Asymmetric alkene epoxidation catalysed by a novel family of chiral metalloporphyrins: Effect of structure on catalyst activity, stability and enantioselectivity

Article abstract of DOI:10.1039/b304212j

A selection of alkenes has been epoxidised with iodosylbenzene, catalysed by three related iron(III) tetraarylporphyrins: 1*, 2* and 3* with four 2,6-di(1-phenylbutoxy)phenyl groups, with one pentafluorophenyl and three 2,6-di(1-phenylbutoxy)phenyl groups and with two pentafluorophenyl and two 2,6-di(1-phenylbutoxy)phenyl groups, respectively. 1* is very sterically hindered and prone to self-oxidation which makes it a relatively poor epoxidation catalyst. Introducing the smaller pentafluorophenyl groups, in place of 2,6-di(1-phenylbutoxy)phenyl, increases catalyst reactivity, stability and selectivity. This change allows easier access of the substrates to the active oxidant and also, by decreasing the electron density on the porphyrin ligand, increases the reactivity of the oxoiron intermediate and its stability towards self-oxidation. A family of five homochiral catalysts, 1, 2 and 3, [the analogues of 1*, 2* and 3*, prepared from (R,R)-2,6-di(1-phenylbutoxy)benzaldehyde] and catalyst 4 with three pentafluorophenyl and one (R,R)-2,6-di(1-phenylbutoxy)phenyl group and 5 the manganese(III) analogueof 3 have been used to epoxidise three prochiral alkenes. All the reactions give low enantioselectivities. Using styrene as the substrate, (S)-styrene epoxide is the major enantiomer obtained with all the catalysts except 1 which leads to the (R)-styrene epoxide being preferred. In contrast cis-hept-2-ene and 2-methylbut-2-ene give the same major epoxide enantiomer with all the catalysts. The dependence of the ee values on catalyst and substrate structure, temperature and solvent is examined and discussed.

Full text of DOI:10.1039/b304212j

View Article Online / Journal Homepage / Table of Contents for this issue
Asymmetric alkene epoxidation catalysed by a novel family of
chiral metalloporphyrins: effect of structure on catalyst activity,
stability and enantioselectivity
John R. Lindsay Smith* and Gloriana Reginato
Department of Chemistry, University of York, York, England YO10 5DD
Received 22nd April 2003, Accepted 30th May 2003
First published as an Advance Article on the web 11th June 2003
A selection of alkenes has been epoxidised with iodosylbenzene, catalysed by three related iron() tetraaryl-
porphyrins: 1*, 2* and 3* with four 2,6-di(1-phenylbutoxy)phenyl groups, with one pentafluorophenyl and three
2,6-di(1-phenylbutoxy)phenyl groups and with two pentafluorophenyl and two 2,6-di(1-phenylbutoxy)phenyl groups,
respectively. 1* is very sterically hindered and prone to self-oxidation which makes it a relatively poor epoxidation
catalyst. Introducing the smaller pentafluorophenyl groups, in place of 2,6-di(1-phenylbutoxy)phenyl, increases
catalyst reactivity, stability and selectivity. This change allows easier access of the substrates to the active oxidant and
also, by decreasing the electron density on the porphyrin ligand, increases the reactivity of the oxoiron intermediate
and its stability towards self-oxidation. A family of five homochiral catalysts, 1, 2 and 3, [the analogues of 1*, 2*
and 3*, prepared from (R,R)-2,6-di(1-phenylbutoxy)benzaldehyde] and catalyst 4 with three pentafluorophenyl and
one (R,R)-2,6-di(1-phenylbutoxy)phenyl group and 5 the manganese() analogue of 3 have been used to epoxidise
three prochiral alkenes. All the reactions give low enantioselectivities. Using styrene as the substrate, (S)-styrene
epoxide is the major enantiomer obtained with all the catalysts except 1 which leads to the (R)-styrene epoxide being
preferred. In contrast cis-hept-2-ene and 2-methylbut-2-ene give the same major epoxide enantiomer with all the
catalysts. The dependence of the ee values on catalyst and substrate structure, temperature and solvent is examined
and discussed.
Introduction
Results and discussion
The substantial success of Jacobsen and Katsuki and their co-
workers,1 with designing effective chiral manganese() salen
catalysts for asymmetric epoxidation of unfunctionalised
alkenes, has provided a strong stimulus for the development of
other catalytic chiral metal ligand systems. In particular, chiral
metalloporphyrins have and continue to be an active field for
catalyst development.2 Since the first report, twenty years ago,
by Groves and Myers3 of enantioselective epoxidation of
styrene using a chiral picket fence porphyrin, a wide variety
of metalloporphyrins with a chiral superstructure has been
prepared and tested.4 The target remains; a simple synthesis
of chiral metalloporphyrins which are robust and efficient
catalysts for asymmetric oxygen-transfer from cheap oxidants,
such as dioxygen, hydrogen peroxide and hypochlorite, to
prochiral substrates.
