, to the desired catalytic behaviour.
High-throughput experimentation (HTE)techniques have
been used to investigate the large parameter space defined
by the various compositions of mixed B, Al and Ga oxides
prepared by means of different synthesis methods. Combina-
torial chemistry and high-throughput experimentation tech-
niques are among the newest and most promising experi-
mental methods in many fields of chemical research, includ-
ing materials science and catalysis.[11–15] These methods are
based on the investigation of broad sets of experimental pa-
rameters through the design, synthesis, and screening of
large numbers of samples. The high throughput of these
techniques is made possible by the use of automated work-
stations and fast-analysis equipment for the preparation and
screening of the samples. The mixed oxides were synthesised
and their catalytic properties were tested in parallel on an
automated workstation equipped with a 60-well stirring and
heating block. The epoxidation activities and selectivities
were screened by means of a GC set-up equipped with a
very fast heating and cooling system, which allowed rapid
analysis of the samples.
Results and Discussion
The composition diagram of boron, aluminium and gallium
mixed oxides can be described using ternary arrays with a
different number of points.[18] The higher the number of in-
terval steps into which the array is divided, the more accu-
rate the description of the three-component system. A ter-
nary array with nine steps would offer a good coverage of
the composition diagram and would require 45 sample
When studying a system by means of HTE, an appropri-
ate experimental strategy should be selected to deal effec-
1
While aluminium and gallium are metals, boron is a semi-metal. The
term “metals” is used in this article when referring to these three ele-
ments together.
Chem. Eur. J. 2007, 13, 6562 – 6572
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
6563

P. P. Pescarmona et al.
points. This number of points is compatible with the
throughput of the available HTE equipment, which allows
performing up to 60 simultaneous parallel experiments. This
means that a full factorial approach can be chosen to study
the system.
The composition triangle formed by B, Al and Ga mixed
oxides was screened for activity in the epoxidation of cis-cy-
clooctene[10] with a 50 wt% aqueous solution of H2O2 after
reaction for 4 h at 808C. Three different synthesis methods
were used to prepare the oxides; these were derived from
the sol–gel procedures used to synthesise the aluminas with
the highest epoxidation activity[10] or procedures used to
prepare mixed Al and Ga oxides,[22] and were adapted so
that they could be performed on the HTE workstation.
The epoxidation activities of the three sets of 45 mixed-
oxide catalysts are reported in Figure 1. A number of obser-
vations can be made concerning these experiments:
*
The catalytic properties of the oxides vary as a function
of the method used for their synthesis: the epoxidation
activity is proportional to the gallium content with
method 1 (Figure 1a)and proportional to the aluminium
content with method 3 (Figure 1c), while two zones of
higher activity for a high gallium and high aluminium
content are found with method 2 (Figure 1b). The influ-
ence of the method of synthesis is relevant to such a
degree that pure aluminium oxide or pure gallium oxide
can show alternatively the highest and almost the lowest
activity according to the preparation route followed
(compare plots (a)and (c)in Figure 1).
*
Besides the good activity found for aluminium oxide, in
agreement with literature data,[5–10] gallium oxide was
identified as a new, promising epoxidation catalyst. This
result can be explained by considering that gallium
oxides have physicochemical properties similar to those
of aluminium oxides and present a similar polymorph-
ism.[23]
*
Despite the good epoxidation activities shown by pure
gallium and pure aluminium oxides, the activities of the
mixed oxides of Al and Ga decrease as their composition
moves away from that of the active pure oxides. This
lack of synergy was observed with all of the synthesis
methods employed and can be ascribed to the structural
and textural features of the mixed oxides rather than to
their chemical properties. Low or moderate activities
were found for most of the tested ternary oxides. The
best synergies were found for mixed oxides with very low
or no Al content and with Ga/B molar ratios larger than
1. These catalysts showed moderate to good epoxidation
activities when synthesised according to methods 1 and 2.
However, the selectivities towards the epoxide of the
most active of these catalysts were only around 50–60%,
as compared to values higher than 90% for the pure gal-
lium oxide and pure aluminium oxide catalysts.
Figure 1. Activity in the epoxidation of cyclooctene with 50 wt% aqueous
H2O2, after 4 h at 808C, of ternary arrays of B, Al, and Ga mixed oxides
prepared according to different synthesis methods [a)method 1;
b)method 2; c)method 3].
*
Catalysts with high boron content show very low epoxi-
dation activities, regardless of the synthesis method em-
ployed.
6564
www.chemeurj.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 6562 – 6572

HTE Screening of Epoxidation Catalysts
FULL PAPER
Besides the identification of gallium oxide as a new lead,
this series of experiments indicated the decisive influence of
the synthesis method on the final performance of the cata-
lysts. The three synthetic methods employed differ from
each other in the concentrations of the metal precursors in
the reaction mixture, in the ratio between water and solvent,
and in the pH of the medium (see the Experimental Sec-
tion). These parameters are known to affect the hydrolytic
condensation process that takes place in sol–gel syntheses.[24]
On the basis of these considerations, it was decided to study
in more detail the effects on the epoxidation activity of var-
iations in the concentration of the metal precursors and in
the amount of basic, neutral, or acid aqueous solution em-
ployed during the synthesis of aluminium and gallium
oxides, that is, the two compositions that led to the best cat-
alytic results in the first set of HTE tests.
For each metal oxide, four different concentrations of
metal precursor and two different amounts of a basic, neu-
tral, or acidic aqueous solution were selected. The combina-
tion of all these parameters defined a parameter space of 48
elements, which could again be studied with the available
HTE equipment using a full factorial approach. The activi-
ties of the catalysts in the epoxidation of cyclooctene with
H2O2 as a function of the variation of the above-mentioned
synthesis parameters are reported in Figure 2. A more com-
plete overview of the catalytic results, including the cyclooc-
tene conversions and the selectivities towards epoxidation,
can be found in Table 1 (first catalytic test).
The results confirm the relevance of the effects of the
varied synthesis parameters on the epoxidation activities of
the catalysts. Some leads can be identified and some general
behaviour can be deduced from the analysis of these HTE
data. The effects of the screened synthesis parameters on
the aluminium oxides are quite different to those on the gal-
lium oxides. Therefore, the results concerning each group of
oxides will be discussed separately.
