Remarkable Effects of Metal, Solvent, and Oxidant on Metalloporphyrin-Catalyzed Enantioselective Epoxidation of Olefins

Article abstract of DOI:10.1021/jo970463w

Three metal complexes of one particular homochiral porphyrin were investigated as catalysts for enantioselective epoxidation of unfunctionalized olefins under various reaction conditions. Much better results were obtained with the iron and ruthenium compl

Full text of DOI:10.1021/jo970463w

5
514
J . Org. Chem. 1997, 62, 5514-5521
Rem a r k a ble Effects of Meta l, Solven t, a n d Oxid a n t on
Meta llop or p h yr in -Ca ta lyzed En a n tioselective Ep oxid a tion of
Olefin s
Zeev Gross* and Santiago Ini
Department of Chemistry, Technion-Israel Institute of Technology, Haifa 32000, Israel
Received March 13, 1997X
Three metal complexes of one particular homochiral porphyrin were investigated as catalysts for
enantioselective epoxidation of unfunctionalized olefins under various reaction conditions. Much
better results were obtained with the iron and ruthenium complexes than with the manganese
derivative. The absence of any effect of amines on the iron porphyrin-catalyzed reaction in benzene,
as well as the superior results in aromatic as opposed to both more and less polar nonaromatic
solvents, suggest that specific association of aromatic molecules to the metalloporphyrin affects its
solution structure. Strong evidence for the involvement of active oxidants that are more selective
than trans-dioxoruthenium(VI) porphyrin is provided by the significant effect of primary oxidants
on the ruthenium porphyrin-catalyzed reactions. Preliminary results with the iron complex of an
only slightly modified porphyrin under the optimized reaction conditions found in this study resulted
in epoxidation of styrene and 4-chlorostyrene to their epoxides with enantiomeric excesses identical
to the best ever reported. These results were obtained with an unprecedented large number of
catalytic turnovers, requiring only 0.01 mol % of catalyst.
In tr od u ction
Enantioselective epoxidation of olefins by homochiral
used as oxidants, as well as the relatively low price of
the catalyst.6 The easy preparation of a large variety of
chiral salen complexes allowed systematic variation of
the steric and electronic environment.7 The major limi-
tation of these catalysts is their quite modest catalytic
activity, usually in the range of 80-120 turnovers.
Metalloporphyrin catalysts are more promising in this
metal complexes is an important subfield of catalytic
asymmetric synthesis, with relevance to organic chem-
1
istry, medicine, and industry. Useful levels of asym-
metric induction of allylic alcoholssproduction of the
corresponding epoxy alcohols with enantiomeric excesses
5
6
aspect, as extremely large turnover numbers (10 -10 )
(
ee’s) greater than 90%swere first achieved in 1980 by
8
have been achieved with achiral derivatives.
2
Katsuki and Sharpless, with significant improvements
and enlargement of the scope of the reaction over the
years.3 An intrinsic limitation of this system is the
essential presence of the hydroxyl group in the organic
substrate, required for coordination to the metal center
of the catalyst. The ultimate challenge in this field
remains enantioselective epoxidation of unfunctionalized
olefins, in which chiral recognition is based solely on
nonbonding interactions between the prochiral substrate
and the homochiral catalyst. The most important early
contribution in this respect was provided by Groves and
The quite extensive research of metalloporphyrin-
catalyzed enantioselective epoxidation of olefins has been
concentrated on the superstructures of the iron(III) and
manganese(III) catalysts and the role of added pyridine
or imidazole ligands.9 Despite considerable efforts, high
ee’ss89% and 96% for 2-nitrostyrene and 3,5-dinitrosty-
rene, respectively, with Naruta’s catalyst10 and 88% for
1
1
,2-dihydronaphthalene with Collman’s catalyst (Scheme
1
1
)
sare the exception rather than the rule. We have
recently prepared a new homochiral porphyrin 1-H
2
(
Scheme 2),12 whose structural features are based on the
4
Meyers in 1983, who extended the principles of cyto-
catalysts of Naruta and Collman. It contains threitol
chiral units as in Collman’s porphyrin, which, however,
cover both faces of the porphyrin plane as in Naruta’s
and other catalysts. Utilization of the ruthenium com-
plex of that porphyrin as a catalyst for the epoxidation
of styrenesthe first example of epoxidation catalysis by
any chiral ruthenium porphyrinsrevealed promising
results.13 We now report our results for the chloroiron-
chrome P-450 mimicking synthetic metalloporphyrin
catalysts to chiral derivatives. Thus, enantiomerically
enriched epoxides were formed in the reaction of simple
olefins with mild oxidants in the presence of catalytic
amounts of a homochiral iron(III) porphyrin. Major
improvements were provided in 1990 by J acobsen and
5
co-workers, who discovered that replacing the metal-
loporphyrins by homochiral manganese(III) salen com-
plexes resulted in much higher enantioselectivities.
Additional advantages were that cheaper and more
available oxidants such as commercial bleach could be
(
(
III), chloromanganese(III), and trans-dioxoruthenium-
VI) complexes of porphyrin 1-H : 1-Fe(Cl), 1-Mn(Cl),
2
(
(
6) Zhang, W.; J acobsen, E. N. J . Org. Chem. 1991, 56, 2296.