Systematic studies aimed at optimising the design of new
homochiral porphyrin ligands involve comparing the catalytic
performance of their metal complexes (typically iron, mangan-
ese and ruthenium). The ligands are usually prepared by one of
two approaches: attaching the chiral superstructure to each of
the four atropisomers of a parent tetraarylporphyrin4p or modi-
fying the substituents on a single atropisomer.4o,4q Less com-
monly, a family of closely related tetraarylporphyrins with one
to four chiral aryl groups have been synthesised and studied.5
The latter procedure was used for this study.
Effect of porphyrin structure on catalyst stability and efficiency
in alkene epoxidation
Initially, the catalytic efficiencies of the optically inactive iro-
n() porphyrins 1*, 2* and 3*† in alkene epoxidations were
investigated. These three catalysts, with respectively four, three
and two 2,6-di(1-phenylbutoxy)phenyl groups, were selected to
evaluate the effect of the different steric and electronic environ-
ments of each on the oxidation processes and on the fate of the
catalyst. The epoxidations were carried out in dry dichloro-
methane at room temperature with iodosylbenzene, using
catalyst, oxidant and substrate in the molar proportions 1 : 100
: 1000. Eight alkenes were used as substrates, namely, styrene,
three cyclic alkenes (cyclooctene, cyclohexene and cyclo-
pentene), and four acyclic alkenes (oct-1-ene, cis-hept-2-ene,
2,3-dimethylbut-2-ene and 2-methylbut-2-ene). Subsequently,
four iron() porphyrins 1, 2, 3 and 4 and the manganese()
porphyrin 5 were used for the asymmetric epoxidation of
styrene, cis-hept-2-ene and 2-methylbut-2-ene and the reaction
conditions were systematically varied to improve the enantio-
selectivity of the epoxidations.
In these systems, iodosylbenzene generates the active oxid-
ant, oxoiron() porphyrin π radical cation or oxomanganese()
porphyrin7
which is consumed in three competing pathways,
epoxidation, catalyst destruction and iodosylbenzene dispro-
portionation (illustrated for the iron porphyrins in Scheme 1).
The product distributions and oxidation yields, based on iodo-
sylbenzene, (path a, Scheme 1) were monitored with time by
capillary GC analysis and the final yields were measured after
70–90 min. Despite the large excess of alkene over catalyst,
We have recently described the preparation of an efficient
and short synthetic route to a new family of metalloporphyrins,
with pentafluorophenyl and (R,R)-2,6-di(1-phenylbutoxy)-
phenyl groups on the porphryin meso positions.6 In this paper
we report the use of these to catalyse alkene epoxidation by
iodosylbenzene and show that the position and number of
pentafluorophenyl groups on the macrocycle ring has a marked
effect on the catalyst’s activity, stability and enantioselectivity.
† Catalysts 1*, 2* and 3* are optically inactive analogues of 1, 2 and 3
and were prepared from ( )-1-phenylbutanol.6
T h i s j o u r n a l i s © T h e R o y a l S o c i e t y o f C h e m i s t r y 2 0 0 3
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 2 5 4 3 – 2 5 4 9
2543

View Article Online
Scheme 1
catalyst destruction, by inter- and/or intramolecular oxid-
ations, was an important side-reaction (path b, Scheme 1). The
catalyst’s stability during the epoxidations was determined by
analysing aliquots from the reaction mixture by UV–Vis
spectroscopy. The catalysed disproportionation of iodosyl-
benzene to iodoxybenzene and iodobenzene (path c, Scheme 1)
was monitored, at the end of the reaction, by measuring the
yield of iodoxybenzene by iodometric titration and by compar-
ing the yields of iodobenzene and epoxide (GC analysis). Fig. 1
clearly shows that during the epoxidation of 2,3-dimethylbut-2-
ene with catalyst 1* and iodosylbenzene, the rate of formation
and the yield of iodobenzene are higher than those of 2,3-di-
methylbut-2-ene epoxide. This arises largely from the catalysed
disproportionation of iodosylbenzene. With a catalyst to oxid-
ant ratio of 1 : 100 in the reactions, oxidant consumption arising
from catalyst destruction is unlikely to exceed a few percent.
Similar graphs have been obtained with all the substrates. The
% yield of epoxide, iodobenzene and iodoxybenzene at the end
of each reaction was measured and the oxidant balance was
determined in terms of iodobenzene, [PhI], and oxidant, [O]
(equations 1 and 2, respectively). The oxidant accountability in
the reactions was generally good showing that no significant
oxidation products had been missed (see for example Table 1).
Total % [PhI] = yield PhI (%) ϩ yield PhIO2 (%) (1)
Total % [O] = yield epoxide (%) ϩ 2 × yield PhIO2 (%) (2)
The rates of epoxidation, product yields and catalyst
stability depend on the structure of the porphyrin ligand, thus
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 2 5 4 3 – 2 5 4 9
2544

View Article Online
Table 1 The accountability of iodosylbenzene in epoxidation of cyclohexene, oct-1-ene and 2-methylbut-2-ene catalysed by iron() porphyrin, 2
Substrate
Oxidation yield (%)
Iodobenzene (%)
Iodoxybenzene (%)
[PhI] (%)a
[O] (%)a
Cyclohexene
Oct-1-ene
2-Methylbut-2-ene
81.3b
32.0
80.0
89.0
66.2
92.8
11.2
29.7
3.5
100.2
95.8
96.3
103.7
91.4
87.0
a The estimated error in the accountability is ∼5%. b Total yield of cyclohexene epoxide, cyclohex-2-en-1-ol and cyclohex-2-en-1-one.