Aluminium oxides: The most active aluminas were obtained
with small amounts of hydrolysing solutions (0.320 mL NH3
in water/ethanol, 0.196 mL H2O, 0.196 mL aqueous oxalic
acid). The catalytic performances of these materials were
not noticeably affected by variations in the concentration of
the aluminium precursor. The activities of aluminium oxides
synthesised with large amounts of hydrolysing solutions
(1.600 mL NH3 in water/ethanol, 0.980 mL H2O, 0.980 mL
aqueous oxalic acid)were found to be lower when a high
water/solvent ratio was employed in their preparation. All
syntheses involving the use of 0.980 mL of aqueous oxalic
acid led to aluminas with low activities. These results indi-
cate that too high a water concentration and too acidic an
environment are detrimental to the formation of active alu-
minium oxides. A high water concentration in the reaction
mixture increases the hydrolysis rate while retarding con-
densation reactions. Moreover, high acid to aluminium
ratios in solution have been reported to hinder condensation
reactions.[24,25] On the other hand, a basic or moderately
Figure 2. Activity in the epoxidation of cyclooctene with 50 wt% aqueous H2O2 after 4 h at 808C as a function of the concentration of the metal precur-
sors and of the amount and type of hydrolysing solution employed during the synthesis of the aluminium oxide and gallium oxide catalysts [first catalytic
test].
Chem. Eur. J. 2007, 13, 6562 – 6572
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
6565

P. P. Pescarmona et al.
Table 1. Catalytic epoxidation of cyclooctene with 50 wt% aqueous H2O2 after 4 h at 808C: conversions, yields and selectivities against the concentration
of the metal precursors and the amount and type of hydrolysing solution used during the synthesis of the aluminium oxide and gallium oxide catalysts.
Synthesis conditions
First catalytic test
Second catalytic test
[a]
[a]
Al
A
GaCl3
Amount of Amount and
2-butanol type of
Cyclo-
octene
Epoxide By-products Epoxide Cyclo- Epoxide By-products Epoxide
yield [%] yield [%]
sel.
octene yield [%] yield
sel.
[mL]
aqueous
solution
conv. [%]
[%]
conv.
[%]
[%]
[%]
1
2
3
4
5
6
7
8
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
0.400
0.400
0.400
0.400
0.400
0.400
1.000
1.000
1.000
1.000
1.000
1.000
2.500
2.500
2.500
2.500
2.500
2.500
5.000
5.000
5.000
5.000
5.000
5.000
0.400
0.400
0.400
0.400
0.400
0.400
1.000
1.000
1.000
1.000
1.000
1.000
2.500
2.500
2.500
2.500
2.500
2.500
5.000
5.000
5.000
5.000
5.000
5.000
0.320 mL NH3 sol. 43
1.600 mL NH3 sol. 13
40
13
35
15
33
1
40
26
36
11
29
2
33
33
36
29
28
12
32
35
31
29
33
7
3
2
2
1
1
9
2
2
35
48
32
5
3
3
46
44
51
22
5
3
1
3
2
3
1
3
3
4
1
5
2
3
3
3
3
3
3
4
4
3
3
4
2
1
1
1
0.4
1
1
2
1
93
95
92
88
92
58
93
91
91
90
86
55
91
92
92
92
89
78
90
90
90
90
89
82
70
75
73
74
67
93
57
62
77
70
80
78
57
66
64
64
80
77
69
64
55
62
65
56
45
18
34
17
32
2
36
31
39
11
26
2
38
37
36
33
30
14
40
32
35
32
31
8
6
13
2
2
2
8
5
3
27
61
27
4
5
9
68
72
56
36
7
42
16
32
15
30
1
34
29
35
10
23
2
35
34
33
30
27
11
38
29
33
30
28
7
5
13
2
1
2
8
4
3
26
60
27
4
4
9
66
71
55
35
5
4
2
2
2
2
0.3
2
2
5
1
3
1
2
2
2
3
3
3
3
2
2
2
3
2
0.2
0.2
1
0.4
1
92
89
93
89
93
79
95
93
88
91
89
64
94
93
94
92
91
78
93
92
94
93
91
80
97
98
70
75
73
93
98
92
98
98
99
91
94
97
98
99
99
97
81
85
97
98
98
97
0.196 mL H2O
0.980 mL H2O
38
17
0.196 mL HOx sol. 36
0.980 mL HOx sol.
2
0.320 mL NH3 sol. 43
1.600 mL NH3 sol. 29
9
0.196 mL H2O
0.980 mL H2O
40
12
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
0.196 mL HOx sol. 34
0.980 mL HOx sol.
4
0.320 mL NH3 sol. 36
1.600 mL NH3 sol. 36
0.196 mL H2O
0.980 mL H2O
39
32
0.196 mL HOx sol. 32
0.980 mL HOx sol. 15
0.320 mL NH3 sol. 36
1.600 mL NH3 sol. 40
0.196 mL H2O
0.980 mL H2O
34
33
0.196 mL HOx sol. 37
0.980 mL HOx sol.
0.320 mL NH3 sol.
1.600 mL NH3 sol.
0.196 mL H2O
9
5
3
2
0.980 mL H2O
2
0.196 mL HOx sol.
0.980 mL HOx sol.
0.320 mL NH3 sol.
1.600 mL NH3 sol.
0.196 mL H2O
2
9
6
3
45
72
1
0.1
0.2
1
11
21
8
1
2
0.980 mL H2O
1
0.196 mL HOx sol. 41
0.2
0.4
0.3
0.2
1
1
1
1
1
0.5
2
1
1
1
0.980 mL HOx sol.
0.320 mL NH3 sol.
1.600 mL NH3 sol.
0.196 mL H2O
6
5
5
76
71
2
26
25
13
7
2
2
32
28
27
12
0.980 mL H2O
0.196 mL HOx sol. 67
0.980 mL HOx sol. 29
0.320 mL NH3 sol.
1.600 mL NH3 sol.
0.196 mL H2O
8
5
74
78
3
3
3
39
46
49
15
71
66
69
29
69
65
68
28
0.980 mL H2O
0.196 mL HOx sol. 76
0.980 mL HOx sol. 27
[a] Molar fraction.
acidic environment catalyses the hydrolytic condensation of
aluminium alkoxides.[24] These considerations suggest that
the condensation process is crucial for the formation of alu-
minium oxide catalysts with good epoxidation activity. If the
hydrolysis is favoured over the condensation, the aluminium
oxides will have highly hydroxylated surfaces: the degree of
hydrophilicity would then be too high, preventing the al-
kenes from approaching the catalytic sites on the surface.
