Mechanistic studies on the epoxidation of alkenes with molecular oxygen and aldehydes catalysed by transition metal-β-diketonate complexes

Article abstract of DOI:10.1039/a801175c

The scope, mechanism and kinetics of the aerobic epoxidation of alkenes with an aldehyde and substituted β-diketonate-transition metal complexes as catalysts were studied. β-Diketonate complexes of nickel(II) proved to be among the best catalysts for this reaction. The epoxidation is not dependent on substrate concentration and is first order in aldehyde, catalyst concentration and oxygen partial pressure. It was shown by reactivity studies and EPR experiments that the reaction is radical in nature. Additional evidence for this was obtained from stereochemical investigations. The metal catalyst is not only an efficient initiator of the reaction, but is also believed to enhance the reactivity of intermediate species in the oxidation process by allowing these to co-ordinate to the metal center. A mechanism is proposed for the catalytic reaction.

Full text of DOI:10.1039/a801175c

View Article Online / Journal Homepage / Table of Contents for this issue
Mechanistic studies on the epoxidation of alkenes with molecular
oxygen and aldehydes catalysed by transition metal–â-diketonate
complexes†
Bastienne B. Wentzel, Patricia A. Gosling, Martin C. Feiters and Roeland J. M. Nolte*
Department of Organic Chemistry/Nijmegen SON Research Center, University of Nijmegen,
6525 ED Nijmegen, The Netherlands
The scope, mechanism and kinetics of the aerobic epoxidation of alkenes with an aldehyde and substituted
β-diketonate–transition metal complexes as catalysts were studied. β-Diketonate complexes of nickel() proved
to be among the best catalysts for this reaction. The epoxidation is not dependent on substrate concentration
and is first order in aldehyde, catalyst concentration and oxygen partial pressure. It was shown by reactivity
studies and EPR experiments that the reaction is radical in nature. Additional evidence for this was obtained
from stereochemical investigations. The metal catalyst is not only an efficient initiator of the reaction, but
is also believed to enhance the reactivity of intermediate species in the oxidation process by allowing these
to co-ordinate to the metal center. A mechanism is proposed for the catalytic reaction.
Molecular oxygen as a cheap, clean and readily available oxid-
O
O2, (catalyst)
ant has received much attention in recent years.1 Mukaiyama
and co-workers2–7 and others8–11 have reported that molecular
RCHO
RCO2H
oxygen can be used as the terminal oxidant in the epoxidation
of alkenes with an aldehyde or primary alcohol as coreactant
and a metal β-diketonate as a catalyst (Scheme 1). There has
been discussion in the literature about the mechanism of the
‘Mukaiyama’ catalytic system and the role of the transition-
metal catalyst in it, which can be omitted as was shown by
Kaneda et al.12 Since peroxyacids, which are formed in the aut-
oxidation of aldehydes, are powerful epoxidizing reagents, the
reaction in Scheme 1 might proceed through the peroxyacid as
the actual epoxidizing agent. The only role of the transition-
metal catalyst in this scenario is to catalyse the formation of the
peracid as shown in Scheme 2.
Scheme 1
O2
O
O
O
O
RCH
RC OOH
RC OH +
(transition-
metal catalyst)
Scheme 2 Peracid epoxidation mechanism
Another possible mechanism proposed in the literature10 is
the formation of a metal–oxygen complex which reacts to form
an oxometal species, similar to species described for manganese
or vanadium.13,14 In Scheme 3 this mechanism is outlined for a
transition-metal()–β-diketonate complex.
O•
O
RCHO, O2
MII
MIII
MII
+
RCO2H
A combination of the mechanisms in Schemes 2 and 3 was
considered by Nam et al.15 They investigated the ‘Mukaiyama’
system using cyclam-type transition-metal complexes and con-
cluded from indirect evidence that the epoxidation reaction in
their system is radical in nature. The peroxyacid and the oxo-
metal mechanisms in Schemes 2 and 3 were believed to play no
role. An acylperoxy radical, rather than a peroxyacid, was pro-
posed to react with the alkene to form an epoxide. Alternatively,
the acylperoxy radical could co-ordinate to the metal first, and
this complex subsequently epoxidizes the alkene. The same
authors reported shortly after16 that cyclam complexes of NiII
are inhibitors of this radical reaction. These complexes were
believed to be sufficiently good reducing agents to react with an
acylperoxy radical and form an unreactive acylperoxy anion
and a nickel() complex.
O
MIV
Scheme 3 Oxo–metal epoxidation mechanism
β-diketonates, are used. In spite of the extensive discussions in
the literature the role of the metal complex is still not entirely
clear. This issue will be addressed as well.
Experimental
Materials
Dichloromethane was dried over CaCl2, distilled from CaH
under dry nitrogen and stored over molecular sieves.