Table 2 Yields of epoxide obtained from the corresponding alkenes
using catalysts 1*, 2* and 3* and iodosylbenzene
Yields (%)
1*
2*
3*
Cyclooctene epoxide
Styrene epoxide
Cyclopentene epoxide
Oct-1-ene epoxide
2-Methylbut-2-ene epoxide
2,3-Dimethylbut-2-ene epoxide
Cyclohexene epoxide
cis-2-Heptene epoxide
51.1
48.5
48.9
6.8
45.1
50.5
48.3
32.4
68.9
79.5
48.8
31.9
80.0
62.3
67.9
68.5
100.0
91.5
59.3
67.6
81.6
90.1
81.9
80.3
Fig. 3 Degradation of catalysts 1*, 2* and 3* versus time during the
epoxidation of 2,3-dimethylbut-2-ene with iodosylbenzene.
catalysts, is very dependent on the structure of the macrocyclic
ligand,8 and that iron() tetra(pentafluorophenyl)porphyrin is
a significantly better catalyst than the non-fluorinated ana-
logue, iron() tetraphenylporphyrin.8a It is clear, from the
results presented here, that replacing a 2,6-di(1-phenylbutoxy)-
phenyl group by pentafluorophenyl leads to a marked improve-
ment in catalyst efficiency and stability. The presence of a
pentafluorophenyl group on the porphyrin ring can increase the
catalytic activity in three ways: first, the electron-withdrawing
fluorines activate the electrophilic high-valent oxoiron inter-
mediate.8,9 Secondly, by removing electron density from the
macrocyclic ring, the halogens also make the porphyrin less
susceptible to attack by the active oxidant leading to less
catalyst self-destruction.8,9 Finally, since a pentafluorophenyl
group is considerably less bulky than 2,6-di(1-phenylbutoxy)-
phenyl not only does this allow easier access of the substrate to
the oxoiron centre during the catalytic cycle but it also removes
oxidisable 1-phenylbutoxy groups, reducing intramolecular
catalyst destruction. It is difficult to determine which of these
effects is the most important, however, it is likely that all three
play a role in these oxidations.
Fig. 1 Formation of 2,3-dimethylbut-2-ene epoxide and iodobenzene
and catalyst degradation with time in the epoxidation of 2,3-
dimethylbut-2-ene with catalyst 1* and iodosylbenzene.
The structures of the catalysts were investigated using com-
puter molecular modelling. To simplify these calculations only
the chiral analogues of the optically inactive catalysts 1*, 2*
and 3* (1, 2 and 3) were used (Figs. 4–6). The structures show
the marked difference in access to the metal centre in the three
Fig. 2 Formation of cyclohexene epoxide as a function of time, in the
epoxidation of cyclohexene with iodosylbenzene catalysed by 1*, 2*,
3*.
on changing the catalysts in the order 1*, 2* and 3* the
reactions become faster (see for example the epoxidation of
cyclohexene, Fig. 2), the maximum yields increase (Table 2 and
Fig. 1) and the catalysts become more stable (see for example
the epoxidation of 2,3-dimethylbut-2-ene, Fig. 3). It is well
documented that the reactivity of iron porphyrins, as oxidation
Fig. 4 Molecular models of top and side views of chiral catalyst 1
showing full van der Waals radii. The oxygens are drawn in red, the
nitrogens in blue and the chlorine atom is in green.
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 2 5 4 3 – 2 5 4 9
2545

View Article Online
part from the lower reactivity of terminal alkenes10 but also
from catalyst deactivation, by N-alkylation of one of the pyr-
role rings of the porphyrin macrocycle by the substrate, that
occurs during the oxidation process. To investigate the latter
reaction further, at the end of the epoxidation of oct-1-ene
using catalyst 3*, the reaction mixture was passed through a
short silica gel column (using dichloromethane as eluant) and
the porphyrin fraction was analysed by ESI-mass spectroscopy.
The mass spectrum had a base peak at m/z 1604 (abundance
100%) which corresponds to the molecular structure 6, with
molecular formula of C92H83ClFeF10N4O5. The literature
reports several examples of deactivation of metalloporphyrins
during terminal alkene epoxidation by iodosylbenzene.11 In
these reactions, a pyrrole nitrogen traps an intermediate, divert-
ing it from epoxide to give an N-hydroxyalkyl-derivative of the
porphyrin. This generates a less active catalyst and leads to
demetallation.11 It is likely that the same deactivation of the
catalyst takes place in the analogous reactions of 1* and 2*. In
agreement with this conclusion, each of the reactions with
oct-1-ene shows a large amount of the catalyst degradation
(>80%).
Fig. 5 Molecular models of top and side views of chiral catalyst 2
showing full van der Waals radii. The oxygens are drawn in red, the
nitrogens in blue, the chlorine is in dark green and the fluorines in light
green.