This hypothesis is in agreement with data reported in the lit-
erature, which showed that aluminas synthesised with higher
acid to aluminium ratios were the most hydrophilic and the
least active.[10]
The best catalytic results were obtained with aluminium
oxides 1 and 7 (Table 1, first catalytic test), synthesised from
concentrated solutions of aluminium sec-butoxide in 2-buta-
nol, hydrolysed with a solution of ammonia in water and
ethanol, with a low water/solvent ratio. The cyclooctene con-
version (43%)and epoxide yield (40%)obtained with these
6566
www.chemeurj.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 6562 – 6572

HTE Screening of Epoxidation Catalysts
FULL PAPER
catalysts were significantly higher than those obtained with
the aluminium oxide 5, prepared according to a similar
method to that employed for the synthesis of the best alumi-
na epoxidation catalyst reported in the literature.[8,10] The se-
lectivities in favour of the formation of the epoxycyclooc-
tane were 90% or higher for the most active of the alumini-
um oxides. The main by-product was 2-cycloocten-1-one, ac-
companied by minor amounts of 2-cycloocten-1-ol.
gallium oxides discussed above contained chlorine atoms
that reacted during the first catalytic test. This change might
generate a different catalytic performance of the recycled
material. Therefore, it was of particular interest to re-evalu-
ate the epoxidation activities of the 48 catalysts reported in
Table 1 after washing them with ethanol and drying at
1108C. The epoxide yields obtained in this second catalytic
run are reported in Figure 3, while a more complete over-
view of the catalytic results is presented in Table 1.
Gallium oxides: Catalysts with high epoxidation activities
were synthesised by the hydrolytic condensation of less con-
centrated solutions of GaCl3 with water or with small
amounts of aqueous solutions of oxalic acid (0.196 mL).
Since the hydrolysis of GaCl3 produces HCl, all of these syn-
theses take place in an acidic environment. Gallium oxides
prepared with a high concentration of precursor or under
basic conditions show very low activities. Similarly to the
observations regarding the aluminium oxides, too high an
acid/metal ratio (0.980 mL aqueous oxalic acid)is detrimen-
tal to the activity of the catalysts. However, the effect is less
pronounced with gallium oxides, which still retain a good ac-
tivity at low water/solvent ratios. These results suggest that a
moderate acidity is the most suitable condition for the syn-
thesis of active gallium oxide catalysts, probably by provid-
ing a good balance between hydrolysis and condensation re-
actions.
The best among the gallium oxide catalysts are in adja-
cent positions in the screened parameter space (Figure 2),
indicating the range and limits of suitable synthesis condi-
tions for the preparation of active gallium oxides. These cat-
alysts showed cyclooctene conversions higher than 70% and
epoxide yields of around 50% (Table 1, first catalytic test).
These catalytic results represent a significant improvement
compared to those obtained with the best aluminium oxides.
Although the epoxidation activities of these catalysts are
very promising, the epoxide selectivities are only between
55 and 80% (Table 1, first catalytic test), which are marked-
ly lower values than those obtained with the aluminium
oxides. The lower selectivity observed with the gallium
oxides can be explained by considering the nature of the by-
products formed during the epoxidation reaction. For the
most active catalysts, the three main by-products are 1,2-cy-
clooctanediol, 3-chlorocyclooctene, and 5-chlorocyclooctene,
with smaller amounts of other chlorinated products as well
as of 2-cycloocten-1-one. The formation of 1,2-cyclooctane-
diol indicates the presence of strongly acidic species that
cause hydrolytic ring-opening of the epoxide (Scheme 1).
The presence of chlorinated products reveals that the cata-
lysts still contain Cl atoms from the gallium source, GaCl3.
These Cl atoms at the surface may react directly with cyclo-
octene to produce the chlorinated species or may be hydro-
lysed when the aqueous H2O2 is added to yield HCl, which
then catalyses the hydrolytic ring-opening of the formed ep-
oxides.
All of the aluminium oxides showed similar catalytic per-
formances in the first and the second runs, with a slight de-
crease in cyclooctene conversion and a slight increase in ep-
oxide selectivity. On the other hand, the catalytic perform-
ances of the gallium oxides varied significantly between the
first and second runs. The conversions of cyclooctene
showed, on average, a slight decrease, while the epoxide
yields and, consequently, the selectivities increased consider-
ably. Sample 40 gave the best results, with an epoxide yield
of 71% and a selectivity of 99% after reaction for 4 h at
808C (Table 1, second catalytic test): this represents a major
improvement compared to the known alumina epoxidation
catalysts.[10] The increased epoxide yields and selectivities of
various recycled gallium oxides can be attributed to the ab-
sence of chlorine species in the recycled catalysts, as con-
firmed by the much reduced presence of 1,2-cyclooctanediol,
3-chlorocyclooctene, 5-chlorocyclooctene or other chlorinat-
ed compounds among the by-products. This indicates that
the chlorine species present in the as-synthesised gallium
oxides were almost completely removed during the first cat-
alytic run and the subsequent washing procedure.
The HTE results presented so far have been based on a
screening of the epoxidation activities of B, Al and Ga
oxides as a function of their compositions and the conditions
of their synthesis. The studied parameter spaces were select-
ed as those expected to have the most marked influences on
the catalytic properties of the oxides. The identification of
new formulations for active catalysts has proved that these
choices were consistent. However, the screening of other pa-
rameters not hitherto explored can prove useful for the opti-
misation of known catalyst formulations and opens the pos-
sibility of the serendipitous discovery of less expected leads.