Acetonitrile was HPLC grade. All other solvents and iso-
butyraldehyde were distilled before use. Oxygen was obtained
from Hoek-Loos and dried over calcium chloride. All alkene
substrates were commercial samples (Aldrich) and were purified
by column chromatography over basic alumina with CH2Cl2 as
eluent or by vacuum distillation. An exception is S-limonene
[1-methyl-4-(1-methylethenyl)cyclohexene] (Aldrich, 96%)
which was used as received. Epoxide products were identified
In the present paper we further explore the scope, kinetics
and mechanism of the epoxidation of alkenes by the ‘Mukai-
yama’ system. The most effective catalysts for this reaction
reported so far, viz. second-row transition-metal complexes of
† Supplementary data available: epoxidation results. For direct elec-
tronic access see http://www.rsc.org/suppdata/dt/1998/2241/, otherwise
available from BLDSC (No. SUP 57387, 3 pp.) or the RSC Library.
See Instructions for Authors, 1998, Issue 1 (http://www.rsc.org/dalton).
Non-SI units employed: atm = 101 325 Pa, G = 10Ϫ4 T.
J. Chem. Soc., Dalton Trans., 1998, Pages 2241–2246
2241

View Article Online
with gas chromatography. The metal complexes were com-
mercial products or were synthesized according to literature
procedures.17,18 Their physical properties were consistent with
their structures and with literature values.17,18
Table 1 Epoxidation of alkene substrates*
Substrate
Conversion (%)
Yield epoxide (%)
α-Pinene
β-Pinene
S-Limonene
Norbornene
Styrene
trans-β-Methylstyrene
Camphene
Allylbenzene
cis-Stilbene
trans-Stilbene
Cyclohexene
Oct-1-ene
93
27
77
100
98
51
68
16
41
6
86
23
72
96
79
47
61
12
36
5
Instrumentation
The GC analyses were performed on a Varian 3700 instrument
with a fused-silica capillary column (25 m length, 25 µm diam-
eter) with a CP-sil stationary phase or a 15 m × 35 µm diameter
column with an FFAP stationary phase. The instrument was
equipped with a flame-ionization detector and coupled to a
Hewlett-Packard 3395 integrator. The UV/VIS spectra were
taken on a Perkin-Elmer Lambda 5 spectrometer, IR spectra on
a Bio-Rad FTS-25 spectrometer, low-temperature EPR spectra
on a Bruker Electron Spin Resonance ER-220D-LR spectro-
meter and room-temperature spectra on a Bruker ESP-300
instrument. The NMR analyses were performed on a Bruker
AC-300 or WH-90 instrument; the solvent was CDCl3.
49
23
45
19
* Reaction conditions: 0.1 mol lϪ1 alkene, 0.3 mol lϪ1 isobutyraldehyde,
1.0 × 10Ϫ3 mol lϪ1 catalyst 1c, 5.0 cm3 CH2Cl2, 1.0 atm O2, 25 ЊC, 4 h.
Table 2 Epoxidation of α-pinene catalysed by substituted nickel()–
β-diketonate complexesa
The catalytic system
Turnover numberb
The standard conditions used in the epoxidation of alkenes by
nickel()–β-diketonate complexes in the presence of an alde-
hyde were as follows. 0.1 mol lϪ1 Alkene, 0.3 mol lϪ1 aldehyde
and 1 mmol lϪ1 catalyst were stirred (1000 revolutions minϪ1)
in CH2Cl2 at 25.0 0.5 ЊC under 1.0 atm of oxygen. Unless
indicated otherwise, kinetic experiments were carried out
with α-pinene as the alkene substrate, isobutyraldehyde as the
coreagent and bis[3-(p-tert-butylbenzyl)pentane-2,4-dionato]-
nickel() 1c as the catalyst. Errors were estimated to be less than
5%. The catalytic reaction was followed by monitoring the dis-
appearance of the substrate and the appearance of product(s)
as a function of time with gas chromatography; the internal
standard was dmf (100.0 µl). Adding dmf in small quantities
did not affect the reaction.
Entry
Complex 1
1
2
3
4
5
6
7
8
9
a
b
c
d
e
f
g
h
i
15
c
21
11
32
39
24
8
1
a Reaction conditions as in Table 1. b Estimated error: 5%. c 74% Yield
in 4 h.
a R = H, R′ = CH3
R′
R′
R′
R′
b R = H, R′ = p-methoxyphenyl
c R = p-tert-butylbenzyl, R′ = CH3
d R = benzyl, R′ = CH3
Determination of CO2 evolved from the epoxidation reaction19
O
O
O
O
R
MII
R
A standard reaction mixture (see above) was prepared with
[Ni(acac)2] 1a as the catalyst and S-limonene as the substrate.