7) J acobsen, E. N. In Catalytic Asymmetric Synthesis; Ojima, I.,
*
To whom correspondence should be addressed. Tel.: (int. code)-
72-4-8293954. Fax: (int. code)-972-4-8233735. E-mail: chr10zg@tx.
technion.ac.il.
Ed.; VCH: New York, 1993; pp 159-202.
9
(8) Mansuy, D. Coord. Chem. Rev. 1993, 125, 129.
(9) Collman, J . P.; Zhang, X.; Lee, V. J .; Uffelman, E.; Brauman, J .
I. Science 1993, 261, 1404.
X
Abstract published in Advance ACS Abstracts, J uly 15, 1997.
1) Catalytic Asymmetric Synthesis; Ojima, I., Ed.; VCH: New York,
993.
(
(10) Naruta, Y.; Ishihara, N.; Tani, F.; Maruyama, K. Bull. Chem.
Soc. J pn. 1993, 66, 158.
1
(
(
2) Katsuki, T.; Sharpless, K. B. J . Am. Chem. Soc. 1980, 102, 5974.
3) Sharpless, K. B. In Catalytic Asymmetric Synthesis; Ojima, I.,
(11) Collman, J . P.; Lee, V. J .; Kellen-Yuen, C. J .; Zhang, X.; Ibers,
J . A.; Brauman, J . I. J . Am. Chem. Soc. 1995, 117, 692.
(12) Ini, S.; Kapon, M.; Cohen, S.; Gross, Z. Tetrahedron: Asymmetry
1996, 7, 659.
(13) Gross, Z.; Ini, S.; Kapon, M.; Cohen, S. Tetrahedron Lett. 1996,
37, 7325.
Ed.; VCH: New York, 1993; pp. 101-158.
(
(
4) Groves, J . T.; Myers, R. S. J . Am. Chem. Soc. 1983, 105, 5791.
5) Zhang, W.; Loebach, L.; Wilson, S. R.; J acobsen, E. N. J . Am.
Chem. Soc. 1990, 112, 2801.
S0022-3263(97)00463-5 CCC: $14.00 © 1997 American Chemical Society

Metalloporphyrin-Catalyzed Epoxidation of Olefins
J . Org. Chem., Vol. 62, No. 16, 1997 5515
Sch em e 1. Str u ctu r es of th e Most
En a n tioselective Ca ta lysts Utilized by th e Gr ou p s
of Na r u ta a n d Collm a n
Ta ble 1. En a n tiom er ic Excesses (ee) a n d Ch em ica l
Yield s in Ep oxid a tion of Styr en e by Iod osylben zen e in
th e In d ica ted Solven ts, Ca ta lyzed by 1-F e(Cl), 1-Mn (Cl),
a n d 1-Ru (CO)a
solvent
CH2Cl2
benzene
catalyst
% ee (% yield)
% ee (% yield)
1
1
1
-Mn(Cl)
-Fe(Cl)
-Ru(CO)
6 (38)
13 (46)
5 (12)
10 (77)
44 (65)
42 (47)
a
2
5 °C, 0.5 M styrene, styrene:iodosylbenzene:catalyst ) 1000:
1
00:1, chemical yields relative to iodobenzene.
on nonbonding interactions of the substrate with the
chiral moieties during its approach to the oxometal bond.
Thus, it is necessary to assure that the oxometal bond is
formed inside the cavity created by the chiral units. This
prerequisite was achieved in monofaced iron and man-
ganese porphyrins such as Collman’s porphyrin by ap-
plying bulky amines or phenoxides that coordinate to the
metal at the opposite and less sterically hindered face.
But since for ruthenium porphyrins the active form of
1
5
the catalyst has a trans-dioxoruthenium(VI) structure,
Sch em e 2. Str u ctu r es of th e Tw o Hom och ir a l
Meta llop or p h yr in s a n d Th eir Meta l Com p lexes
Utilized in th e Cu r r en t Stu d ies
both faces of the porphyrin plane must be identical. For
that purpose, we have prepared the new homochiral
porphyrin 1-H
discussed in our earlier paper,12 the etherfication of
,10,15,20-tetrakis(2,6-dihydroxyphenyl)porphyrin with
2
, in which both faces are the same. As
5
the commercially available (S,S)-(-)-1,4-di-O-tosyl-2,3-
O-isopropylidene-L-threitol proceeded in high yields.
Purification was also very easy since only one of the two
possible isomers was formed. The three required metal
complexess1-Fe(Cl), 1-Mn(Cl), and 1-Ru(O)
2
in Scheme
with FeCl , MnCl
3
(CO)12, respectively, by standard methods with
2
swere prepared by reaction of 1-H
2
2
2
,
1
6
and Ru
2
and 1-Ru(O) , respectively (Scheme 2). This paper is
surprising ease and high yields, probably because of the
very large deviation from planarity of the free base. In
addition, the acquired X-ray crystal structures of both
dedicated to the elucidation of the effect of metals,
primary oxidants, and solvents on the enantioselectivity
of epoxidation. The examinations were devoted mostly
to styrene, whose enantioselective epoxidation still re-
mains a significant challenge: the highest reported ee
of styrene oxide by any metalloporphyrin-based system
is 69%; with chiral salen complexes usually only up to
1
2
1
-H
2
and its ruthenium(II) carbonyl complex 1-Ru(CO)
reveal the chiral cavities around the center of the
porphyrin.1
2,13
Although variations of the porphyrin’s
superstructure were not part of the main research
program, we have prepared one additional porphyrin and
its iron(III) complex 2-Fe(Cl) by similar synthetic routes.