Among the parameters influencing the hydrolytic condensa-
tion leading to the formation of B, Al and Ga oxides, the
nature of the anion of the metal precursor and the nature of
the solvent in which the reaction takes place had not yet
been investigated in this study. These two parameters are in-
terrelated, since each metal precursor displays different de-
grees of solubility in different solvents. The best performing
alumina catalysts reported in the literature have been pre-
pared using 2-butanol as the solvent.[10] Therefore, 2-butanol
was chosen as solvent for the preparation of all of the
above-mentioned HTE samples. For this further study, a
number of metal precursors of the type MX3 were selected
(where M is B, Al, or Ga and X is chloride, nitrate, or an
alkoxy group). The group X reacts during the hydrolysis
step of the synthesis and, therefore, influences the reaction
rate. Most of the chosen metal precursors are soluble in eth-
An important feature of a good catalyst is its stability
under the reaction conditions, which would imply that its ac-
tivity is retained when it is reused in successive runs. The
Chem. Eur. J. 2007, 13, 6562 – 6572
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
6567

P. P. Pescarmona et al.
Figure 3. Activity in the epoxidation of cyclooctene with 50 wt% aqueous H2O2 after 4 h at 808C of the recycled aluminium and gallium oxides, as a
function of the concentration of the metal precursors and of the amount and type of hydrolysing solution employed during their synthesis [second cata-
lytic test].
anol, which was thus used as the solvent. Aluminium iso-
propoxide and aluminium sec-butoxide are only sparingly
soluble in ethanol; consequently, 2-propanol and 2-butanol,
respectively, were used as solvents for these two precursors
(Table 2). Although boron-containing oxides and mixed alu-
minium and gallium oxides generally gave poor results in
the previous HTE tests, it was decided to check whether
using different precursors might bring about an improve-
ment. Therefore, the selected metal precursors were com-
bined in reduced-size ternary arrays (Figure 4). The other
synthesis parameters were chosen to be as in the case of
samples 15 and 39 (Table 1), that is, a set of conditions that
led to very active catalysts for both aluminium and gallium
oxides. A parameter space of 44 elements was generated.
Each combination was tested for its activity in the epoxida-
tion of cyclooctene with H2O2 (Table 2). Pure gallium oxides
gave very poor results, in striking contrast with the very
high activities found before. Sample 92 was prepared under
the same conditions as sample 39, the only difference being
the use of ethanol instead of 2-butanol as solvent. The much
better results obtained with 2-butanol instead of ethanol can
be ascribed to the lower miscibility of 2-butanol with water:
this can prevent too high a hydrolysis rate during the forma-
tion of the oxides, which has been proposed to be detrimen-
tal to the formation of active epoxidation catalysts (vide
supra). Similarly, the effect of the interaction between water
and the solvent or the group X can be used to explain the
results obtained with the pure aluminium oxides: the highest
and most satisfactory activities were found for the pure alu-
minium oxides 85 and 88 (equivalent to sample 15), which
were synthesised from Al
ACHTRE(UNG iPrO)3 in 2-propanol and Al-
AHCTREUNG
periments, most of the mixed oxides showed low activities.
However, some effect of the nature of the precursor could
be observed. For instance, 1:1 boron to aluminium or 1:1
boron to gallium mixed oxides showed higher activities
when AlCl3 and GaCl3 were used as precursors, probably
owing to the fact that the chlorides rapidly hydrolyse to gen-
erate an acidic environment, which might favour the interac-
tion with the boron species. Among these catalysts, the 1:1
boron to gallium mixed oxide 70 proved to be the most
active mixed oxide identified in this work, with an epoxide
yield of 32% and a selectivity of 84%. The best combina-
tion of precursors was found to be different for each molar
ratio of boron, aluminium, and gallium: while aluminium
and gallium chlorides proved to be the best precursors for
1:1 mixed oxides with boron, the only significant activities
for an equimolar combination of boron, aluminium, and gal-
lium were obtained with aluminium sec-butoxide and galli-
um nitrate as precursors (samples 62 and 77).
The HTE study presented here led to the discovery of
various promising epoxidation catalysts. Ideally, the results
should have general validity and not be dependent on the
equipment employed to obtain them.[26] However, the data
generated using HTE might be affected by the small volume
scale at which the experiments are performed and by techni-
6568
www.chemeurj.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 6562 – 6572

HTE Screening of Epoxidation Catalysts
FULL PAPER
Table 2. Catalytic epoxidation of cyclooctene with 50 wt% aqueous H2O2 after 4 h at 808C: conversions, yields and selectivities against the composition
of the B, Al and Ga mixed-oxide catalysts and of the precursors and solvents used for their synthesis.
[a]
[a]
[a]
[a]
[a]
[a]
[a]
[a]
B
G
B
U
Al
(NO3)3
AlCl3
Al
in
U
Al
(sBuO)3
Ga
U
GaCl3
in
Cyclooctene Epoxide By-prod-
Epoxide
selectivity
in 2-butanol in ethanol
conversion
yield
[%]
ucts
ethanol 2-propanol
ethanol [%]
yield [%] [%]
49
50 0.5
51 0.5
52 0.5
53 0.5
54 0.5
55 0.5
56 0.333
57 0.333
58 0.333
59 0.333
60 0.333
61 0.333
62 0.333
63 0.333
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
0
0.5
0
0
0
0
0
0.333
0.333
0
0
0
0
0
0
0
0.5
0
0
0
0
0
0.333
0.333
0
0
0
0
0
0
1
0.5
0.5
0
0
0
0
0
0
0
0
0
0
0
0
0
0.5
0
0
0
0
0
0
0
0
0.5
0
0
0
0
0
0
0
0
0.5
0
0
0
0
0
0
0
0
0
0
0
0
0.5
0
0.333
0
0.333
0
0.333
0
0.333
0
0
0
0
0
0
0.5
0
0.333
0
0.333
0
0.333
0
0.333
0
0
0
0
0
0
0
0
0.5
0
0.333
0
0.333
0
0.333
0
0.333
0
0
0
0
0
0
0.5
0
0.333
0
0.333
0
0.333
0
0.333
0
0
3
2
10
4
5
3
6
1
3
3
1
2
3
16
3
1
5
7
5
6
2
38
3
3
2
3
5
3
18
4
2
2
4
6
3
6
33
3
2
2
8
2
3
3
4
1
1
1
1
2
1
14
1
1
3
6
2
3
2
32
2
1
1
2
3
2
16
2
2
2
2
5
1
2
1
1
2
1
1
0.3
2
1
2
2
1
1
2
2
2
0.2
2
1
2
2
0.3
6
1
2
0.5
1
2
2
2
2
1
0.3
2
1
2
3
4
65
72
79
68
69
91
70
64
45
45
48
68
36
90
39
83
61
83
55
59
87
84
71
37
68
56
58
47
90
45
70
85
49
81
34
42
89
79
37
92
100
100
100
48
0
0
0
0
0.333
0.333
0
0
0
0
0
0
0.5
0
0
0
0
0
0
0.333
0.333
0
0
0
0
0
0
0
1
0.5
0.5
0
0
0
0
0
0.333
0.333
0
0
0
0
0
0.5
0
0
0
0
0
0
0
0.333
0.333
0
0
0
0
0
0
0
0
1
0.5
0.5
0
0
0
0
0.333
0.333
0
0
0
0
0.5
0
0
0
0
0
0
0
0
0.333
0.333
0
0
0
0
0
0
0
0
0
1
0.5
0.5
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0.5
0.5
0.5
0.5
0.5
0.5
0.333
0.333
0.333
0.333
0.333
0.333
0.333
0.333
0
0
0
0
0
0
0
0
0
0.5
0
0
0.5
0
0
0.5
0
0
0.5
0
1
0.5
0
0
0.5
0
0
0.5
0
0
0.5
0
1
29
2
2
36
1
0.5
1
1
3
3
0.0
0.0
0.0
0.3
5
39
1
0.5
1
0
0
0
0
0
0
0
0
0
0
0
0
0.5
0.2
[a] Molar fraction.