The exhaust gas of the reaction was bubbled though a BaCl2
solution [in 4 mol lϪ1 NaCl (aq)–ethanol–glycol (1.5:2:1 v/v/v)
at pH 11], and the CO2 formed precipitated as BaCO3. The
turbidity of regularly taken samples from the barium solution
was measured with UV spectroscopy at 360 nm, and the
amount of CO2 that had evolved was calculated from a cali-
bration curve.
e R = p-methoxybenzyl, R′ = CH3
f
R = p-nitrobenzyl, R′ = CH3
g R = p-fluorobenzyl, R′ = CH3
h R = ethyl, R′ = CH3
1
i
R = H, R′ = CF3
Electronic and steric effects
In order to study electronic effects a new series of 3-substituted
nickel()–β-diketonate complexes (1a–1h) was synthesized and
tested as epoxidation catalysts. The results for the epoxidation
of α-pinene are shown in Table 2. As can be seen in Table 2
there is a small but significant effect on the turnover rate of
α-pinene by the catalyst when the para position of the benzene
ring of the nickel complex is altered. Substitution of the
benzene ring with an electron-withdrawing nitro group (entry 6)
yields a catalyst with a relatively high turnover number (39)
when compared to 11 turnovers found for the catalyst unsubsti-
tuted at the aromatic ring (entry 4). Remarkably, the catalyst
substituted with an electron-rich methoxybenzyl group (entry
5) gives a high turnover number as well (32). Apparently, elec-
tronic effects do not play a major role in the reaction catalysed
by the nickel() complexes 1.
Results
Scope of the epoxidation reaction
A number of substrates were tested in the epoxidation reaction
using the ‘Mukaiyama’ conditions and compound 1c (MII =
NiII) as the catalyst. The results are collected in Table 1. Substi-
tuted alkenes, especially α-pinene, limonene and norbornene
(bicyclo[2.2.1]hept-2-ene) gave very good yields as was expected
based on other studies (see for example Yamada et al.4 and Fdil
et al.11). Also styrene is a very good substrate for this epoxid-
ation reaction. Doubly substituted and electron-rich alkenes
such as cis-stilbene, cyclohexene, β-pinene, trans-β-methylstyr-
ene and camphene gave poorer but still respectable yields of
epoxides between 23 (β-pinene) and 61% (camphene).
We also tested a variety of other metal complexes and metal
salts in the epoxidation of α-pinene and S-limonene (see SUP
57387). In agreement with literature studies, nickel and cobalt
complexes gave the highest epoxide yields (e.g. Fdil et al.11).
It is noteworthy in this respect that Nam et al.15,16 found nickel
cyclam-type complexes to be inactive in their epoxidation reac-
tions. These complexes were shown to reduce the generated acyl
peroxy radical to the peroxy anion which is inactive as an
oxidant.
Aldehyde reactivity
The reactivity of a variety of aldehydes as coreactants in
Scheme 1 was tested under standard conditions (see SUP 57387).
Straight-chained and branched aldehydes such as pivaldehyde
(68% α-pinene epoxide after 4 h) and isobutyraldehyde (91%
epoxide) were the most active coreactants. Aromatic or con-
jugated aldehydes such as benzaldehyde and cinnamaldehyde
were completely inactive under the reaction conditions. Investi-
gations into a heterogeneous epoxidation system by Laszlo and
2242
J. Chem. Soc., Dalton Trans., 1998, Pages 2241–2246

View Article Online
Table 3 Stereoselectivity of the epoxidation of cis-stilbene with differ-
ent oxidants and nickel complexesa
Catalyst
(%)
Conversion
(%)
Epoxide cis:trans
Entry
(%)
ratio
1
2
Complex 1c
Complex 1c and
pyridine
40
47
36
46
1:13
1:45
3
Complex 1c and
m-ClC6H4CO3Hb
m-ClC6H4CO3Hc
[Ni(salophen)]d
[Ni(tpp)]e
63
65
3:1
4
5
6
7
60
70
80
71
60
53
68
69
100% cis
1:10
1:12
[Ni(O2CMe)2]
1:22
a Reaction conditions as in Table 1. b 0.1 mol lϪ1 m-ClC6H4CO3H
under N2. c 0.1 mol lϪ1 m-ClC6H4CO3H, no catalyst. d H2 salophen =
N,NЈ-Bis(salicylidene)-o-phenylenediamine. e H2tpp = 5,10,15,20-Tetra-
phenylporphyrin.
Fig. 1 Generation of CO2 in the epoxidation of S-limonene. Reaction
conditions: (a) 0.01 mmol [Ni(acac)2] in CH2Cl2 (5 cm3); (b) 0.01 mmol
[Ni(acac)2] and 3 mmol isobutyraldehyde in CH2Cl2 (5 cm3); (c) 1 mmol
S-limonene, 3 mmol isobutyraldehyde, 0.01 mmol [Ni(acac)2], CH2Cl2
(5 cm3), 1 atm O2, 25 ЊC
Levart,20 using kaolinite and an aldehyde in the presence of O2,
yielded similar aldehyde reactivities.
If the acylperoxy radical is the active oxidizing agent, this
radical might yield a carboxyl radical after epoxidation. The
latter radical can decompose into CO2 and an alkyl radical. As
Lassila et al.21 have suggested, epoxidation and decomposition
may occur in a concerted process. A relatively stable alkyl rad-
ical is formed in the case of isobutyraldehyde and pivaldehyde,
whereas the alkyl or aryl radicals generated from the other
aldehydes will be less stable. This decomposition could create a
driving force for the epoxidation reaction.