In these derivatives, the 1,3-dioxolane rings carry 2,2-
diphenyl substituents instead of the 2,2-dimethyl groups
7% ee are obtained.7 The present results reveal that
enantioselectivity is sensitive to the identity of the metal
Ru g Fe . Mn), the solvent (aromatic > chlorinated),
5
(
and for the ruthenium complex also to the primary
oxidant (N-oxides > iodosylarenes). Up to 6000 catalytic
turnovers were achieved for epoxidation of olefins with
the iron complex in aromatic solvents, without a decrease
in enantioselectivity. We trust that these and other
observations of the current studies will have a significant
impact on the enantioselective epoxidation of olefins by
metal complexes of other homochiral porphyrins as well.
This is demonstrated by preliminary results for a slightly
modified iron(III) porphyrin, 2-Fe(Cl), which catalyzes
the epoxidation of styrene to its oxide with 68% ee.
2
in 1-H (Scheme 2).
Ep oxid a tion s: Meta l a n d Solven t Effects. The
enantioselective epoxidation of styrene by iodosylbenzene
in the presence of catalytic amounts of the three metal
complexes 1-Fe(Cl), 1-Mn(Cl), and 1-Ru(CO) were first
compared in the most commonly used solvent, CH
(Scheme 3). The results, which are summarized in Table
1, show that in CH Cl the enantiomeric excesses (ee’s)
2 2
Cl
2
2
of styrene oxide were very low for all complexes and that
both the chemical yields and the ee’s decreased in the
order of Fe > Mn > Ru. Higher chemical yields were
obtained in benzene for all three complexes, accompanied
by significantly larger ee’s for the 1-Fe(Cl)- and 1-Ru-
(CO)-catalyzed reactions, but not for that of 1-Mn(Cl):
Resu lts
P r ep a r a tion of th e Ca ta lysts. The active oxidants
in metalloporphyrin-catalyzed epoxidations are high-
valent oxometal species,14 in which the oxygen atom that
is transferred to the substrate is located above one of the
porphyrin’s faces. Enantioselective recognition is based
the solvent change from CH Cl2 to benzene induced
2
changes in the ee’s from 13% to 44% for 1-Fe(Cl) and
from 5% to 42% for 1-Ru(CO). The 1-Fe(Cl)-catalyzed
(
15) Groves, J . T.; Quinn, R. Inorg. Chem. 1984, 23, 3844; J . Am.
(14) McMurry, T. J .; Groves, J . T. In Cytochrome P-450: Structure,
Chem. Soc. 1985, 107, 5790.
Mechanism, and Biochemistry; Ortiz de Montellano, P. R., Ed.; Plenum
Press: New York, 1986; Chapter 1.
(16) Buchler, J . W. The The Porphyrins; Dolphin, D., Ed.; Academic
Press: New York, 1978; Chapter 10.

5
516 J . Org. Chem., Vol. 62, No. 16, 1997
Sch em e 3. Ca ta lytic En a n tioselective Ep oxid a tion of Styr en e Der iva tives
Gross and Ini
Ta ble 2. En a n tiom er ic Excesses (ee) a n d Ch em ica l
Yield s in 1-F e(Cl)-Ca ta lyzed Ep oxid a tion of Styr en e by
a
Iod osylben zen e in Va r iou s Solven ts
solvent
CH3CN
CH2Cl2
benzene
toluene
CCl4
ETb
46.0
10
46
41.1
13
46
34.5
44
65
33.9
44
54
32.5
36
78
%
%
ee
yield
a
b
Same reaction conditions as in Table 1. Solvent polarizability
2
6
parameter.
reaction was further examined in a variety of solvents
Table 2). It is clear that in the aromatic solvents
benzene and toluene the ee’s (44% ee) are higher than
in both the more polar acetonitrile (10% ee) and CH Cl
13% ee) and the less polar CCl (36% ee). Similarly, for
(
2
2
(
4
the 2-Fe(Cl)-catalyzed epoxidation of styrene, while the
ee was 59% in benzene, only 31% and 38% ee were
2 2
obtained in CH Cl and heptane, respectively.
F igu r e 1. Effect of temperature on the chemical yields and
enantiomeric excesses (ee) in 1-Fe(Cl)- and 2-Fe(Cl)-cata-
lyzed epoxidation of styrene by iodosylbenzene.
Effect of Am in es. In many cases, inclusion of amine
ligands in the reaction mixtures was found to have a
large positive effect on both the chemical yields and the
enantioselectivities.9 For the present cases, no enhanced
enantioselectivity was found for the 1-Mn(Cl)-catalyzed
reactions in both CH
2 2
Cl and benzene solutions or for the
1
-Fe(Cl)-catalyzed reactions in benzene (Table 3). The
very low yields in the presence of unsubstituted imidazole
are reasonably accounted for by bis-coordination to the
metal. For the 1-Fe(Cl)-catalyzed reactions in CH
Table 3), almost all pyridine ligands had a positive effect
on the enantioselectivity, most pronounced for pyridine,
,4-dimethylpyridine, and quinoline. Still, the 36% ee
for 1-Fe(Cl) in the presence of 2,4-dimethylpyridine in
CH Cl is lower than the 44% ee for the same catalyst in
benzene without added amine. Finally, no enhanced
enantioselectivities were found in CH Cl with the im-
2 2
Cl
(
2
2
2
F igu r e 2. Time-dependent changes in chemical yields and
enantiomeric excesses (ee) in 1-Fe(Cl)-catalyzed epoxidation
of styrene by iodosylbenzene at 20 °C.