Fig. 6 Molecular models of top and side views of chiral catalyst 3
showing full van der Waals radii. The oxygens are drawn in red, the
nitrogens in blue, the chlorine is in dark green and the fluorines in light
green.
iron() porphyrins. With 1, the catalyst core is completely bur-
ied by the substituents; this provides a very constrained
environment which allows intramolecular catalyst destruction
to compete effectively with alkene epoxidation. Replacing one
of the 2,6-di(1-phenylbutoxy)phenyl groups by pentafluoro-
phenyl (catalyst 2) creates a space on each porphyrin face,
accounting for the increased rate of epoxidation with catalyst
2*. In catalyst 3 the alternating pentafluorophenyl and 2,6-di-
(1-phenylbutoxy)phenyl groups on the porphyrin meso posi-
tions generate a well defined corridor, on each porphyrin face,
for the substrate to approach the metal centre. This results in a
further increase in the rate of epoxidation at the expense of
catalyst destruction.
The competition of epoxidation and catalyst destruction
(Scheme 1) for the oxoiron() porphyrin π radical cation is
well illustrated by catalysts 1*, 2* and 3*. Thus as the catalyst
becomes less involved in self-oxidation reaction paths, the rate
of the epoxidation and the yields increase. This is apparent both
when different substrates are epoxidised by the same catalyst
and when a single substrate is epoxidised by different catalysts.
The latter is well illustrated by the epoxidation of 2,3-di-
methylbut-2-ene, where the increased epoxidation yields that
arise from changing the catalyst from 1* to 2* to 3* (Table 2)
parallel a decrease in catalyst deactivation. Fig. 3 shows that
during the reaction, catalyst 1* is significantly destroyed whilst
the concentration of 2* is only slightly affected and 3* is
unchanged. This trend is observed with all the substrates used
in this study. It appears that alkene epoxidation is particularly
sensitive to steric effects around the porphyrin macrocycle so
that, if the oxoiron centre is not readily accessible to the sub-
strate, the side-reactions (destruction of the catalyst and form-
ation of iodoxybenzene) can compete with the epoxidation.
With all three catalytic systems, particularly those using 1*
and 2*, the yield of epoxide from oct-1-ene compared with
those from the other substrates is low (Table 2). This arises in
The oxygen transfer from the active oxoiron species of
catalyst 1* gives approximately a 50% yield of epoxide with all
the alkenes, except oct-1-ene (Table 2). This lack of discrimin-
ation between the alkenes is unexpected and not observed with
the other two catalysts. With the two less hindered catalysts, 2*
and 3*, where the active oxidant is more accessible to the
substrates, the epoxide yields are higher and show a greater
discrimination in substrate reactivity. The results also reveal an
interesting example of tuning of the catalyst’s properties. With
the least hindered catalyst 3*, as expected the more electron-
rich 2,3-dimethylbut-2-ene gives a higher epoxide yield than
2-methylbut-2-ene,10 however, with the more hindered catalyst
2* the less electron-rich alkene is more reactive. This reversal
of reactivity can be accounted for in terms of steric effects
outweighing electronic preferences, since the less hindered
2-methylbut-2-ene is able to approach the more hindered oxo-
iron active centre of catalyst 2* more readily than 2,3-di-
methylbut-2-ene.
Enantioselective epoxidations
Styrene, cis-hept-2-ene and 2-methylbut-2-ene were used in
enantioselective epoxidations catalysed by the chiral metallo-
porphyrins 1, 2, 3, 4 and 5. These prochiral substrates were
chosen as representative of three different types of alkene.
Styrene has been widely used previously as a substrate to
compare the enantioselective efficiency of different metallo-
porphyrin-catalysed epoxidations,4 whereas aliphatic alkenes
have been much less extensively studied.4n For each epoxidation,
the product distribution was monitored several times, by chiral
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 2 5 4 3 – 2 5 4 9
2546

View Article Online
Table 3 Effect of temperature on the yield and ee value of styrene epoxidation (catalyst 1 and 3) and cis-hept-2-ene (catalyst 2) with iodosylbenzene
in dichloromethane
Substrate
Catalyst
T/ЊC
Epoxide yield(%)
Epoxide ee(%)
Styrene
Styrene
Styrene
Styrene
cis-Hept-2-ene
cis-Hept-2-ene
cis-Hept-2-ene
1
1
3
3
2
2
2
20
0
20
0
20
0
Ϫ20
48.5
39.5
91.5
99.0
68.5
76.7
65.4
16.3
15.0
16.6
23.2
16.4
20.0
21.0
GC analysis, during the course of the reaction and when the
oxidation was complete.
chiral catalyst remained effectively constant (or even increased
slightly). It was expected that catalyst self-oxidation might
involve the chiral 1-phenylbutoxy groups, since heteroatoms on
the chiral appendages of metalloporphyrins are reported to
facilitate the oxidative degradation of the superstructure,2c and
this would lead to less selective iron porphyrin catalysts and to a
decrease in the enantioselectivity as the reactions progressed.
However, this was not observed and we conclude that the main
pathway for loss of 1 and to a lesser extent with 2 is oxidative
destruction of the porphyrin macrocycle.1a,2a,8c,12
2-Methylbut-2-ene was epoxidised using catalysts 1, 2 and 3
and the same major enantiomer (unidentified) was produced
in each reaction (Fig. 9). The small size of the alkyl groups
attached to the double bond makes chiral selectivity with this
alkene extremely challenging and not surprisingly, of the three
substrates studied, 2-methylbut-2-ene gave the lowest ee values.
