Aerobic epoxidation of alkenes using polymer-bound Mukaiyama catalysts

Article abstract of DOI:10.1039/b005066k

Nickel(II) acetylacetonate bound to polybenzimidazole has been shown to be an efficient catalyst for the aerobic epoxidation of alkenes with an aldehyde as co-reactant. Rates of reaction and yields of epoxide are higher with this system than with dissolved Ni(II) acetylacetonate under otherwise identical conditions. The polymer-bound complex acts as a heterogeneous catalyst and can be recovered from the reaction mixture. Some loss of activity is observed upon re-use of the catalyst due to leaching of the metal complexes.

Full text of DOI:10.1039/b005066k

View Article Online / Journal Homepage / Table of Contents for this issue
Aerobic epoxidation of alkenes using polymer-bound Mukaiyama
catalysts
Bastienne B. Wentzel,a Salla-M. Leinonen,b Stephen Thomson,b David C. Sherrington,b
Martinus C. Feitersa and Roeland J. M. Nolte*a
1
a Department of Organic Chemistry, NSR Center, University of Nijmegen, Toernooiveld 1,
6525 ED, Nijmegen, The Netherlands. E-mail: nolte@sci.kun.nl
b Department of Pure and Applied Chemistry, University of Strathclyde, 295 Cathedral Street,
Glasgow, UK G1 1XL
Received (in Cambridge, UK) 26th June 2000, Accepted 14th August 2000
First published as an Advance Article on the web 2nd October 2000
Nickel() acetylacetonate bound to polybenzimidazole has been shown to be an efficient catalyst for the aerobic
epoxidation of alkenes with an aldehyde as co-reactant. Rates of reaction and yields of epoxide are higher with this
system than with dissolved Ni() acetylacetonate under otherwise identical conditions. The polymer-bound complex
acts as a heterogeneous catalyst and can be recovered from the reaction mixture. Some loss of activity is observed
upon re-use of the catalyst due to leaching of the metal complexes.
In contrast to the oxidation reactions mentioned above,
aerobic oxidations using polymer-bound catalysts are not
Introduction
The aerobic epoxidation of alkenes with a transition metal
catalyst has been widely studied over the past decade. A well
known, efficient method of alkene epoxidation in solution is the
so-called ‘Mukaiyama’ procedure, in which an unfunctional-
widespread in the literature. A polyaniline-supported cobalt()
catalyst was prepared by Das and Iqbal27 and used for the
aerobic epoxidation of alkenes. More recently, polystyrene-
bound nickel complexes were synthesized by Han and Lei.28
ized alkene is epoxidized very efficiently using a transition metal
These catalytic systems showed activity in the aerobic oxidation
of cyclohexene. A robust polymer-bound catalyst for the
aerobic epoxidation of alkenes under Mukaiyama’s conditions
would be an interesting alternative for the epoxidation catalysts
β-diketonate complex as catalyst, molecular oxygen as oxidant
and an aliphatic aldehyde as co-reactant.1 For example, when
bis(acetylacetonato)nickel() (Ni(acac)2) is used as a catalyst,
and isobutyraldehyde as a co-reagent, aliphatic and cyclic
already described in the literature. It should have the high activ-
alkenes can be converted into the corresponding epoxides
ity of a homogeneous catalytic system and the advantages of
with yields ranging from 72% [(S)-limonene†] to 96%
filtration and re-usability of a heterogeneous catalyst. As part
(norbornene†).2 Other metal β-diketonate complexes, e.g.
of a project concerning the oxidation of terpenes for use in the
flavors and fragrances industry we have used limonene as a
model substrate. The oxidation of this compound is the first
CoII(acac)2,3 MnII(acac)2,3,4 and FeIII(acac)3 are effective as
5
catalysts as well. Transition metal complexes such as Schiff’s
base complexes are also suitable for the aerobic epoxidation of
step in a synthetic route toward carvone, a spearmint flavor
alkenes with a co-reacting aldehyde. For example, cobalt()
used in foodstuffs. We present our first studies aimed at the
Schiff’s bases give good results but these catalysts are not
development of such a catalytic system in this paper.