2
2
idazole ligands, and the chemical yields were much lower
than without ligand.
Tem p er a tu r e Effect. The effect of temperature on
the chemical yields and the enantioselectivity of the
epoxidation of styrene was examined for the reactions
catalyzed by both 1-Fe(Cl) and 2-Fe(Cl) in toluene. The
results presented in Figure 1 show that in both cases the
highest ee’ss57 and 68%, respectivelyswere obtained at
based systems relative to salen complexes is the much
higher turnover numbers obtainable with the former
catalysts. In many cases, however, the enantioselectivity
decreases with time due to chemical modification of the
catalyst. This aspect was studied with catalyst 1-Fe-
(Cl). Results are shown in Figure 2 for the reaction of
styrene with iodosylbenzene in benzene at 23 °C in the
presence of 0.01 mol % of 1-Fe(Cl). While the chemical
yield increased with time, the ee remained constant at
48 ( 2%. Furthermore, the final yield of 60% corresponds
to 6000 turnovers, which to our knowledge is the highest
ever reported number for enantioselective epoxidation of
olefins. In addition, the 50% yield after 30 min corre-
sponds to 167 turnovers/min. Thus, reduction of the
catalyst’s concentration from 1.0 to 0.1 and even 0.01 mol
% had practically no effect on either the chemical yields
(64, 67, and 60%, respectively) or the ee’s (44, 42, and
46%, respectively).
-
20 °C. Fortunately, at this temperature the chemical
yields were still quite reasonable, 47% and 59%, respec-
tively. The same temperature effectsincreased enantio-
selectivity with only slightly lower chemical yieldsswas
also displayed by other olefins, as shown in Table 5 for
2
-Fe(Cl) catalysis at 23 °C and -20 °C. It must be
emphasized that ee’s of 68% and 70% for epoxidation of
styrene and 4-chlorostyrene, respectively, are practically
identical to the best ever reported results (69% and 70%
ee, respectively) by metalloporphyrin catalysis.11 In
addition, these results were obtained with an extremely
low concentration of catalyst, 0.01 mol % .
Ca ta lytic Tu r n over s. As already mentioned in the
introduction, the main advantage of metalloporphyrin-
Effect of Oxid a n t. Similar ee’s were obtained for
epoxidation of styrene under 1-Fe(Cl) catalysis with two

Metalloporphyrin-Catalyzed Epoxidation of Olefins
J . Org. Chem., Vol. 62, No. 16, 1997 5517
Ta ble 3. Effect of Am in e Liga n d s on th e En a n tiom er ic Excesses (ee) a n d Ch em ica l Yield s in Ep oxid a tion of Styr en e by
Iod osylben zen e, Ca ta lyzed by 1-F e(Cl) a n d 1-Mn (Cl) in CH 2Cl2 a n d Ben zen ea
Ta ble 4. En a n tiom er ic Excesses (ee) a n d Ch em ica l
Yield s in 1-Ru (O)2-Ca ta lyzed Ep oxid a tion of Styr en e by
a
Va r iou s Oxid a n ts in Ben zen e
NO2
CH3
Cl
N
O
Cl
N
O
N
O
oxidant none
IO
IO
%
%
ee
35b
42
47
29
6
50c
20
49
3
54
0.4f
d
e
yield 21
a
2
5 °C, 2 h, 0.165 M styrene, styrene:oxidant:catalyst ) 330:
3
30:1. Yields determined by GC integration against external
standard, unless otherwise noted. Stoichiometric epoxidation,
b
F igu r e 3. Hammett plot of the enantiomeric excesses (ee) in
-catalyzed epoxidation of para-substituted styrenes
(substituents are indicated) by 2,6-dichloropyridine N-oxide.
1
.45 µmol of 1-Ru(O)2 and 9 µmol of styrene-d8 in 1 mL of
1
-Ru(O)
2
c
d
benzene-d6. 57% ee and 15% chemical yield at 15 °C. Deter-
mined by GC integration against iodobenzene. The low yields are
probably due to the insolubility of this oxide. The low yields are
due to inhibition by the produced amine.
e
f
3
4
on styrene, several substituted derivatives were also
investigated. The results for catalysis by 1-Fe(Cl) and
2-Fe(Cl), which are summarized in Table 5, clearly show
that larger enantioselectivities are obtained for the less
oxidants, iodosylbenzene and iodosylmesitylene. This is
in contrast with the results for epoxidation of styrene by
iodosylbenzene, iodosylmesitylene, and three different
aromatic N-oxides in the presence of catalytic amounts
reactive, electron-poor olefins, in accord with studies of
other research groups.9
-11
The low enantioselectivities
of 1-Ru(O)
2
in benzene, which are presented in Table 4.
obtained for the more reactive cis-â-methylstyrene and
1,2-dihydronaphthalene probably also fall into this cat-
egory. The enantioselectivities of epoxidation of the ring-
substituted styrenes was also examined for the 1-Ru-
The large difference between iodosylbenzene and iodo-
sylmesitylene, the significantly higher ee’s obtained with
all N-oxides, and the small differences between them
clearly indicate that 1-Ru(O) is not the sole reactive
2
intermediate in these systems. This is also emphasized
2
(O) -catalyzed reaction with 2,6-dichloropyridine N-oxide
and iodosylbenzene. These very different results are
shown as Hammett plots in Figures 3 and 4, respectively.