If decomposition of the carboxyl radical into carbon dioxide
and an alkyl radical plays a major role in the reaction, it should
be possible to detect this by measuring the amount of CO2
evolving from the reaction; CO2 was determined as described in
the Experimental section. The results are shown in Fig. 1. After
3 h 67% of the aldehyde had reacted and only 10% had evolved
as CO2. Isopropyl hydroperoxide might be anticipated as an
oxidation product of isobutyraldehyde.21 Decomposition of
this hydroperoxide would yield acetone or isopropyl alcohol,
neither of which was detected by GC. Finally, only a part of
the converted aldehyde was retrieved as isobutyric acid (approxi-
mately 10% by GC), indicating a difference in reaction mechan-
ism of the present catalytic system and the systems reported by,
for example, Mizuno et al.,22 Yanai et al.10 (the ‘Mukaiyama’
system) and Nam et al.15 (see also Discussion).
Fig. 2 Effect of aldehyde concentration on the initial epoxidation rate
under standard conditions
Kinetics
The order in substrate concentration was determined for
α-pinene in dichloromethane at 25 ЊC using complex 1c as the
catalyst and isobutyraldehyde as the coreactant. For aldehyde:
alkene ratios >2:1 the order was zero. Below this ratio the
reaction was first order.
The order in aldehyde concentration was calculated from the
initial epoxidation rates measured at various aldehyde concen-
trations (0 to 0.6 mol lϪ1 isobutyraldehyde) and the results are
shown in Fig. 2. When the concentration of aldehyde was under
2.0 mol equivalents with respect to the substrate alkene little
epoxide was formed. When it was equal to or greater than this a
first-order dependence on the aldehyde was found. These results
were observed regardless of the concentrations of the reactants.
Therefore, it appears that the aldehyde must be present in the
reaction mixture at a concentration of approximately twice that
of the substrate for epoxidation to occur. The origin of this
effect is not yet clear. The mechanism outlined in Scheme 4 does
not, for instance, explain the need for ca. 2 equivalents of alde-
hyde in the reaction. A tentative explanation might be the
following. For epoxidation to occur the alkene must be close to
the active epoxidizing species, which we propose to be the acyl-
peroxy metal complex. It is conceivable that the alkene first co-
ordinates to the nickel center. In that case its position will be
trans to that of the co-ordinated peroxy radical. When enough
acid has been formed from the aldehyde one of the acetylaceto-
nate ligands of the acylperoxy–nickel–alkene complex can dis-
sociate allowing the alkene to move cis to the active oxidizing
species. This displacement will not be possible when not enough
acid formed by autoxidation of the aldehyde is present, and
thus will not take place at low aldehyde concentrations. For a
Stereochemistry
In order to obtain information about the stereochemistry of the
reaction we studied the epoxidation of cis-stilbene with various
nickel() complexes and several oxidants. The results are shown
in Table 3. From a comparison of the products obtained with
m-chloroperbenzoic acid (entry 4) with the products obtained
with the nickel() catalysts (other entries) it can be concluded
that a free peroxyacid cannot be the main oxidizing species in
our system; in that case the stereochemistry of the epoxide
would have been retained, which is not observed. In the pres-
ence of a small quantity of pyridine the cis:trans ratio was
shifted from 1:13 (entry 1) to 1:45 (entry 2), indicating that,
although the main pathway for epoxidation is not concerted,
a small fraction of the products are formed via a concerted
pathway, which is inhibited by the presence of a co-ordinating
ligand such as pyridine. We found that it is not possible to
induce chirality in the epoxidation products by using chiral
nickel()–β-diketonate complexes such as camphor {(1R)-1,7,7-
trimethylbicyclo[2.2.1]heptan-2-one} and carvone [2-methyl-
5-(1-methylethenyl)-2-cyclohexen-1-one] derived complexes
synthesized by Fdil et al.,11 in agreement with their results.
J. Chem. Soc., Dalton Trans., 1998, Pages 2241–2246
2243

View Article Online
H
OCR
RCHO
NiII
H+
•
OCR
NiII
NiI
R
OCOO•
RCO2H
H+
O
*
NiI
O2
Fig. 3 Influence of the catalyst concentration on the epoxidation rate
under standard conditions
Scheme 4 Proposed mechanism for the epoxidation of alkenes with
O2 and an aldehyde, catalysed by nickel()–β-diketonate complexes;
* = rate-determining step
methylphenol was added to the reaction mixture during the
reaction epoxidation stopped immediately. When it was added
at the beginning no epoxide was formed. These results indicate
that the formation of a radical species in the reaction mixture is
crucial for epoxidation to occur. Furthermore, in the presence
of a radical inhibitor, no conversion of the substrate into other
oxidation products was observed.
When cyclobutanol was used as the substrate in a reaction
with complex 1c as catalyst and isobutyraldehyde as reductant
only 4-hydroxybutyraldehyde was produced, indicating an
oxidizing species of a radical nature as opposed to a two-
electron oxidant.25
related situation involving the dissociation of an acetylaceto-
nate ligand from [Ni(acac)2] see ref. 23.