always selective for epoxidation.6–12 Manganese–salen com-
plexes have been used for the enantioselective aerobic epoxid-
Results and discussion
ation of alkenes with molecular oxygen and aldehydes as
reagents. Enantiomeric excesses (ee’s) as high as 92% are
possible with this catalytic system.13–17
The results of limonene epoxidation after 4 hours are collected
in Table 1. It can be seen that the polymer-bound catalyst PBI-
Ni is more active (entry 1) than the homogeneous catalyst
(entry 11) and also displays better selectivity toward formation
of limonene oxide. PBI-Co is also very active but somewhat less
selective (entry 2). PBI-Mo is not active at all with oxygen as an
oxidant. No reaction was observed without catalyst at ambient
temperature or with PBI resin without an attached metal
complex. Neither did the reaction proceed when no aldehyde
co-reactant was present. However, at 45 ЊC in the autoclave set-
up, the reaction did proceed without catalyst after a variable
induction period of approximately 50–100 minutes (entry 12).
The main product of limonene epoxidation was in all cases the
1,2-epoxide, i.e. the endocyclic epoxide. The isopropenyl ring
was found to be much less reactive toward epoxidation under the
present conditions, in agreement with our earlier studies using
the corresponding homogeneous nickel catalyst.2 GC analysis
showed the 1,2-epoxide to be a 60:40 mixture of the cis and
trans isomers. In all cases the main by-product of the reaction
It has been reported in the literature that the epoxidation of
alkenes by various oxidants can proceed very efficiently when a
polymer-supported catalyst is used. For example, the epoxid-
ation of propene, octene and other alkenes with hydroperoxide
using a polymer-bound MoVIO2(acac)2 catalyst was shown to
be possible with yields up to 100%.18–22 A Jacobsen-type
manganese catalyst has been anchored to various polymer
supports and has been used with different oxidants to yield
epoxides with ee’s up to 95%.23–25 The polymer-supported
catalysts were thermo-oxidatively very stable and could be
recycled without loss of activity.26 An overview of polymer-
supported alkene epoxidation catalysts has been published
recently by Sherrington.22
† The IUPAC name for limonene is 4-isopropenyl-1-methylcyclo-
hexene, for norbornene is bicyclo[2.2.1]hept-2-ene and for carveol is
4-isopropenyl-1-methylcyclohex-6-en-2-ol.
3428
J. Chem. Soc., Perkin Trans. 1, 2000, 3428–3431
This journal is © The Royal Society of Chemistry 2000
DOI: 10.1039/b005066k

View Article Online
Table 1 Aerobic epoxidation of (S)-limonene with various polymer-
bound catalystsa
Conversion Yield of
(%)
Entry Catalyst
Substrate
epoxideb (%)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
PBI-Ni
PBI-Co
(S)-Limonene
93
99
0
51
0
0
96
74
61
0
31
0
0
73
49 (25)
46 (73)
11 (44)
65
45 (13)
88
PBI-Mo
DowexA-Ni
DowexB-Ni
BP-Ni
PBI-Nic
DowexA-Nid
DowexB-Nid
BP-Nid
75 (100)
50 (100)
17 (100)
70
94 (100)
100
88
43
Fig. 1 Activity of the polymer-supported PBI-Ni catalyst after
filtration. Conditions: 5 ml CH2Cl2, 1 mmol (S)-limonene, 3 mmol
isobutyraldehyde, 1 mol% NiII, 25 ЊC, O2-filled balloon. Filtration was
carried out after 120 min (50% conversion) using a glass funnel.