cal features of the HTE equipment. For this reason, it is
very important to check the reproducibility of the most in-
teresting HTE results on a larger volume scale by repeating
the experiments in a conventional manner. Therefore, galli-
um oxide was prepared under the same conditions as em-
ployed for the synthesis of the best catalyst discovered in
this work (sample 40)but on a tenfold larger volume scale.
The material was tested over various cycles for the epoxida-
tion of cyclooctene with 50 wt% aqueous H2O2 (Table 3).
The results of the first two runs are in good agreement with
those obtained by HTE, confirming the general validity of
the HTE results reported here. After the second run, the cy-
clooctene conversion and the epoxide yield are seen to grad-
ually decrease, while the selectivity towards the epoxide is
retained. Calcination of the sample for 24 h at 4008C did
not serve to restore the features of the original catalyst
(run 5). This slow deactivation upon recycling is similar to
that observed for aluminium oxides, and is attributed to
structural changes of the material during the catalytic pro-
cess.[8,9] Alongside their epoxidation activity, transition-
metal-free oxides might catalyse the unwanted decomposi-
tion of H2O2 into O2 and H2O. The amount of H2O2 decom-
posed during the catalytic test (run 2)was determined by
means of titration with Ce4+ as 37% of the H2O2 employed,
corresponding to a selectivity of H2O2 towards the epoxide
of 49%. These values are of the same order as those found
Chem. Eur. J. 2007, 13, 6562 – 6572
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
6569

P. P. Pescarmona et al.
which active gallium oxides can be prepared is quite narrow
and could have been overlooked if a conventional research
approach had been followed. Another relevant result is the
identification of new, improved formulations for the synthe-
sis of aluminium oxides as active epoxidation catalysts. Fi-
nally, the screening of large parameter spaces characteristic
of an HTE approach has provided an understanding of the
effect of the synthesis parameters on the catalytic activity of
the transition-metal-free oxides.
Gallium oxide is a promising heterogeneous catalyst for
the sustainable epoxidation of alkenes, since it is active with
aqueous H2O2, it is selective, and it does not contain transi-
tion metals. However, its stability and its efficiency in using
H2O2 need to be improved before it can meet the require-
ments for industrial application. Future work can also be ex-
pected to extend the scope of application of the best cata-
lysts discovered here to other alkenes. Furthermore, these
catalysts will be fully characterised in order to shed more
light on the relationship between their physicochemical fea-
tures and their catalytic properties.
Figure 4. Ternary array representing the combination of the different
boron, aluminium, and gallium precursors used to synthesise metal oxide
catalysts for epoxidation reactions.
Table 3. Catalytic test of the larger volume scale version of catalyst 40 in
the epoxidation of cyclooctene with 50 wt% aqueous H2O2 [conditions:
4 h at 808C; H2O2:cyclooctene=2; solvent: ethyl acetate].
Experimental Section
The HTE samples were prepared by means of a Tecan Genesis RSP 100
liquid-handling robotic workstation coupled with a personal computer
supplied with Gemini software enabling programming of the workstation.
Liquids were dispensed and transferred using two kinds of needles: fixed
metal needles to dispense water and organic solvents and disposable plas-
tic needles suitable for handling viscous or contaminant liquids, that is,
the solutions of the metal precursors. High accuracy of the operations
performed with the needles was ensured by their prior calibration with
different classes of liquid: the error was generally no higher than 1%,
even when dealing with volumes as small as 100 mL. The workstation was
equipped with a reaction block developed in-house, consisting of a heat-
ing unit containing 60 wells for 10 mL glass vials (maximum working
temperature of 808C)and a Variomag plate for individual magnetic stir-
ring of the 60 parallel reaction vessels.
Run
Cyclooctene
conversion [%]
Epoxide
yield [%]
By-products
yield [%]
Epoxide
selectivity [%]
1
2
3
4
5
67
66
39
28
27
46
64
39
28
26
19
1
0.4
0.3
1
71
98
99
99
95
when using aluminium oxide as catalyst.[5,9] Finally, a leach-
ing experiment was performed: the solid catalyst was fil-
tered off from the reaction mixture after it had been stirred
for 30 min while heating from room temperature to 808C.
The clear filtrate was stirred for 4 h at 808C and epoxide
formation was monitored by GC analysis. An extremely low
epoxide yield was observed, proving that no active species
had leached from the catalyst and confirming that the pro-
cess is truly heterogeneous.
Chemicals used in the synthesis of the boron, aluminium, and gallium
oxides were as follows: boron methoxide, B
isopropoxide, B(OCH(CH3)2)3 [B(iPrO)3], aluminium nitrate nonahy-
drate, Al(NO3)3·9H2O, aluminium chloride, AlCl3, aluminium isopropox-
ide, Al(OCH(CH3)2)3 [Al(iPrO)3], aluminium sec-butoxide, Al(OCH-
(CH3)CH2CH3)3 [Al(sBuO)3], gallium nitrate hydrate, Ga(NO3)3·xH2O,
ACHRTUNEG(OCH3)3 [BCAHTRE(UGN OMe)3], boron
G
A
ACHTREUNG
AHCTREUNG
U
R
N
ACHTREUNG
G
G
ACHTREUNG
and gallium chloride, GaCl3, as metal precursors; 2-butanol, 2-propanol,
and ethanol as solvents; deionised water; oxalic acid [HOx]; and ammo-
nia, NH3, 25 wt% in aqueous solution. Chemicals used in the catalytic
tests were as follows: cis-cyclooctene (95%); di-n-butyl ether as an inter-
nal standard for the GC analyses; ethyl acetate as solvent; and hydrogen
peroxide, H2O2, 50 wt% in aqueous solution, as oxidant.