The UV/VIS experiments revealed that isobutyraldehyde
binds to the nickel center of complex 1c as could be concluded
from the change from pink to green and the disappearance of
the broad band at 520 nm in the spectrum of 1c. From a
UV/VIS titration experiment the binding constant of the 1:1
1c–isobutyraldehyde complex was calculated to be Kb =
0.68 0.08 l molϪ1. The epoxide reaction could be inhibited by
adding competing Lewis bases to the reaction mixture. Reac-
tions carried out in the presence of pyridine (a strong Lewis
base relative to isobutyraldehyde) slowed the reaction and
changed the order in substrate concentration from one to zero
if the aldehyde:alkene ratio was 2:1. This suggests that co-
ordination of the aldehyde to the nickel complex is an import-
ant step in the reaction sequence.
The concentration of the catalyst was varied and the effect on
the initial rate was determined, as shown in Fig. 3. The rate
increases linearly with the concentration of catalyst up until a
concentration of 5.0 × 10Ϫ4 mol lϪ1. At higher than 10.0 × 10Ϫ4
mol lϪ1 (which is 1 mol% with respect to the alkene concen-
tration) the rate decreased dramatically. Since this catalyst (1c)
can only exist in the monomeric form under the reaction condi-
tions used,24 aggregation of the nickel complex into a trimer
(as in the case of 1a) is not a factor in the decrease in activity.
It is possible that an increased amount of nickel catalyst acts as
a radical trapping compound (see for example Nam et al.16),
which would inhibit the reaction by preventing the formation of
the nickel–peroxo radical species or converting it into a non-
radical peroxy anion.
The epoxidation reaction was investigated at two different
oxygen pressures, viz. 0.21 and 1.0 atm. No effect on the select-
ivity for epoxide was found. At both oxygen concentrations the
reaction followed zero-order kinetics in substrate concentration
with rate constants k0 = 1.73 × 10Ϫ5 and 6.32 × 10Ϫ5 mol lϪ1 sϪ1,
respectively. Without O2 the reaction did not proceed. Based on
these data it can be concluded that the reaction is approxi-
mately first order in oxygen concentration. The epoxidation of
α-pinene was carried out at various temperatures between 18
and 32 ЊC and found to follow Arrhenius behavior, with an
activation energy Ea = 48 6 kJ molϪ1. Using the Eyring
relationship the parameters ∆H‡ and ∆S‡ were calculated to be
46 6 kJ molϪ1 and Ϫ116 6 J KϪ1 molϪ1, respectively. The
large negative value of ∆S‡ points to a rigid transition state for
the rate-determining step.
Kaneda et al.12 have found conditions (using temperatures
slightly above ambient) under which a metal catalyst is not
necessary in the epoxidation of alkenes with O2 and aldehyde.
Initiation in the absence of metal complex can only take place
by light. We investigated these conditions and found that
the observations of Kaneda were not valid for our reactions.
The epoxidation under standard conditions (see Experimental
section) did not proceed without a catalyst, even at higher tem-
peratures (80 ЊC). In neat (freshly distilled) isobutyraldehyde
the reaction did proceed but slower than with nickel catalyst
present. For comparison reactions with conventional initiators
were also performed. The exceptionally reactive radical initiator
di-tert-butyl peroxalate26 was tested in the epoxidation of
S-limonene without catalyst under otherwise standard con-
ditions.‡ The reaction proceeded much more slowly than with
1a as the catalyst. The product distribution was identical, i.e.
the cis:trans ratio of the epoxide was in both cases 2:3. The
selectivity for epoxide was, however, much lower: 67% as
opposed to 93% when 1a was used. The epoxidation reaction
did not proceed when either an initiator or a catalyst was
present, although it was determined by iodide–thiosulfate
titration27 that the aldehyde used contained a small amount of
peroxide (ca. 0.5%).
EPR studies using a spin trap
Electron spin spectroscopy using ‘spin traps’ can provide evi-
dence for the presence of radicals in a reaction. We investigated
the epoxidation of different alkenes by the nickel complexes 1a
and 1c with the radical trap 5,5-dimethyl-1-pyrroline N-oxide
2.28
‡ Di-tert-butyl peroxalate was synthesized according to a literature pro-
cedure.26 For use in the epoxidation reaction a 16.4 mmol lϪ1 solution
in CH2Cl2 was prepared, and 5 cm3 of it was used as the solvent for
the reaction. The amount of peroxide present was checked by iodide–
thiosulfate titration.27 Thus, in the reaction mixture 1.5% of peroxide
with respect to the alkene was present.
Probes for a radical reaction mechanism
When a radical trapping compound such as 2-tert-butyl-4-
2244
J. Chem. Soc., Dalton Trans., 1998, Pages 2241–2246

View Article Online
Table 4 The EPR hyperfine splitting constants from spin-trap experiments with complexes 1a and 1c and compound 2 in CH2Cl2 at 25 ЊC
Entry
Sample
No substratea
Isobutyraldehyde
O2 present
aN/G
aHβ/G
aHγ/G
1
2
3
4
5
6
7
8
Yes
Yes
Yes
No
No
Yes
Yes
No
Yes
No
Yes
Yes
Yes
Yes
No
Yes
12.7
13.35 0.17
13.3
9.8
7.37 0.12
8.7
8.4
0.3
1.5
α-Pinenea
13.1
Stilbenea
No signal
12.69 0.35
13.30 0.20
No signal
S-Limoneneb
8.03 0.37
7.16 0.12
0.3 0.1
a Catalyst 1c. b Catalyst 1a.