Ni(acac)2
Noned
PBI-Ni
PBI-Ni
PBI-Ni
α-Pinene
Oct-1-ene
Styrenee
52
23
a Reaction conditions: 5 ml CH2Cl2, 1 mmol substrate, 3 mmol iso-
center that the reaction is prevented. It is known that addition
of an excess of pyridine to a reaction mixture containing
Ni(acac)2 completely inhibits the reaction2 since two pyridine
molecules are coordinated to the NiII center. The DowexA resin
has only one pyridine ligand which seems to decrease the
reactivity of the Ni center, but does not completely inhibit
the reaction.
butyraldehyde, 25.0 ЊC, O2-filled balloon,
4
h. b In the case of
limonene the 1,2-epoxide is formed as a 60:40 cis–trans mixture. In the
case of α-pinene the reaction mixture was not analyzed for diastereo-
meric products. c At 25.0 ЊC in toluene, 7 bar pressure, 8% O2 in N2,
18 h. d At 45.0 ЊC in toluene, 7 bar pressure, 8% O2 in N2, ~2 h (in
parentheses: 18 h). e 2,3-Dimethylvaleraldehyde was used instead of
isobutyraldehyde.
At higher pressure in the autoclave set-up the reactivity of
the polymer-bound catalysts was quite different (Table 1,
entries 7–10). Using PBI-Ni, (S)-limonene could be oxidized
efficiently at 25 ЊC and 7 bar pressure (entry 7). Rather remark-
ably, the other three supported catalysts were not active at
25 ЊC and the temperature had to be raised to 45 ЊC in order to
obtain some catalytic activity. The uncatalyzed (free radical)
reaction at 45 ЊC (entry 12) displayed a similar epoxidation
selectivity to DowexA-Ni (entry 8) whereas DowexB-Ni
and BP-Ni (entries 9 and 10) were less active and selective.
We therefore conclude that the latter two polymer supports
inhibit the reaction as was suggested above (vide infra). The
selectivity of these catalysts is very low at high conversions,
which might also be due to the high temperature which favors
uncatalyzed free radical oxidations at the expense of a catalyzed
reaction.
was the 7,8-epoxide in yields to 10%, depending on the condi-
tions. Diepoxide formation (at the cost of the 1,2-epoxide) was
observed when the epoxidation was run for a long period of
time (more than 5 hours). Otherwise, the diepoxide was formed
in less than 1%. Some other by-products of the reaction
included carveol† and unidentified isomers of carveol, all in
small amounts up to 1%.
The differences between the various polymer resins used as
the support are shown in entries 1 and 4–6. Apart from PBI,
only the DowexA polymer 2 is a suitable support for the
Some other alkenes were tested as substrates for the aerobic
epoxidation with the most active catalyst PBI-Ni (Table 1,
entries 13–15). The epoxidation of α-pinene and oct-1-ene was
very efficient. Styrene epoxidation however was slow, and
probably polymers were also formed as reaction products.
The PBI-Ni catalyzed epoxidation of (S)-limonene was
shown to be heterogeneous (Fig. 1) since the reaction stopped
immediately when the polymer catalyst was filtered off. The
filtered reaction mixture was analyzed for nickel content using
atomic absorption spectroscopy (ICP). It appeared that ~20%
of the original NiII present on the polymer had leached into the
solution, but this unidentified Ni species was not active as an
epoxidation catalyst. Clearly, the ligand environment of NiII is
very important for the catalytic activity. ICP analysis of the
used polymer confirmed the ~20% loss of nickel content. When
the used PBI-Ni was washed with CH2Cl2 and tested again in a
second epoxidation run, only 30% of the original epoxidation
activity was left (Fig. 2). This is in contrast to a MoVI loaded
PBI support which has been shown to be very thermo-
oxidatively stable and retains activity on recycling.21,26 The NiII
loaded resin described here appears not to be as stable towards
oxidative radical conditions.
Ni(acac)2 catalyst under ambient conditions, although the
epoxidation reaction proceeds less selectively and more slowly
than with PBI-Ni. It is understandable that the BP resin is not
an effective support for the catalyst (entry 6) since it contains a
secondary amine which has been observed to inhibit aerobic
epoxidation using NiII and an aldehyde.29 Why the DowexB
resin 3 does not afford an active catalyst is not clear. Possibly
the two pyridine ligands are both so tightly bound to the NiII
The polymer-catalyzed epoxidation appears to display the
same features as the homogeneous Ni(acac)2 catalyzed
reaction.2 First, addition of a radical trapping compound
(2-tert-butyl-4-methylphenol) during the reaction inhibited the
epoxidation completely. The presence of this reagent at the start
of the reaction prevented epoxidation from occurring at all.