Conclusion
All samples discussed in this article were prepared by the hydrolytic con-
densation of a metal precursor or of a mixture of two or three of them.
The total amount of metal precursors was the same in all of the reported
experiments (1.00 mmol per sample). The remaining synthesis conditions
were different in the various experiments and will be discussed separately
(see below).
High-throughput experimentation techniques have been
used to explore the synthesis and application of transition-
metal-free oxides as epoxidation catalysts. The aim was to
discover improved pathways for the effective and sustaina-
ble catalytic epoxidation of alkenes. Our study has yielded
various interesting results. The most promising result is the
discovery of gallium oxide as a very active heterogeneous
catalyst for the epoxidation of alkenes using aqueous hydro-
gen peroxide. In this respect, the use of HTE proved to be
particularly useful: the range of synthesis conditions under
The synthesis mixtures were prepared by using the HTE workstation to
dispense the desired liquids into 10 mL glass vials placed in the 60-well
reaction block. The vials were sealed with caps and heated to 808C for
3 h while their contents were stirred at 500 rpm. The vials were uncapped
and left to stand at room temperature for 16 h in a fume hood. Next, the
samples were transferred to an oven and calcined according to the fol-
lowing temperature program: 25 to 1008C at 108Cminꢀ1; 24 h at 1008C;
6570
www.chemeurj.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 6562 – 6572

HTE Screening of Epoxidation Catalysts
FULL PAPER
100 to 2008C in 30 min; 12 h at 2008C; 200 to 4008C in 60 min; 24 h at
4008C. The samples were placed in the oven in special racks holding the
10 mL glass vials. It is considered that this arrangement might prevent
the circulation of air inside the vials and, therefore, affect the calcination
process.
to the formation of a dark-brown solution.[27] An aqueous solution of
oxalic acid was prepared by dissolving the acid (56.3 mmol)in deionised
H2O (75.0 mL). By means of the HTE workstation, 45 samples were pre-
pared, each containing a solution of metal precursors (5.000 mL)to
which an aliquot (1.000 mL)of the aqueous solution of oxalic acid was
added with stirring. Each sample had a different ratio of the three metal
precursors, in order to describe the desired ternary array (Figure 1). The
preparation of the 45 samples required around 2 h. After calcination, the
materials exhibited colours ranging from white to brown. Low yields of
the final solid were observed for the samples with high boron content.
The materials obtained were ground to fine powders and tested for their
activities in the epoxidation of cyclooctene with hydrogen peroxide
under similar conditions to those employed in the literature for testing
aluminas.[9,10] However, a 50 wt% aqueous solution of hydrogen peroxide
was used instead of a 70 wt% solution. Cyclooctene (2.50 mmol), di-n-
butyl ether (1.25 mmol), ethyl acetate (2.50 g), and hydrogen peroxide
(5.00 mmol)were added to each sample. First, a solution containing cy-
clooctene, di-n-butyl ether, and ethyl acetate was added to each solid cat-
alyst with stirring. Then, the aqueous solution of hydrogen peroxide was
added. Both solutions were dispensed using the HTE workstation. The
reaction mixtures were stirred for 4 h at 500 rpm and 808C in capped
vials placed in the 60-well reaction block. The rubber septa of the caps
were pierced with a sharp needle to prevent the development of an over-
pressure in the reactors during the catalytic test.
Method 2:[8,10] Three solutions of metal precursors were prepared by dis-
solving B
8.000 mL). Each solution had a different density: 1.00 mmol of B
was contained in 0.633 mL of a clear, colourless solution; 1.00 mmol of
Al(sBuO)3 was contained in 0.667 mL of a clear, colourless solution;
ACHRTUNEG(iPrO)3, AlACHTRENUG
ACHTREUNG
AHCTREUNG
1.00 mmol of GaCl3 was contained in 0.487 mL of a dark brown solution.
An aqueous solution of oxalic acid was prepared by dissolving the acid
(11.3 mmol)in deionised H O (15.0 mL). By means of the HTE worksta-
2
tion, 45 samples were prepared, each containing a solution of the metal
precursors (total 1.00 mmol)in 2-butanol (0.400 mL), to which an aliquot
(0.196 mL)of the aqueous oxalic acid solution was added with stirring.
Each sample had a different ratio of the three metal precursors, in order
to describe the desired ternary array (Figure 1). The preparation of the
45 samples required around 40 min. After calcination, the materials ex-
hibited colours ranging from white to brown, the intensity of the colour
being proportional to the gallium content.
Each HTE series of experiments, from the synthesis of the oxides to the
catalytic tests, was carried out in a single set of glass vials, thus avoiding
transfers of the samples that may have led to a loss of material.
Cyclooctene conversions and the epoxycyclooctane and by-product yields
were determined by gas chromatography (GC)analysis on a Finnigan
Trace GC Ultra chromatograph from Interscience, equipped with an
RTX-5 fused silica column (10 m; 0.1 mm). The analysis time for each
sample was just 2.25 min by virtue of the rapid heating and cooling
system of the column (Ultra-Fast Module). The temperature profile
during the analysis was as follows: 45 s at 708C, 70 to 2508C at
1808Cminꢀ1, 30 s at 2508C. The GC samples were prepared by adding an
aliquot of the reaction mixture (ꢁ0.3 mL)to an equal volume of decane:
any H2O that might have been present in the sample would have separat-
ed in a phase at the bottom of the GC flask, preventing its injection into
the chromatograph and consequent deterioration of the column. The con-
versions and yields were calculated by normalising the areas of the GC
peaks by means of the area of the internal standard peak.