While in the reactions with 2,6-dichloropyridine N-oxide
a linear plot was obtained, the Hammett plot describing
the reactions with iodosylbenzene is clearly curved.
by the stoichiometric epoxidation of styrene by 1-Ru-
(
O)
Effect of Su bstr a tes a n d LF ER Rela tion sh ip s.
Although the emphasis of the present studies was placed
2
, which produced styrene oxide with only 35% ee.

5
518 J . Org. Chem., Vol. 62, No. 16, 1997
Gross and Ini
Ta ble 5. En a n tiom er ic Excesses (ee) a n d Ch em ica l Yield s in Ep oxid a tion of Ar om a tic Olefin s, Ca ta lyzed by 1-F e(Cl),
-F e(Cl), a n d 1-Ru (O)2
2
catalyst
T
substrate
1-Fe(Cl)
2-Fe(Cl)
23 °C
2-Fe(Cl)
-20 °C
1-Ru(CO)
25 °C
1-Ru(O)2
25 °C
25 °C
a
ee (yield)a
ee (yield)b
ee (yield)c
d
ee (yield)
ee (yield)
4
5
4 (76)
6 (68)
59 (75)
65 (70)
68 (59)
48 (39)
51 (39)
54 (29)
53 (9)
73 (64)
F
5
5
3
3
3 (55)
1 (79)
1 (73)
2 (99)e
63 (56)
55 (10)
37 (56)
45 (99)e
11 (16)
70 (26)
33 (24)
53 (43)
60 (55)e
30 (39)
45 (15)
41 (22)
20 (41)
57 (2)
60 (2)
29 (9)
36 (12)
23 (6)
Cl
Br
Me
3
(60)
a
Reaction conditions: 0.5 M styrene/benzene, styrene:iodosylbenzene:catalyst ) 100 000:10 000:1 for 3 h. Chemical yields are reported
relative to consumption of iodosylbenzene. In toluene, under otherwise identical reaction conditions. c Styrene:iodosylbenzene:catalyst
b
)
1000:100:1, under otherwise identical reaction conditions. d Reaction conditions: 330 µmol of styrene, 330 µmol of 2,6-dichloropyridine
e
N-oxide, and 1 µmol of catalyst in 2 mL of m-xylene for 5 h. Chemical yields are reported relative to an external standard. The ee and
yields refers to the cis-epoxide (the enantiomers of the trans-epoxide could not be separated). The ratio of cis-epoxide/(trans-epoxide +
benzyl methyl ketone) were 99, 1.3, and 0.3 at 25, -20, and -40 °C, respectively.
complex was more enantioselective than the manganese-
(III) derivative. Also, in series of nonporphyrinic metal
complexes, better results were obtained with ruthenium
than with manganese derivatives.20 Accordingly, the
main initiative of the present studies was to compare the
enantioselectivity of epoxidation catalysis by the three
most important metal complexessiron, manganese, and
2
rutheniumsof one particular porphyrin, 1-H .
The results for the 1-Fe(Cl)- and 1-Mn(Cl)-catalyzed
epoxidation of styrene were significantly different: higher
ee’s were obtained for 1-Fe(Cl) under all reaction condi-
2 2
tions, the solvent effect (benzene vs CH Cl , Table 1) was
much larger for 1-Fe(Cl), and the effect of amines on
the enantioselectivity was different for the two catalysts
F igu r e 4. Hammett plot of the enantiomeric excesses (ee) in
-catalyzed epoxidation of para-substituted styrenes
1
-Ru(O)
2
(substituents are indicated) by iodosylbenzene.
(Table 3). All these results are readily understandable,
considering known phenomena in iron(III) and manga-
Discu ssion
14,19,21
nese(III) porphyrin chemistry:
(a) Formation of the
most reactive intermediate in iron(III) porphyrin-cata-
In principle, three variables are of primary importance
2
2
lyzed reactions does not rely on amine coordination, but
oxomanganese(V) porphyrins are the main active oxidant
only in the presence of trans-coordinated amines, while
in their absence oxomanganese(IV) porphyrin intermedi-
ates are dominant.23 (b) The equilibrium constants for
coordination of amines to iron(III) porphyrins are much
larger than for manganese(III) porphyrins.24 (c) Imida-
zole binds to iron(III) porphyrins more strongly than
pyridine by 3 orders of magnitude.25 (d) The equilibrium
constants for coordination of imidazoles and pyridines to
in enantioselective epoxidation of unfunctionalized olefins
by homochiral metal complexes: the chiral environment
created by the superstructure of the catalyst, the steric
and/or electronic effect of the olefin, and the identity of
the metal. Traditionally, the research in this field has
7
,9
focused on the first two factors, while much less is
known about the effect of the metal. To our knowledge,
in no previous case was the same porphyrin studied with
more than two different metals,1
7,18
and homochiral
ruthenium porphyrins were not reported as epoxidation
catalysts prior to our recent contribution.13 Since it is
iron(III) porphyrins are smaller in benzene than in CH
2
-
Cl
2
by about 3 orders of magnitude.25 Thus, the reduced
well known that for metalloporphyrin catalysts the
reactivity is Mn > Fe >> Ru,19 it may be anticipated that
(20) Kureshy, R. I.; Khan, N. H.; Abdi, S. H. R.; Bhatt, A. K. J . Mol.
enantioselectivity will increase in the opposite order.