J. Chem. Soc., Perkin Trans. 1, 2000, 3428–3431
3429

View Article Online
Table 2 Analytical data of polymer-bound metal complexes
Ni-1
Co-1
Mo-1
Ni-2
Ni-3
Ni-4
Metal loading (mmol gϪ1 resin)
Metal weight (mg gϪ1 resin)
% Metal
0.33
19.4
1.9
0.28
16.5
1.6
1.78
26.8
2.7
1.36
79.8
8.0
1.26
74.0
7.4
0.44
25.8
2.6
Ratio ligand:metal
9.8:1
11.6:1
1.8:1
2.6:1
2.5:1
9.1:1
%C
Calculateda
Found
77.9
Nd
77.9
Nd
77.9
Nd
76.57
65.8
79.97
65.6
62.4
54.0
%H
Calculateda
Found
3.9
Nd
3.9
Nd
3.9
Nd
7.85
7.0
6.71
6.1
7.24
7.0
%N
Calculateda
Found
18.2
Nd
18.2
Nd
18.2
Nd
9.92
5.3
13.33
6.7
11.2
4.1
a Calculated for the uncomplexed polymer. Decreased weight percentages after complexation indicate that metal is complexed to the ligand.
Experimental
Materials
CH2Cl2 was dried over CaCl2 and distilled from CaH2 under
dry nitrogen and stored over molecular sieves. All other
solvents and aldehydes were distilled before use. Oxygen was
obtained from Air Liquide. All alkene substrates were com-
mercial samples (Aldrich or Fluka, 96–98%) and were purified
by vacuum distillation. Epoxide products were identified using
gas chromatography and were compared to authentic samples.
Ni(acac)2 was purchased from Fluka (95%). PBI resin 1 was a
gift from Hoechst-Celanese, DowexA and DowexB resins (2
and 3) were from the Dow Chemical company and BP resin 4
was from the BP company.
Fig. 2 Re-use of the PBI-Ni catalyst. Conditions: 5 ml CH2Cl2,
1 mmol (S)-limonene, 3 mmol isobutyraldehyde, 1 mol% NiII, 25 ЊC,
O2-filled balloon. a: fresh PBI-Ni catalyst, b: 2nd run with the same
batch of PBI-Ni catalyst.
Instrumentation
GC analyses were performed on a Varian 3800 instrument with
a Supelco fused-silica capillary column (15 m length, 35 µm ID,
df = 1.0) with a FFAP stationary phase. Data were analyzed
with Varian Star 5.2 software. Metal analyses were carried out
using simultaneous/sequential inductively coupled plasma
atomic emission spectroscopy (ICP-AES) on a Spectroframe
instrument (Spectro Analytical Instruments). A 500 ppb
nickel() nitrate solution in dilute nitric acid (Spectrosol, BDH
Chemicals) was used as a standard.
A mixture of cis- and trans-stilbene oxide was obtained when
(Z)-stilbene was used as a substrate. This means that the
epoxidation reaction takes place in two steps as observed with
autoxidation chemistry, and not via a concerted peracid
mechanism. Furthermore, the reaction did not proceed in the
absence of a branched aliphatic aldehyde; without a Ni catalyst
it proceeded only very slowly after addition of MCPBA. These
factors point to a radical reaction via the same mechanism as
the Ni(acac)2-catalyzed reaction in homogeneous solution. The
experiments presented above have demonstrated that the ligand
environment of the NiII catalyst in the aerobic epoxidation
using the Mukaiyama system is crucial. This leads us to believe
that the reaction largely proceeds in the coordination sphere of
NiII, as we have proposed earlier.2
Preparation of the polymer-supported complexes
The four resins (PBI 1, DowexA 2, DowexB 3 and BP 4) were
stirred overnight in aqueous 1 M NaOH, washed with H2O
until neutral, then extracted in a Soxhlet with acetone (24 h)
and finally dried under vacuum. The attachment of the
catalysts was achieved by refluxing a mixture of the resin with
an excess of the appropriate M(acac)n complex in toluene (50
ml) for 48 h. The resulting resins were purified by extraction in a
Soxhlet with toluene for 48 h and dried in vacuo at 40 ЊC (48 h).