Method 3:[22] Three solutions of metal precursors were prepared by dis-
solving BACHTERNGU(iPrO)3, AlACHTREU(GN sBuO)3 or GaCl3 (20.0 mmol)in 2-butanol to give a
total volume of 100.0 mL in each case. GaCl3 was used as gallium source
in order to avoid the washing step that would be needed to remove the
explosive NH4NO3 that would have been formed in the presence of aque-
(NO3)3 had been used as gallium source.[22] By means of
ous NH3 if GaACHTREUNG
the HTE workstation, 45 samples were prepared, each containing a solu-
tion of the metal precursors (5.000 mL)to which a 1:1 (v/v)mixture of
ethanol and a 25 wt% aqueous solution of ammonia (1.600 mL)was
added with stirring. The total volume of 1.600 mL was added in separate
aliquots of 0.160 mL each, over a period of 1 h, to mimic dropwise addi-
tion. Each sample had a different ratio of the three metal precursors, to
describe the desired ternary array (Figure 1). The preparation of the 45
samples required around 3 h. After calcination, the materials displayed
colours ranging from white for samples containing aluminium but no gal-
lium, to dark brown and black for samples with high gallium content.
The by-products were identified by means of gas chromatography-mass
spectrometry (GC-MS)analysis using an Agilent 6890N gas chromato-
graph coupled with an Agilent 5973 MSD mass spectrometer. The GC
was equipped with a WCOT fused-silica column (30 m; 0.25 mm)coated
with a 0.25 mm thick HP-5 MS film. The temperature program was analo-
gous to that employed for the GC analysis.
Samples 1–48: Two solutions of metal precursors were prepared by dis-
solving AlACHTRE(UNG sBuO)3 or GaCl3 (35.0 mmol)in 2-butanol (11.214 g;
14.000 mL). An aqueous solution of oxalic acid was prepared by dissolv-
ing the acid (11.3 mmol)in deionised H 2O (15.0 mL). By means of the
HTE workstation, 48 samples were prepared. First, a solution of either
For recycling of the HTE samples, a washing procedure using the auto-
mated workstation was developed. After the catalytic tests, the reaction
solution was removed from each of the samples. Ethanol (5 mL)was
then added to each sample, and the suspensions obtained were stirred for
5 min. Next, the samples were centrifuged for 10 min to deposit the solid
catalysts. The supernatant ethanolic solution was then removed from
each sample; to avoid loss of solid during this operation, the liquid was
aspirated from above the level of the solid, implying that a thin layer of
ethanol could not be removed. This washing procedure was repeated
four times. Finally, the samples were dried for 16 h in an oven at 1108C.
AlACHTRE(UNG sBuO)3 or GaCl3 (1.00 mmol)in 2-butanol (0.400 mL)was dispensed
into each glass vial. Some of the samples (see Figure 2)were diluted by
adding 0.600 mL, 2.100 mL, or 4.600 mL of 2-butanol with stirring. Next,
a basic, neutral, or acidic hydrolysing solution was added to each sample
with stirring, that is, a 1:1 (v/v)mixture of ethanol and 25 wt% aqueous
ammonia (0.320 mL; 2.14 mmol NH3 or 1.600 mL; 10.7 mmol NH3), de-
ionised H2O (0.196 mL or 0.980 mL), or aqueous oxalic acid (0.196 mL;
0.147 mmol HOx or 0.980 mL; 0.736 mmol HOx)(Figure 2.) The hydro-
lysing solutions were dispensed in separate aliquots of 0.160 mL (basic
solutions)or 0.196 mL (neutral and acidic solutions)to mimic dropwise
addition. The preparation of the 48 samples required about 1 h. Gallium
samples prepared with 0.980 mL of H2O or aqueous HOx exhibited two
distinctly separated liquid phases: a clear, colourless aqueous phase at
the bottom and a dark-brown 2-butanol phase at the top. The volume of
the aqueous phase progressively decreased as the volume of 2-butanol
was increased from 0.400 to 5.000 mL. After calcination, the aluminium
oxides were white or off-white while the gallium oxides displayed colours
ranging from brown to dark grey and black. The yield of the final solid
was approximately the same for all of the aluminium oxides, whereas
lower yields were observed among the gallium oxides for the samples
synthesised under basic conditions.
A
similar but non-automated procedure was used to recycle larger
volume scale samples.
The amount of H2O2 decomposed into H2O and O2 during the catalytic
tests was determined by titrating the reaction solution (after 4 h at 808C)
with a 0.1m solution of CeACTHERUNG(SO4)2, obtained by dissolving CeCAHTREU(GN SO4)2·4H2O
(50.0 mmol)in H 2SO4 (28 mL)and doubly-distilled H 2O (28 mL)and di-
luting to a total volume of 500 mL with H2O. The reaction solution was
diluted with H2O (18 mL)and with a 7 vol% aqueous solution of H 2SO4
(2 mL). The obtained colourless solution was titrated with the 0.1m CeIV
solution until it turned yellow (2Ce4+ + H2O2 ! 2Ce3+ + 2H+ + O2).
Synthesis of the catalysts
Method 1: Three solutions of metal precursors were prepared by dissolv-
ing BACHTREUNG(iPrO)3, AlACHTREUNG(sBuO)3 or GaCl3 (20.0 mmol)in 2-butanol to give a
total volume of 100.0 mL in each case. The dissolution of GaCl3 is a fast,
exothermic process, accompanied by the evolution of HCl gas, and leads
Chem. Eur. J. 2007, 13, 6562 – 6572
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
6571

P. P. Pescarmona et al.
Samples 49–92: Solutions of boron and gallium precursors were prepared
by dissolving B(OMe)3, B(iPrO)3, Ga(NO3)3·xH2O or GaCl3 (8.00 mmol)
in ethanol to give a total volume of 20.0 mL in each case. Solutions of
aluminium precursors were prepared by dissolving Al(NO3)3·9H2O or
AlCl3 (6.00 mmol)in ethanol, Al (iPrO)3 (6.00 mmol)in 2-propanol, or
Al(sBuO)3 (6.00 mmol)in 2-butanol to give a total volume of 15.0 mL in
[4] G. Grigoropoulou, J. H. Clark, J. A. Elings, Green Chem. 2003, 5, 1–
7.
[5] D. Mandelli, M. C. A. van Vliet, R. A. Sheldon, U. Schuchardt,
Appl. Catal. A 2001, 219, 209–213.