The most sterically hindered catalyst 1 gave the highest ee value
(10.1%) whereas catalyst 3, with more open access to the active
oxidant, led to negligible asymmetric induction. For this reason
the study was not extended to include catalysts 4 and 5.
The catalysed epoxidations of styrene and cis-hept-2-ene
were initially carried out at room temperature, using iodosyl-
benzene and 1, 2, 3, 4 and 5. The yields for the iron()
porphyrin-catalysed oxidations of both substrates increase on
going from catalyst 1 to 4 (Figs. 7 and 8), as discussed above,
this is attributable to a decrease of steric hindrance around
the catalytic centre and the increased stability of the catalysts
arising from the electron-withdrawing effects of the pentafluoro-
phenyl groups. Low enantiomeric excesses of the product
epoxides were obtained with all the catalysts, with the highest
values for styrene arising from 1 and 3 (16.3 and 16.6%, respect-
ively) and for cis-hept-2-ene from 2 (16.4%) (Figs. 7 and 8). The
most interesting feature of the styrene oxidations is that, with
catalyst 1, the (R)-styrene epoxide is favoured whereas with all
the other catalysts, the (S)-isomer predominates. With cis-hept-
2-ene, all the catalysts gave the same epoxide enantiomer as the
major product although its configuration was not determined.
The manganese() catalyst, 5, with both styrene and cis-hept-
2-ene gave comparable epoxide yields to its iron() analogue,
3, however, it was less stereoselective giving ee values approx-
imately half those from 3.
Fig. 9 Yields and enantiomeric excesses of 2-methylbut-2-ene epoxide
obtained with catalysts 1, 2 and 3 and iodosylbenzene in dichloro-
methane at room temperature.
Fig. 7 Yields and enantiomeric excesses of styrene epoxide obtained
in the epoxidation of styrene by iodosylbenzene catalysed by 1, 2, 3, 4
and 5.
Effect of temperature on asymmetric epoxidation. The effect
of temperature on the yield and enantioselectivity of the
epoxidation of styrene and cis-hept-2-ene was examined using
the most selective catalysts for each substrate (styrene, catalysts
1 and 3 and cis-hept-2-ene, catalyst 2) (Table 3). With catalysts 3
and 2, lowering the temperature from 20 to 0 ЊC resulted in a
higher overall epoxidation yield and a small but significant
increase in the ee values. Improved enantioselectivities were
anticipated based on previous work on epoxidations using high
valent oxometalloporphyrins4i,13 and can be attributed to a
slower more selective epoxidation mechanism which leads to
an increased differentiation between the Re and Si faces of
the prochiral alkene. Interestingly the epoxidation of styrene
using the very hindered catalyst 1 led to a decrease in the overall
yield and no improvement in the enantioselectivity. Lowering
the temperature of the cis-hept-2-ene epoxidation further to
Ϫ20 ЊC led to a further small increase in the ee value but the
overall yield was reduced.
Fig. 8 Yields and enantiomeric excesses of cis-hept-2-ene epoxide
obtained with catalysts 1, 2, 3, 4 and 5 and iodosylbenzene at room
temperature.
To assess whether or not the enantioselectivity of the oxygen
transfer was maintained throughout the epoxidation process,
the products were monitored at selected times during the
course of the reactions. The enantiomeric excesses with each
Solvent dependence of the enantioselective epoxidation of
styrene and cis-hept-2-ene. The influence of solvent on the
catalysed asymmetric epoxidations was examined with styrene
(with 1 and 3) and cis-hept-2-ene (with 2) at 0 ЊC in dichloro-
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 2 5 4 3 – 2 5 4 9
2547

View Article Online
Table 4 Solvent dependence of the yield and enantiomeric excess of epoxidation of styrene (catalysts 1 and 3) and cis-hept-2-ene (catalyst 2) by
iodosylbenzene at 0 ЊC
Substrate
Catalyst
Solvent
Epoxide yield(%)
Epoxide ee(%)
Major enantiomer
Styrene
Styrene
Styrene
Styrene
Styrene
Styrene
Styrene
Styrene
cis-Hept-2-ene
cis-Hept-2-ene
cis-Hept-2-ene
cis-Hept-2-ene
1
1
1
1
3
3
3
3
2
2
2
2
CH2Cl2
Toluene
CH3OH
CH3CN
CH2Cl2
Toluene
CH3OH
CH3CN
CH2Cl2
Toluene
CH3OH
CH3CN
39.5
64.3
4.5
15.0
17.5
0.0
15.5
23.2
9.7
12.9
16.3
21.0
18.5
26.2
18.0
(R)
(R)
68.5
99.0
84.0
67.8
86.5
76.7
69.5
36.7
72.5
(R)
(S)
(S)
(S)
(S)
—
—
—
—
methane, toluene, methanol, and acetonitrile (Table 4).
Although the overall yields and ee values showed significant
solvent dependence, the major enantiomer formed with each
substrate/catalyst combination was unaffected by the changes.