Metal contents were determined by microanalysis. The results
of all metal analyses and microanalytical analyses are summar-
ized in Table 2.
Conclusions
The polymer-supported nickel() complex PBI-Ni turns out
to be a very active and selective catalyst for the epoxidation
of alkenes under ‘Mukaiyama’s conditions’. The epoxidation of
(S)-limonene proceeds more efficiently and gives a higher yield
when it is catalyzed by PBI-Ni than by the soluble Ni(acac)2
catalyst. The reaction is heterogeneous and the catalyst can be
recovered from the reaction mixture. Unfortunately, upon
regeneration of the polymer-supported nickel complex, the
catalyst loses some activity which is due to leaching of some
inactive metal complex. Further studies will be aimed at
reducing this leaching effect.
Description of the catalytic system
Epoxidation at ambient conditions. Alkene (1 mmol), alde-
hyde (3 mmol), and the appropriate amount of polymer com-
plex corresponding to 0.01 mmol of metal as the catalyst, were
stirred (1000 rpm) in CH2Cl2 (5 ml) at 25.0 0.5 ЊC under 1.0
atm of oxygen. Unless indicated otherwise, the experiments
were carried out with (S)-limonene as the substrate and iso-
3430
J. Chem. Soc., Perkin Trans. 1, 2000, 3428–3431

View Article Online
butyraldehyde as the co-reagent. The catalytic reaction was
followed by monitoring the disappearance of the substrate and
the appearance of product(s) as a function of time with gas
chromatography using 1,2,4-trichlorobenzene as the external
standard.
References
1 T. Mukaiyama, Aldrichimica Acta, 1996, 29, 59.
2 B. B. Wentzel, P. A. Gosling, M. C. Feiters and R. J. M. Nolte,
J. Chem. Soc., Dalton Trans., 1998, 2241.
3 N. Fdil, A. Romane, S. Allaoud, A. Karim, Y. Castanet and
A. Mortreux, J. Mol. Catal. A: Chem., 1996, 108, 15.
4 T. Mukaiyama, T. Yamada, T. Nagata and K. Imagawa, Chem.
Lett., 1993, 327.
Epoxidation at high pressure. The epoxidation reactions at
high pressure were performed in a Premex autoclave reactor
with Hastelloy C276 wet parts, equipped with a HC276 Dean–
Stark water separator, a 4-blade stirrer (1500 rpm), a sintered
HC276 gas-inlet (5 µm frit) and a sampling tube. The temper-
ature was regulated with a Premex C-M2 control unit to
0.1 ЊC. The nitrogen and oxygen gas inlet was regulated by
mass-flow controllers (MFC’s) and could be controlled to
0.1% v/v O2 in N2. After depressurizing, the exhaust gas was
cooled (Ϫ80 ЊC) to condense any vapor that was present. The
gas was then analyzed for carbon dioxide and carbon monoxide
with a Maihak Multor 610 CO2/CO analyzer (IR detection)
and for oxygen content with a Servomex 570A O2 analyzer.
Dipentene [60 mmol, a mixture of (R)- and (S)-limonene] and
toluene (150 ml) were loaded into the autoclave and the Dean–
Stark cooler was filled with toluene. This mixture was equi-
librated for at least 1 h at 25.0 ЊC in an atmosphere of 8.5% v/v
O2 in N2 (7 bar total pressure). No reaction was observed during
this period. The autoclave was opened and the appropriate
amount of polymer-bound catalyst (0.06 mmol, 0.1 mol% of
metal catalyst), and isobutyraldehyde (180 mmol) were added
quickly and the autoclave was closed and pressurized again.