Catal. 1996, 110, 33 and references therein.
Indeed, for Groves and Viski’s porphyrin,18 the iron(III)
(21) Mlodnicka, T.; J ames, B. R. In Metalloporphyrins Catalyzed
Oxidations; Montanari, F., Casella, L., Eds.; Kluwer Academic Publish-
ers: Dordrecht, 1994; pp 121-148.
(22) Gross, Z.; Nimri, S. Inorg. Chem. 1994, 33, 1731. Groves, J . T.;
Watanabe, Y. J . Am. Chem. Soc. 1986, 108, 507.
(23) Groves, J . T.; Stern, M. K. J . Am. Chem. Soc. 1988, 110, 8628.
(24) Boucher, L. J . Coord. Chem. Rev. 1972, 7, 289.
(25) Walker, F. A.; Lo, M.; Tutram Ree, M. J . Am. Chem. Soc. 1976,
98, 5552.
(17) Maillard, P.; Guerquin-Kern, J . L.; Momenteau, M. Tetrahedron
Lett. 1991, 32, 4901.
(
18) Groves, J . T.; Viski, P. J . Org. Chem. 1990, 55, 3628.
(19) Meunier, B. Metalloporphyrins Catalyzed Oxidations; Mon-
tanari, F., Casella, L., Eds.; Kluwer Academic Publishers: Dordrecht,
994; Chapter 1.
1

Metalloporphyrin-Catalyzed Epoxidation of Olefins
J . Org. Chem., Vol. 62, No. 16, 1997 5519
Sch em e 4. P la u sible Mech a n ism for Ra cem iza tion
th r ou gh a Lon g-Lived Ra d ica l In ter m ed ia te in
Ep oxid a tion Ca ta lysis by Hom och ir a l
Meta llop or p h yr in s
F igu r e 5. Plot of log â
coordination of N-methylimidazole to (tpp)Fe(Cl) in DMF, CH
Cl , CHCl , and benzene vs the polarizability parameter E of
2
, the equilibrium constant for bis-
2
-
2
3
T
the solvents. The solid line describes the best linear fit through
all points, while the broken line is generated by ignoring the
data point for benzene. The data are taken from Walker et
2
5
chemical yields for the 1-Fe(Cl)-catalyzed reactions with
all imidazole ligands in CH Cl and with unsubstituted
2 2
al.
lutidineswas surprising. This pronounced solvent effect,
which was also apparent in the 1-Ru(O) -catalyzed
epoxidation of styrene (5% ee in CH Cl vs 42% in
imidazole in benzene are probably due to formation of
bis-ligated iron(III). The enhancement of the ee’s by most
pyridine ligands suggests catalysis by amine-coordinated
2
2
2
benzene, Table 1), called for special attention.
2 2
oxoiron(IV) porphyrin cation radical in CH Cl . On the
First, we have reexamined the correlation found by
other hand, the absence of a real effect of amine ligands
on the 1-Mn(Cl)-catalyzed reactions is consistent with
the small “aminophilicity” of manganese porphyrins.
Furthermore, it also suggests that oxygen atom transfer
2
5
Walker et al. between a polarizability parameter of the
2
7
solvents (E
tion of N-methylimidazole to (tpp)Fe(Cl). The results
of their study with DMF, CHCl , CH Cl , and benzene
T
) and the equilibrium constant for coordina-
2
8
from oxidized catalyst to substrate occurs via an oxoman-
3
2
2
ganese(IV) porphyrin intermediate.23 As 2 equiv of Mn-
are shown in Figure 5 with two linear fit lines, the full
line taking into account all four solvents and the broken
line only three points, ignoring the result for benzene.
The negative deviation of the point for benzene from the
broken line might suggest that the low equilibrium
constant in benzene is due not only to its low polarity.
This proposal was checked by comparing the coordination
(IV) is required for oxidation of styrene to its oxide, the
reaction must proceed stepwise, most probably with a
relatively long-lived benzylic radical intermediate (Scheme
2
6
4
).