Interestingly no single solvent was optimum for all three sys-
tems. Styrene with 1 gave the best epoxide yields in acetonitrile
or toluene and the highest ee value in the latter. In contrast with
3, dichloromethane led to a virtually quantitative yield of
epoxide and the highest ee value. With cis-hept-2-ene and 2,
methanol gave the highest ee value but the overall yield was less
than half that of the reaction in dichloromethane.
Methanol is not an innocent solvent for these oxidations. It
reacts with iodosylbenzene (Scheme 2)14 giving (hydroxy-
methoxyiodo)benzene (7) and (dimethoxyiodo)benzene (8) and
the former is thought to be the species that transfers oxygen to
the metalloporphyrin to generate the active oxidant.15 Fur-
thermore, it is a competing substrate for the active oxidant
and consequently epoxidation yields in methanol are generally
lower than those in more robust solvents.15 The particularly low
epoxide yield from styrene with catalyst 1 probably arises from
steric hindrance restricting the bulky iodosylbenzene derivative
7 from approaching the metal centre.
metal centre.5 These restrictions can make highly substituted
metalloporphyrins both inefficient and unselective catalysts
whereas, by contrast, analogues with more open access to the
metal centre can be both effective and selective.4n,4p,5
Rose and Collman and co-workers5 synthesised a series of
chiral metallotetraarylporphyrins based on the parent, double
picket fenced iron() tetra-(2,6-di-Mosher’s amidophenyl)-
porphyrin, with one to three of the four chiral 2,6-disubstituted
aryl groups systematically replaced by pentafluorophenyl
rings. Comparison of these as catalysts for the asymmetric
epoxidation of styrene showed that, despite the low ee values
obtained (<6%), there was a trend; the least hindered catalyst
with three pentafluorophenyl substituents gave the highest and
the most hindered parent compound the lowest enantio-
selectivity. It was argued that placing fewer chiral groups on
each face of a porphyrin can, by allowing better access to the
metal centre, generate a more selective catalyst. The ee values
from the present study show that no one catalyst is optimum for
the epoxidation for all three types of alkene used. This
emphasises the importance of substrate–catalyst interactions in
differentiating between the Re and Si faces of the alkene. It also
suggests that to obtain optimum ee values a specific catalyst
may need to be designed for the epoxidation of each class of
alkene.
The low ee values obtained in this study most probably arise
from a lack of structural rigidity of the chiral appendages, lead-
ing to poor spatial definition of the catalytic site. Collman
et al.4n have shown that iron porphyrins with a rigid chiral
superstructure, even with relatively open access to the metal
centre, can be very efficient and selective catalysts for asym-
metric epoxidation. In the present study, attempts to improve
control of the approach of the substrate to the active oxidant
by restricting access to the metal centre, led to lower epoxide
yields and more catalyst destruction. This is well illustrated for
the epoxidation of cis-hept-2-ene where catalysts 1 and 4 gave
effectively the same ee values although there was a marked dif-
ference in yield. With 1 the yield was 32.4% because the active
site of the catalyst is masked by the presence of four bulky
chiral substituents limiting the access of the substrate whereas
with 4, more ready access, led to an 84.9% yield.
Comparison of the results obtained with 3 and 5 shows that
although the manganese catalyst gives similar epoxide yields to
the iron analogue it is less stereoselective. Comparable observ-
ations have been reported previously for the asymmetric epox-
idation of styrene using chiral strapped4i and chiral wall
metalloporphyrins.16 However, this not a general phenomenon
since chiral manganese() porphyrins4 and the extensively
studied chiral manganese() salens1 can be very effective
catalysts for asymmetric alkene epoxidation. We suspect that
the difference we and others have observed will be dependent on
the structure of the porphyrin and may reflect a different
approach of the substrate towards the oxo-metal active oxidant
with the iron and manganese complexes (see for example
reference 13b).
Scheme 2
Previous work on the solvent effects on oxidations with
metalloporphyrin systems has used aprotic solvents and found
that aromatic hydrocarbons, such as benzene or toluene, give
the highest ee values.4i,13 This has been interpreted as arising
from solvent induced changes of the structure of the chiral
cavity surrounding the oxo-metal group of the active oxidant,
leading to greater enantioselectivity. In the present study,
however, the best solvent is dependent on the chiral catalyst and
shows no obvious trends with solvent type or polarity.
General discussion on catalyst design
In designing chiral metalloporphyrin catalysts for asymmetric
alkene epoxidation, Collman and co-workers2a,4n and Campbell
and Kodadek2c have pointed out that, to optimise enantio-
selectivity without losing epoxide yields, there is a balance
between control of the approach of the substrate by the por-
phyrin superstructure and access to the active oxidant. System-
atic studies of the influence of metalloporphyrin structure on
catalytic activity and enantioselectivity have led to the conclu-
sion that the failure of some catalysts arises from the hindrance
of both enantioapproaches of the prochiral substrate to the
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 2 5 4 3 – 2 5 4 9
2548

View Article Online
113, 87–107; (d ) E. N. Jacobsen and M. H. Wu, in Comprehensive
Experimental
Asymmetric Catalysis, eds. E. N. Jacobsen, A. Pfaltz and
H. Yamamoto, Springer-Verlag, Berlin, 1999, pp. 649–675;
(e) T. Katsuki, in Catalytic Asymmetric Synthesis, ed. I. Ojima,
Wiley-VCH, Weinheim, 2000, pp. 287–325.