[6] M. C. A. van Vliet, D. Mandelli, I. W. C. E. Arends, U. Schuchardt,
R. A. Sheldon, Green Chem. 2001, 3, 243–246.
A
G
ACHTREUNG
AHCTREUNG
C
ACHTREUNG
each case. By means of the HTE workstation, 44 samples were prepared,
each containing a solution of metal precursors (2.500 mL), to which de-
ionised H2O (0.196 mL)was added with stirring. The precursors of each
of the three Group 13 elements were combined with each precursor of
the other two elements in the ratios indicated in Table 2. The preparation
of the 44 samples required around 1 h. After calcination, the materials
containing exclusively boron and those prepared using AlCl3 and/or
GaCl3 displayed grey/brown colours, while all of the other solids were
white. Very low yields of the final solid were observed for the samples
containing exclusively boron.
[7] V. R. Choudhary, N. S. Patil, N. K. Chaudhari, S. K. Bhargava, J.
Mol. Catal. A 2005, 227, 217–222.
[8] R. G. Cesquini, J. M. de Silva, C. B. Woitiski, D. Mandelli, R. Rinal-
di, U. Schuchardt, Adv. Synth. Catal. 2002, 344, 911–914.
[9] R. Rinaldi, J. Sepffllveda, U. Schuchardt, Adv. Synth. Catal. 2004,
346, 281–285.
[10] R. Rinaldi, U. Schuchardt, J. Catal. 2004, 227, 109–116.
[11] P. P. Pescarmona, J. C. van der Waal, I. E. Maxwell, T. Maschmeyer,
Catal. Lett. 1999, 63, 1–11.
[12] B. Jandeleit, D. J. Schaefer, T. S. Powers, H. W. Turner, W. H. Wein-
berg, Angew. Chem. 1999, 111, 2648–2689; Angew. Chem. Int. Ed.
1999, 38, 2494–2532.
Note: It is important to keep in mind that whereas the trends identified
in high-throughput experiments can always be trusted, the level of relia-
bility of each individual experiment constituting an HTE series is gener-
ally lower than that of a conventional laboratory experiment. This is due
to many factors: the experimental errors that can arise during an HTE
run can be specific to the HTE equipment being used (e.g., contamina-
tion with the liquid used to operate the syringes connected to the nee-
dles), but may also be similar to those occurring in standard laboratory
experiments (e.g., due to inefficient stirring, leaching of volatile com-
pounds through the septa of the reactor caps). Simple observations that
help to prevent such errors are often more difficult when dealing with
many simultaneous experiments performed in parallel vessels inside a re-
action block. To increase the reliability of the HTE results, some samples
were prepared in duplicate or in triplicate: in such cases, the average
values of the catalytic results are reported. Another means of checking
and improving the reliability of the HTE results is to include in each
HTE series one or more individual experiments from the previous series:
the level of reproducibility of the results obtained for the repeated ex-
periments in different series can provide a measure of the consistency of
the data in each series.
[13] S. Senkan, Angew. Chem. 2001, 113, 322–341; Angew. Chem. Int.
Ed. 2001, 40, 312–329.
[14] J. G. de Vries, A. H. M. de Vries, Eur. J. Org. Chem. 2003, 799–811.
[15] P. P. Pescarmona, J. C. van der Waal, L. G. A. van de Water, T.
Maschmeyer, in High-Throughput Screening in Heterogeneous Catal-
ysis (Eds.: A. Hagemeyer, P. Strasser, A. F. Volpe, Jr.), Wiley-VCH,
Weinheim, 2004, pp. 211–237.
[16] J. N. Cawse, Acc. Chem. Res. 2001, 34, 213–221.
[17] A. Holzwarth, P. Denton, H. Zanthoff, C. Mirodatos, Catal. Today
2001, 67, 309–318.
[18] Experimental Design for Combinatorial and High-Throughput Mate-
rials Development (Ed.: J. N. Cawse)Wiley Interscience, Hoboken,
2003.
[19] S. R. M. Pereira, F. Clerc, D. Farrusseng, J. C. van der Waal, T.
Maschmeyer, C. Mirodatos, QSAR Comb. Sci. 2005, 24, 45–57.
[20] P. P. Pescarmona, J. C. van der Waal, I. E. Maxwell, T. Maschmeyer,
Angew. Chem. 2001, 113, 762–765; Angew. Chem. Int. Ed. 2001, 40,
740–743.
[21] A. Corma, J. M. Serra, P. Serna, S. Valero, E. Argente, V. Botti, J.
Catal. 2005, 229, 513–524.
[22] C. Otero Areµn, M. Rodríguez Delgado, V. Montouillout, D. Mas-
siot, Z. Anorg. Allg. Chem. 2005, 631, 2121–2126.
[23] C. Otero Areµn, A. López Bellan, M. PeÇarroya Mentruit, M. Ro-
dríguez Delgado, G. Turnes Palomino, Microporous Mesoporous
Mater. 2000, 40, 35–42.
[24] C. J. Brinker, G. W. Scherer, Sol-Gel Science: The Physics and
Chemistry of Sol-Gel Processing, Academic Press, Boston, 1990,
Chapter 2.
Acknowledgements
PPP acknowledges the Fund for Scientific Research (FWO)for a post-
doctoral fellowship. The authors acknowledge sponsorship in the frame
of the following research programs: GOA, VIRCAT, IDECAT, and
CECAT.
[25] L. F. Nazar, L. C. Klein, J. Am. Ceram. Soc. 1988, 71, C85–C87.
[26] M. Baerns, C. Mirodatos, NATO Science Series, Ser. II 2002, 69,
469–479.
[27] S. R. Bindal, V. K. Mathur, R. C. Mehrotra, J. Chem. Soc. A 1969, 6,
863–867.
[1] G. Sienel, R. Rieth, K. T. Rowbottom, Ullmannꢀs Encyclopedia of
Organic Chemicals, Wiley-VCH, Weinheim, 1999.
[2] D. E. De Vos, B. F. Sels, P. A. Jacobs, Adv. Synth. Catal. 2003, 345,
457–473.
Received: January 17, 2007
[3] B. S. Lane, K. Burgess, Chem. Rev. 2003, 103, 2457–2473.
Published online: May 16, 2007
6572
www.chemeurj.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2007, 13, 6562 – 6572