Since for terminal olefins rotation around the
C-phenyl bond necessarily leads to racemization, the low
ee’s (2-15%) for epoxidation of styrene under 1-Mn(Cl)
catalysis are reasonable. Interestingly, for a monofaced
version of the same catalyst,11 which due to its available
coordination site can form a trans-amine-oxomanganese-
of N-methylimidazole (CH
3 4
-imid) to (tmp)Fe(Cl) in CCl
and benzene solutions. Considering the lower E value
T
4
of CCl (32.5) relative to that of benzene (34.5), a lower
equilibrium constant would be expected in the former
solvent if only solvent polarizability is important. But,
as shown in Figure 6, coordination is much stronger in
(V) porphyrin intermediate, the enantioselectivity is
significantly larger. Finally, a similar explanation is
proposed for the optimum enantioselectivity observed at
CCl
imid)
CCl , but not in benzene (2.5 M is required). These data
4 3
. Thus, the bis-imidazole complex [(tmp)Fe(CH -
-
20 °C in the reactions catalyzed by 1-Fe(Cl) and 2-Fe-
+
] is fully formed at 0.6 M N-methylimidazole in
(
Cl) (Figure 1). The anticipated increased differentiation
2
2
9
between the re and si approaches of styrene to the
oxometal center at lower temperaturessenhancing the
enantioselectivitysis probably compensated for by an
increased lifetime of the racemization-causing radical
intermediate. Support for this proposal is provided by
the epoxidation results of cis-â-methylstyrene, in which
the identity of the major product changes from the cis-
epoxide at high temperature to the trans-epoxide at
temperatures lower than -20 °C.
4
clearly indicate specific interactions of benzene with the
iron(III) porphyrin. Further evidence comes from the
crystal structures of aromatic solvates of iron and man-
ganese porphyrins, as well as from solution characteris-
tics of a ruthenium porphyrin. Thus, a comparison of
the X-ray structures of the isoelectronic (tpp)Mn‚
+
(toluene)
2
and [(tpp)Fe‚(p-xylene)
2
]
complexes clearly
shows that in the iron complex the aromatic molecules
have a much stronger and more specific association (with
The presence and absence of a pyridine-ligand effect
on the enantioselectivity of the 1-Fe(Cl)-catalyzed ep-
oxidation reactions in CH
Cl and benzene, respectively,
2 2
3
0,31
both the porphyrin π-system and the metal d orbitals).
The interaction of benzene with ruthenium porphyrins
suggest that these amines do not coordinate to the
catalyst in benzene. This was confirmed by adding the
(
27) Reichardt, C. Angew. Chem., Int. Ed. Engl. 1965, 4, 29.
(28) (a) It was recently shown (Nesset, M. J . M.; Shokhirev, N. V.;
same concentration of pyridine to CH
solutions of 1-Fe(Cl). Indeed, the UV-vis spectrum of
the CH Cl solution changed upon addition of pyridine,
2 2
Cl and benzene
Enemark, P. D.; J acobson, S. E.; Walker, F. A. Inorg. Chem. 1996, 35,
5188) that, in contrast to common belief, the equilibrium constants
for binding of substituted pyridines and imidazoles (including 2-meth-
ylimidazole) to sterically hindered iron(III) porphyrins such as (tmp)-
2
2
while the benzene solution was unaffected. Still, the 44%
ee in benzene without addition of amines relative to the
best result in CH Cl s36% ee in the presence of 2,4-
2 2
4
Fe(ClO ) is actually very large. This justifies the utilization of the
sterically hindered amines included in this study. (b) Abbreviations:
tpp, 5,10,15,20-tetraphenylporphyrin dianion; tmp, 5,10,15,20-tet-
ramesitylporphyrin dianion.
(
29) The equilibrium constant for bis-coordination (â
2
) of N-meth-
-1
(26) For evidence for the proposed radical intermediate in catalysis
ylimidazole to (tmp)Fe(Cl) in CCl was determined as 17.4 M , but
4
by achiral metalloporphyrins, see: (a) Groves, J . T.; Gross, Z.; Stern,
M. Inorg. Chem. 1994, 33, 5065. (b) Gross, Z.; Nimri, S. J . Am. Chem.
Soc. 1995, 117, 8021. (c) Gross, Z. J . Biol. Inorg. Chem. 1996, 1, 368.
no reliable equilibrium constant could be measured in benzene because
of the very large amount of N-methylimidazole (2.5 M) required for
full formation of the bis-adduct.

5
520 J . Org. Chem., Vol. 62, No. 16, 1997
Gross and Ini
more recent investigation of that system, Groves et al.
have provided evidence for involvement of oxoruthenium-
(
V) under these conditions.35
The almost identical results obtained for 1-Fe(Cl)-
catalyzed epoxidation of styrene with either iodosylben-
zene or iodosylmesitylene indicate that the reaction
proceeds through only one common intermediate. On the
contrary, the different ee’s obtained in the reactions of
styrene with different oxidants in the presence of cata-
lytic amounts of 1-Ru(O) (Table 4) clearly demonstrate
2
that the trans-dioxoruthenium(VI) porphyrin is not the
only oxygen-atom-transfer intermediate. Actually, the
stoichiometric oxidation of styrene by 1-Ru(O)
5% ee reveals that epoxidation by the other not yet
characterized intermediates is more enantioselective
49-57% ee for reactions with the pyridine N-oxides).
2
with only
3
(
Another demonstration of the differences in the reactions
with iodosylbenzene and N-oxides is shown in Figures 3
and 4. While construction of a Hammett plot for the ee’s
in 1-Ru(O)
2
-catalyzed epoxidation of ring-substituted
+
styrenes by 2,6-dichloropyridine N-oxide vs the σ val-
ues36 of the substituents yields a reasonable straight line;
for the corresponding reactions with iodosylbenzene a
curved line is obtained. These data indicate that the
reaction mechanisms of these two reactions are signifi-
cantly different, due to the involvement of different
oxygen-atom-transfer intermediates.