Hγ
Hγ
Discussion
Me
Me
Me
Me
From our results, and from those in the literature, we may con-
clude that the [Ni(acac)2]–isobutyraldehyde–alkene system is a
very useful catalytic system for the epoxidation of alkenes. The
mechanism of the epoxidation reaction seems to be different
from that of other epoxidations, such as peroxide-initiated
reactions or reactions catalysed by other transition-metal
complexes (e.g. salen or cyclam-type complexes).
First, an oxometal mechanism (Scheme 3) can probably be
ruled out. No spectroscopic evidence was found for oxometal
complexes in our reactions, and β-diketonate complexes do not
form oxo-complexes easily. There is no evidence that NiII, when
co-ordinated by oxygen ligands, is capable of directly binding
molecular oxygen and forming an oxonickel species, based on
the extensive studies carried out in this area.14,32 In addition, as
reported by Nam et al.,15 the epoxidation reaction can proceed
when it is initiated with a perester giving the same product
distributions as the nickel()-catalysed reaction. If an oxometal
species were to play an important role in the epoxidation
reaction the perester-initiated reaction would most probably
yield a different product distribution. Finally, no 4 was detected
by EPR spectroscopy which can be taken as additional evidence
that no oxometal complexes were generated in solution.
A second possibility is a mechanism in which peracid is
formed in situ, which then oxidizes the alkene, as shown in
Scheme 2. However, this would be at variance with the reactiv-
ity of the aldehydes as observed in our investigations as well as
in those of Mukaiyama and co-workers.2–7 Benzaldehyde, which
has been reported to be easily converted into its corresponding
peroxyacid in the presence of a number of transition metals,33–37
is unreactive under the reaction conditions described here.
Examination of the stereochemistry of the products (see Table
3) provides even stronger evidence discounting an in situ
formed peroxyacid as the active epoxidizing agent. Peroxyacids
normally carry out the epoxidation of alkenes by a concerted
oxygen-transfer step,33,34,38 so that the stereochemistry of the
substrate is conserved in the product. We found that with cis-
stilbene the opposite occurs: trans-stilbene oxide is formed
almost exclusively. Thus peroxyacid epoxidation is not a major
oxidation pathway in the system we studied, though it might
very well be in the system Kaneda et al.12 investigated. Their
results regarding the stereochemistry and kinetics of the reac-
tion are reminiscent of an autoxidation process.
The combined results of our kinetic and mechanistic studies
lead us to conclude that the epoxidation reaction is radical in
nature with the formation of a nickel-bound acyl radical as the
first step and the formation of a nickel–acylperoxy intermedi-
ate, which might be cyclic for stability reasons, as the second
important step. When compared to the literature (Nam et al.15),
the most compelling, new evidence for a radical nature of the
reaction intermediates was found with EPR spectroscopy. Our
investigations pointed to the presence of two radical species.
One is formed when the catalyst and isobutyraldehyde are pres-
ent but oxygen is absent. We propose that this is the trapped
acyl radical bound to the nickel center of the catalyst. The
second radical species is observed when oxygen is added to the
R
+
H
O
N
N
Me
Hβ
Me
N
O –
O•
O•
2
3
4
A solution containing compound 2 and complex 1a (MII =
NiII) in dichloromethane was prepared. This did not give any
signal (other than that of the cavity). To it were added the
components of the reaction, viz. isobutyraldehyde and/or
alkene substrate, so that the concentrations were identical to
those of the standard reaction mixture (see Experimental
section). No signal or only very broad signals were observed in
a glass at 15 K. At 298 K signals were found when at least 2, NiII
and aldehyde or alkene were present. No signals were observed
without 2, strongly suggesting the presence of spin adducts 3 in
the EPR-active samples. The nitrogen and hydrogen hyperfine
splittings aN and aH observed for these samples are summarized
in Table 4. Comparison of the aN and aH values with those
reported in the literature28 with benzene as the solvent allows a
tentative identification of some of the spin adducts and there-
fore of the radicals trapped. Keeping the difference in solvent in
mind, the resemblance of the set of hyperfine splittings of the
signals in entry 1 to those reported for a benzoyloxy radical
adduct with 2 in benzene (aN and aHβ of approximately 12.24
and 9.63 G, respectively) points to the trapping of an acyloxy
radical28 in our case.
Our trapping experiment cannot distinguish between an
adduct derived from an acyloxy and an alkoxy radical. Alkoxy
radicals are reported to have an aN value of 13–13.5 G and an
aHα value of 7–8 G.28 The observed splitting parameters in entries
2, 3, 6 and 7 are very similar to these values, even without any
O2 present (entries 2 and 7), indicating the trapping of a radical
derived from aldehyde, as an O- rather than a C-centered radical.