Instrumental methods
UV–Visible spectra were recorded on a Hewlett Packard model
HP8453 diode array spectrophotometer and analysed using a
PC running Hewlett Packard A.02.05 UV–Vis ChemStation
software.
2 (a) J. P. Collman, X. Zhang, V. J. Lee, E. S. Uffelman and
J. I. Brauman, Science, 1993, 261, 1404–1410; (b) Y. Naruta, in
Metalloporphyrins in Catalytic Oxidations, ed. R. A. Sheldon,
Marcel Dekker, New York, 1994, pp. 241–259; (c) L. A. Campbell
and T. Kodadek, J. Mol. Catal. A: Chem., 1996, 113, 293–310.
3 J. T. Groves and R. S. Myers, J. Am. Chem. Soc., 1983, 105, 5791–
5796.
Isothermal and temperature-programmed gas chromato-
graphy was carried out on an AMS94 gas chromatograph with
a flame ionisation detector and helium as the mobile phase. The
data were analysed on a PC with Jones Chromatography
JCL6000 (revision 005) analytical software. The non-chiral
analyses used an Alltech Carbowax capillary column (30 m, i.d.
0.25 mm, film thickness 0.25 µm) and the chiral separations
employed a Chiraldex γ-cyclodextrin propionyl capillary col-
umn (50 m, i.d. 0.25 mm) or a Supelco β-DEX 120 capillary
column (30 m, i.d. 0.25 mm, film thickness 0.25 µm). The
methods were optimised and calibrated using racemic epoxides.
The (R) and (S) styrene epoxides were identified by comparison
of retention times of authentic samples. The product yields
were quantified using 1,3-dichlorobenzene as the internal
standard. The identities of all the oxidation products were con-
firmed by GC-MS using a VG Autospec S Series A027 mass
spectrometer linked to a Hewlettt Packard 5890 Series 2 gas
chromatograph. The spectra were analysed using a VAX3100
Workstation.
4 (a) D. Mansuy, P. Battioni, J.-P. Renaud and P. Guerin,
Chem. Commun., 1985, 155–156; (b) J. T. Groves and P. Viski, J. Org.
Chem., 1990, 55, 3628–3633; (c) K. Konishi, K.-I. Oda, K. Nishida,
T. Aida and S. Inoue, J. Am. Chem. Soc., 1992, 114, 1313–1317;
(d ) J. P. Collman, X. Zhang, V. J. Lee and J. I. Brauman, Chem.
Commun., 1992, 1647–1649; (e) Y. Naruta, N. Ishihara, F. Tani
and K. Maruyama, Bull. Chem. Soc. Jpn., 1993, 66, 158–166;
( f ) J. P. Collman, V. J. Lee, C. J. Kellen-Yuen, X. Zhang,
J. A. Ibers and J. I. Brauman, J. Am. Chem. Soc., 1995, 117, 692–703;
(g) S. Vilain-Deshayes, P. Maillard and M. Momenteau, J. Mol.
Catal. A: Chem., 1996, 113, 201–207; (h) S. Ini, M. Kapon, S. Cohen
and Z. Gross, Tetrahedron Asymmetry, 1996, 7, 659–662; (i) Z. Gross
and S. Ini, J. Org. Chem., 1997, 62, 5514–5521; (j) J. F. Barry,
L. Campbell, D. W. Smith and T. Kodadek, Tetrahedron, 1997,
53, 7753–7776; (k) R. L. Halterman, S.-T. Jan, H. L. Nimmons,
D. J. Standlee and M. A. Khan, Tetrahedron, 1997, 53, 11257–11276;
(l ) R. L. Halterman, S.-T. Jan, A. H. Abdulwali and D. J. Standlee,
Tetrahedron, 1997, 53, 11277–11296; (m) T.-S. Lai, R. Zhang,
K.-K. Cheung, H.-L. Kwong and C.-M. Che, Chem. Commun.,
1998, 1583–1584; (n) J. P. Collman, Z. Wang, A. Straumanis,
M. Quelquejeu and E. Rose, J. Am. Chem. Soc., 1999, 121, 460–461;
(o) C. Perollier, J. Pecaut, R. Ramasseul and J.-C. Marchon,
Inorg. Chem., 1999, 38, 3758–3759; (p) G. Reginato, L. Di Bari,
P. Salvadori and R. Guilard, Eur. J. Org. Chem., 2000, 1165–1171;
(q) B. Boitrel, V. Baveux-Chambenoit and P. Richard, Eur. J. Inorg.
Chem., 2002, 1666–1672.
The molecular models for the metalloporphyrins 1–5 were
generated using the program Cerius2 (BIOSYM/Molecular
Simulations) and minimised using the Universal force field.17
Materials
All reagents and solvents were used as purchased (Aldrich,
Lancaster) unless otherwise stated. Iodosylbenzene was pre-
pared from iodobenzene diacetate as described previously.14c
The syntheses of the metalloporphyrin catalysts have been
reported.6 The alkenes were purified by passing them through
a short column of activated alumina prior to use and their
purities were checked by GC analysis.