This was taken as time t = 0. During the reaction the pressure
was maintained at 7.0 0.1 bar and the temperature at
25.0 0.1 ЊC. The stirring rate was 1500 rpm. Samples were
taken regularly and analyzed with GC using 1,2,4-trichloro-
benzene as the external standard and 2-tert-butyl-4-methyl-
phenol as a stabilizer (radical trapping agent). The oxygen
content of the exhaust gas was registered and the CO2 and
CO content monitored on-line in real-time using a personal
computer.
5 T. Takai, E. Hata, T. Yamada and T. Mukaiyama, Bull. Chem. Soc.
Jpn., 1991, 64, 2513.
6 R. A. Budnik and J. K. Kochi, J. Org. Chem., 1976, 41, 1384.
7 K. Yorozu, T. Takai, T. Yamada and T. Mukaiyama, Chem. Lett.,
1993, 1579.
8 T. Takai, E. Hata, K. Yorozu and T. Mukaiyama, Chem. Lett., 1992,
2077.
9 D. E. Hamilton, R. S. Drago and A. Zombeck, J. Am. Chem. Soc.,
1987, 109, 374.
10 K. Kato, T. Yamada, T. Takai, S. Inoki and S. Isayama, Bull. Chem.
Soc. Jpn., 1990, 63, 179.
11 S. Bhatia, T. Punniyamurthy, B. Bhatia and J. Iqbal, Tetrahedron,
1993, 49, 6101.
12 T. Punniyamurthy, B. Bhatia, M. M. Reddy, G. C. Maikap and
J. Iqbal, Tetrahedron, 1997, 53, 7649.
13 T. Yamada, K. Imagawa, T. Nagata and T. Mukaiyama, Chem.
Lett., 1992, 2231.
14 T. Yamada, K. Imagawa, T. Nagata and T. Mukaiyama, Bull. Chem.
Soc. Jpn., 1994, 67, 2248.
15 K. Imagawa, T. Nagata, T. Yamada and T. Mukaiyama, Chem.
Lett., 1994, 527.
16 T. Nagata, K. Imagawa, T. Yamada and T. Mukaiyama, Bull. Chem.
Soc. Jpn., 1995, 68, 1455.
17 T. Nagata, K. Imagawa, T. Yamada and T. Mukaiyama, Bull. Chem.
Soc. Jpn., 1995, 68, 3241.
18 M. M. Miller, D. C. Sherrington and S. Simpson, J. Chem. Soc.,
Perkin Trans. 2, 1994, 2091.
19 G. Olason and D. C. Sherrington, Macromol. Symp., 1998, 131, 127.
20 S.-M. Leinonen and D. C. Sherrington, J. Chem. Res. (S), 1999, 572.
21 S.-M. Leinonen, D. C. Sherrington, A. Sneddon, D. McLoughlin,
J. Corker, C. Canevali, F. Morazzoni, J. Reedijk and S. B. D. Spratt,
J. Catal., 1999, 183, 251.
22 D. C. Sherrington, Catal. Today, 2000, 57, 87.
23 L. Canali and D. C. Sherrington, Chem. Soc. Rev., 1999, 28, 85.
24 L. Canali, E. Cowan, H. Deleuze, C. L. Gibson and D. C.
Sherrington, Chem. Commun., 1998, 2561.
25 C. E. Song, E. J. Roh, B. M. Yu, D. Y. Chi, S. C. Kim and K.-J. Lee,
Chem. Commun., 2000, 615.
26 D. C. Sherrington, Chem. Commun., 1998, 2275.
27 B. C. Das and J. Iqbal, Tetrahedron Lett., 1997, 38, 1235.
28 X. Han and Z. Lei, Pure Appl. Chem., 1999, 36, 1337.
29 B. B. Wentzel, M. C. Feiters and R. J. M. Nolte, unpublished results.
Acknowledgements
This research was financed by the Netherlands IOP Foundation
for Innovative Research and the Netherlands Ministry of
Economic Affairs (project no. IKA 94025).
J. Chem. Soc., Perkin Trans. 1, 2000, 3428–3431
3431