The hyperfine splitting constants of the EPR signals of
the radical adducts in entries 1, 3, 4 and 6 are sufficiently differ-
ent from each other to exclude the possibility that the same
radical, derived, for example, from the acetylacetonate
ligand, is trapped in these cases. The radical in entry 1 is likely to
come from the isobutyraldehyde, i.e. it is probably the trapped
acyl radical or, more likely, the trapped acylperoxy radical. The
signals in entries 3, 4 and 6 might result from the alkene.
In the hematin/cumene hydroperoxide system reported
previously29 no radicals were trapped; all the EPR signals
observed were due only to oxidation of the spin trap, resulting
in 5,5-dimethylpyrrolidin-2-one N-oxyl, 4.30 The possibility that
this compound was formed in our system could be ruled out,
as the aN and aH values reported for this radical in solvents
related in polarity to dichloromethane are different from those
observed by us (benzene, aN = 6.45 and aH = 3.28; chloroform,
aN = 6.58 and aH = 3.60 G30). It should be mentioned here that
it was recently shown31 that 4 is always formed when oxometal
complexes are present as intermediate species, e.g. in the case of
Mn. The absence of 4 in our experiments provides additional
evidence against an oxometal epoxidation mechanism as shown
in Scheme 3.
J. Chem. Soc., Dalton Trans., 1998, Pages 2241–2246
2245

View Article Online
reaction mixture; this may be the nickel–acylperoxy radical.
Both radicals were trapped with 2 as oxygen-centered radicals.
Based on the data presented in the Results section the rate
equation (1) holds, provided that the molar ratio of aldehyde to
under mild conditions. It is not yet clear what the fate of
the aldehyde is, since neither large amounts of carboxylic acid
nor of CO2 could be detected. The proposed mechanism is con-
sistent with the kinetic and EPR data and with the observed
stereochemistry of the reaction.
r = kobs[O2][Sub]0[RCHO][Cat]
(1)
Acknowledgements
substrate is greater than 2:1. Here, kobs is the observed rate
constant, [Sub] the concentration of alkene substrate, [RCHO]
the concentration of isobutyraldehyde and [Cat] the concen-
tration of nickel() complex. The reaction is first order in sub-
strate concentration when the ratio of aldehyde to substrate is
equal to or less than 2:1, changing to zero order when a Lewis
base (pyridine) is added.
The authors would like to thank Gerrit Jansen and Dr. Paul van
Kan for performing the EPR measurements. This research is
financed by the Innovation Oriented research Programmes of
the Ministry of Economic Affairs (project nos. IKA 90037 and
94025). R. J. M. N. and M. C. F. thank B. B. W. and P. A. G. for
their equal contribution to this paper.
The overall rate law that usually applies to the autoxidation
of aldehydes is given by equation (2),39,40 provided that the
References
1 See Proc. 6th Int. Symp. Activation of Dioxygen, J. Mol. Catal. A,
Chemical, 1997, 117.
2 T. Yamada, T. Takai, O. Rhode and T. Mukaiyama, Bull. Chem. Soc.
Jpn., 1991, 64, 2109.
3 T. Takai, E. Hata, T. Yamada and T. Mukaiyama, Bull. Chem. Soc.
Jpn., 1991, 64, 2513.
4 T. Yamada, K. Imagawa and T. Mukaiyama, Chem. Lett., 1991, 1.
5 T. Yamada, K. Imagawa and T. Mukaiyama, Chem. Lett., 1992,
2109.
6 T. Mukaiyama, T. Yamada, T. Nagata and K. Imagawa, Chem.
Lett., 1993, 327.
7 T. Mukaiyama and T. Yamada, Bull. Chem. Soc. Jpn., 1995, 68, 17.
8 R. Irie, Y. Ito and T. Katsuki, Tetrahedron Lett., 1991, 32, 6891.
9 S. I. Murahashi, Y. Oda and T. Naota, J. Am. Chem. Soc., 1992, 114,
7913.
10 K. Yanai, R. Irie, Y. Ito and T. Katsuki, Mem. Fac. Sci., Kyushu
Univ., Ser. C, 1992, 18, 213.
11 N. Fdil, A. Romane, S. Allaoud, A. Karim, Y. Castanet and
A. Mortreux, J. Mol. Catal. A: Chemical, 1996, 108, 15.
12 K. Kaneda, S. Haruna, T. Imanaka and M. Hamamoto,
Tetrahedron Lett., 1992, 45, 6827.
¹
²
ki
¹
(2)
²
r = k ͩ ͪ [In] [RCHO]
2kt
oxygen and aldehyde concentrations are sufficiently high. Here
k is the rate constant of the rate-limiting propagation reaction,
ki that of the initiation reaction, kt that of the termination reac-
tion and [In] is the initiator concentration. The rate-limiting
step is hydrogen abstraction from the aldehyde. We assume that
our epoxidation reaction (Scheme 1) is sufficiently fast as not to
interfere with the autoxidation reaction. Our results do not con-
form to the rate law (2), because the order in catalyst concen-
tration (assuming that its only role is initiating the reaction) is 1
instead of ¹. The rate law (2) depends on the type of termin-
¯
²
ation step that is operative which may explain the dependence
on oxygen pressure40 of our reaction. The role of our metal
catalyst is likely to be more than just an initiator of the reac-
tion. If a radical chain mechanism, initiated by the nickel()
complex, takes place an order of ¹ is expected, according to
¯
²
13 D. Srinivasan, P. Michaud and J. K. Kochi, J. Am. Chem. Soc., 1986,
108, 2309.