5 E. Rose, M. Soleilhavoup, L. Christ-Tommasino, G. Moreau,
J. P. Collman, M. Quelquejeu and A. Straumanis, J. Org. Chem.,
1998, 63, 2042–2044.
6 S. P. Foxon, J. R. Lindsay Smith, P. O’Brien and G. Reginato,
J. Chem. Soc., Perkin Trans. 2, 2001, 1145–1153.
7 B. Meunier, Chem. Rev., 1992, 92, 1411–1456.
8 (a) C. K. Chang and F. Ebina, J. Chem. Soc., Chem. Commun., 1981,
778–779; (b) P. S. Traylor, D. Dolphin and T. G. Traylor, J. Chem.
Soc., Chem. Commun., 1984, 279–280; (c) M. J. Nappa and
C. A. Tolman, Inorg. Chem., 1985, 24, 4711–4719; (d ) S. Banfi,
F. Montanari and S. Quici, J. Org. Chem., 1989, 54, 1850–1859;
(e) D. Dolphin, T. G. Traylor and L. Y. Xie, Acc. Chem. Res., 1997,
30, 251–259; ( f ) J.-F. Bartoli, V. Mouries-Mansuy, K. Le Barch-
Ozette, M. Palacio, P. Battioni and D. Mansuy, Chem. Commun.,
2000, 827–828.
Oxidation systems
The achiral epoxidations were carried out at room temperature
by adding iodosylbenzene (22 mg, 0.1 mol) to a solution of the
substrate (1 mmol), metalloporphyrin (1 × 10Ϫ6 mol) and 1,3-
dichlorobenzene (0.01 cm3, GC internal standard) in dry di-
chloromethane (3.1 cm3). For chiral epoxidations the quantities
employed were a fifth of those above. The product distributions
and yields were monitored by removing aliquots from the
alkene oxidations, with a syringe at selected time intervals, and
analysing them by GC. The final yields of epoxide, based on
iodosylbenzene, were calculated after 70–90 min.
The catalyst stabilities during the epoxidations were deter-
mined by diluting aliquots (50 µl) from the reaction mixtures
with dichloromethane (3 cm3) and analysing them by UV–Vis
spectroscopy by monitoring the metalloporphyrin Soret band.
The yield of iodoxybenzene was determined by adding
glacial acetic acid (30 cm3), excess of potassium iodide (10%
aqueous solution) and ice to the reaction mixture dissolved in
methanol (20 cm3). After 30 min in the dark with occasional
shaking, the solution was titrated against sodium thiosulfate
solution using starch as indicator.18
9 P. E. Ellis and J. E. Lyons, Coord. Chem. Rev., 1990, 105,
181–193.
10 J. R. Lindsay Smith and P. R. Sleath, J. Chem. Soc., Perkin Trans. 2,
1982, 1009–1015.
11 T. Mashiko, D. Dolphin, T. Nakano and T. G. Traylor, J. Am. Chem.
Soc., 1985, 107, 3735–3736; D. Mansuy, L. Devocelle, I. Artaud
and P. Battioni, Nouv. J. Chim., 1985, 7, 711–716; J. P. Collman,
P. D. Hampton and J. I. Brauman, J. Am. Chem. Soc., 1990, 112,
2977–2986; J. P. Collman, P. D. Hampton and J. I. Brauman, J. Am.
Chem. Soc., 1990, 112, 2986–2998.
12 A. El-Kasmi, D. Lexa, P. Maillard, M. Momenteau and
J.-M. Savéant, J. Am. Chem. Soc., 1991, 113, 1586–1595.
13 (a) R. Zhang, W.-Y. Yu, T.-S. Lai and C.-M. Che,
Chem. Commun., 1999, 409–410; (b) R. Zhang, W.-Y. Yu, H.-Z. Sun,
W.-S. Liu and C.-M. Che, Chem. Eur. J., 2002, 8, 2495–2507.
14 (a) B. C. Schardt and C. L. Hill, Inorg. Chem., 1983, 22, 1563–1565;
(b) D. R. Leanord and J. R. Lindsay Smith, J. Chem. Soc.,
Perkin Trans. 2, 1991, 25–30; (c) P. R. Cooke and J. R. Lindsay
Smith, J. Chem. Soc., Perkin Trans. 2, 1994, 1913–1923.
15 P. Inchley and J. R. Lindsay Smith, J. Chem. Soc., Perkin Trans. 2,
1995, 1579–1587.
16 G. Reginato, unpublished results.
References
1 (a) E. N. Jacobsen, in Catalytic Asymmetric Synthesis, ed. I. Ojima,
VCH, New York, 1993, pp. 159–202; (b) T. Katsuki, Coord. Rev.,
1995, 140, 189–214; (c) T. Katsuki, J. Mol. Catal. A: Chem., 1996,
17 A. K. Rappe, C. J. Casewit, K. S. Colwell, W. A. Goddard and
W. M. Skiff, J. Am. Chem. Soc., 1992, 114, 10024–10035.
18 G. Gray and J. R. Lindsay Smith, unpublished method.
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 2 5 4 3 – 2 5 4 9
2549