14 R. H. Holm, Chem. Rev., 1987, 87, 1401.
15 W. Nam, H. J. Kim, S. H. Kim, R. Y. N. Ho and J. S. Valentine,
Inorg. Chem., 1996, 35, 1045.
equation (3) where [Ni(acac)2] is taken as an example.
II
I
ؒ
[Ni (acac)2] → [Ni (acac)] ϩ acac
(3)
16 W. Nam, S. J. Baek, K. A. Lee, B. T. Ahn, J. G. Muller, C. J. Burrows
and J. S. Valentine, Inorg. Chem., 1996, 35, 6632.
17 D. P. Graddon, Coord. Chem. Rev., 1969, 4, 1.
18 J. P. Fackler, jun., Prog. Inorg. Chem., 1966, 7, 361.
19 A Textbook of Quantitative Inorganic Analysis, ed. A. I. Vogel,
Longmans, New York, 3rd edn., 1961.
20 P. Laszlo and M. Levart, Tetrahedron Lett., 1993, 34, 1127.
21 K. R. Lassila, F. J. Waller, S. E. Werkheiser and A. L. Wressell,
Tetrahedron Lett., 1994, 35, 8077.
22 N. Mizuno, H. Weiner and R. G. Finke, J. Mol. Catal. A: Chemical,
1996, 114, 15.
23 R. J. M. Nolte and W. Drenth, Recl. Trav. Chim. Pays-Bas, 1973, 92,
788.
The mechanism we propose for the epoxidation of substi-
tuted alkenes with molecular oxygen and isobutyraldehyde,
catalysed by nickel()–β-diketonate complexes, is a catalytic
cycle in which the active oxidizing species is an acylperoxy
radical which stays bound to the metal complex for stabiliz-
ation (Scheme 4). Based on the observed rate equation, we may
tentatively conclude that the rate-limiting step is the formation
of the nickel acylperoxy species, formally a nickel() species (see
Scheme 4). The products evolving from the aldehyde are small
amounts of carboxylic acid and CO2 (both ca. 10% with respect
to converted aldehyde). Other aldehyde oxidation products
could not be identified. It should be noted that Scheme 4 prob-
ably is a simplified picture of the actual process. Studies further
to unravel the details of the reaction are currently in progress.
The role of the metal complex is to promote both hydrogen
abstraction from the aldehyde and acceleration of the oxidation
reaction. The nickel() complex is proposed to take up an elec-
tron from the co-ordinated aldehyde which then loses a proton.
Alternatively, one could imagine the formation of a nickel()
hydride complex. These possibilities are currently under
investigation.
24 J. P. Fackler, jun. and F. A. Cotton, J. Am. Chem. Soc., 1961, 83,
3775.
25 J. Rocek and A. E. Rodkowsky, J. Am. Chem. Soc., 1973, 95, 7123.
26 N. A. Milas and P. C. Panagiotakos, J. Am. Chem. Soc., 1946, 68, 534.
27 P. D. Bartlett, E. P. Benzing and R. E. Pincock, J. Am. Chem. Soc.,
1960, 82, 1762.
28 E. G. Janzen and C. A. Evans, J. Magn. Reson., 1973, 91, 510.
29 R. A. Floyd and L. M. Soong, Biochem. Biophys. Res. Commun.,
1977, 74, 79.
30 H. G. Aurich and J. Trosken, Liebigs Ann. Chem., 1971, 745, 159.
31 T. Sciarone and J. Reek, unpublished work.
32 J. S. Valentine, Chem. Rev., 1973, 73, 235.
33 D. Swern, Chem. Rev., 1949, 45, 1.
34 D. Swern, Org. React., 1953, 378.
Conclusion
35 W. P. Jorissen and P. A. A. van der Beek, Recl. Trav. Chim. Pays-Bas,
1926, 45, 245.
36 P. A. A. van der Beek, Recl. Trav. Chim. Pays-Bas, 1928, 47, 286.
37 T. Mlodnicka, J. Mol. Catal. A: Chemical, 1986, 36, 205.
38 D. Swern and T. W. Findley, J. Am. Chem. Soc., 1950, 72, 4315.
39 Free Radical Chain Reactions, ed. E. S. Huyser, Wiley-Interscience,
New York, 1970.
We have proposed further details of the mechanism for the
epoxidation of alkenes with molecular complex and an alde-
hyde, catalysed by β-diketonate complexes of NiII. It is shown
that the mechanism is radical in nature, with the metal complex
acting as an initiator of the reaction and a promoter of the
oxidation. Our mechanism is in this respect different from that
of Mukaiyama and others. The catalytic system is very useful to
prepare a variety of multiply substituted epoxides from alkenes
40 L. Bateman, Q. Rev. Chem. Soc., 1954, 8, 147.
Received 10th February 1998; Paper 8/01175C
2246
J. Chem. Soc., Dalton Trans., 1998, Pages 2241–2246