F igu r e 6. UV-vis spectra of identical concentrations of (tmp)-
Fe(Cl) in CCl and benzene solutions: without base (s), with
.6 M CH -imid (‚ ‚ ‚), and with 2.5 M CH -imid (- - -).
4
0
3
3
is so strong that it persists even in solution, as was
1
II
disclosed from the H NMR spectrum of (tmp)Ru in
3
2
benzene-d
6
.
Most important, the relative strength of
The last point of interest is the catalytic activity of the
homochiral iron porphyrin complexes. Most asymmetric
epoxidations of styrene and similar substrates deal with
less than 100 catalytic turnovers (we are aware of only
interaction of aromatic molecules with the metallopor-
phyrins in the order of Ru > Fe > Mn parallels the effect
of increased enantioselectivity in changing the solvent
from CH Cl to benzene (Table 1). Accordingly, we
2 2
propose that the strong association of benzene and the
3
7
38
three exceptions, with 2800 (20% ee), 1800 (52% ee),
and 485 (58% ee)10 turnovers).7 Moreover, in many cases,
the enantioselectivity changes during the process. For
example, while 78% ee are obtained in Collman’s man-
ganese porphyrin-catalyzed epoxidation of cis-â-methyl-
styrene after 89 turnovers, as the process continues the
ee gradually decreases down to 57% ee after 1000
turnovers.11 Such phenomena are usually ascribed to
irreversible chemical modification of the catalyst’s su-
perstructure.9 In the present case, the ee remained
practically constant even at a large turnover number,
with both the 1-Fe(Cl) and 2-Fe(Cl) catalysts. For
other aromatic solvents with the 1-Fe(Cl) and 1-Ru-
2
(O) catalysts affects the solution structure of the chiral
cavity. This can explain the 44% ee in benzene relative
to 36% ee in CCl
4
for 1-Fe(Cl) and the 59% vs 38% ee
for 2-Fe(Cl)-catalyzed epoxidation of styrene in benzene
and heptane, respectively.
In the present studies, the effects of different oxidants
were checked for the 1-Fe(Cl)- and 1-Ru(O) -catalyzed
2
epoxidations for the following reason. The concept of
oxometal porphyrins as the oxygen-atom-transfer inter-
mediates implies that the identity of the primary oxidant
must have no effect on the chemical yields, shape, or
enantioselectivity. This is especially true for iron por-
phyrins, for which the oxoiron(IV) porphyrin radical
2
-Fe(Cl) catalyzed epoxidation of styrene by iodosylben-
zene, 68% ee were obtained after 800 turnovers at -20
C and 59% ee after 6000 turnovers at room temperature.
°
To our knowledge, this is the largest ever reported
turnover number for enantioselective epoxidation of
unfunctionalized olefins. In addition, the 68% ee for
epoxidation of styrene and 70% ee for 4-chlorostyrene are
practically identical to the highest ever reported values
for these olefinss69% and 70% ee with 86 and and 83
turnovers, respectively, by the most advanced Collman’s
porphyrin.11 Finally, the enantioselectivity is also sig-
nificantly higher than that obtained with homochiral
intermediate is known to be a much more potent oxidant
than all other high valent complexes.33 As mentioned
earlier, the situation is more complex for catalysis by
manganese(III) porphyrins because of the potential in-
volvement of both oxomanganese(V) and oxomanganese-
IV) intermediates.22 Finally, trans-dioxoruthenium(VI)
(
derivatives are the best characterized epoxidizing inter-
mediates in the catalytic cycle of ruthenium porphyrins.15
But, Hirobe et al. have recently demonstrated that more
reactive intermediates must be formed in the combina-
tion of either (carbonyl)ruthenium(II) or dioxoruthenium-
7
salen catalysts, with the exception of 86% ee for epoxi-
dation of styrene by a combination of m-CPBA and
N-methylmorpholine N-oxide at -78 °C with 22 turn-
overs.39
VI) porphyrins with aromatic N-oxides.34 In an even
(
(
30) Kirner, J . F.; Reed, C. A.; Scheidt, R. J . Am. Chem. Soc. 1977,
9
3
9, 1093.
(35) Groves, J . T.; Bonchio, M.; Carofiglio, T.; Shalyaev, K. J . Am.
Chem. Soc. 1996, 118, 8961.
(31) Xie, Z.; Bau, R.; Reed C. A. Angew. Chem., Int. Ed. Engl. 1994,
3, 2433.
(36) Advances in Linear Free Energy Relationships; Chapman, N.
B., Shorter, J ., Eds.; Plenum Press: New York, 1972.
(37) O’Malley, S.; Kodadek, T. J . Am. Chem. Soc. 1989, 111, 9116.
(38) Halterman, R. L.; J an, S.-T. J . Org. Chem. 1991, 56, 5253.
(39) Palucki, M.; Pospisil, P. J .; Zhang, W.; J acobsen, E. N. J . Am.
Chem. Soc. 1994, 116, 9333.
(
32) Camenzind, M. J .; J ames, B. R.; Dolphin, D. J . Chem. Soc.,
Chem. Commun. 1986, 1137.
33) Watanabe, Y.; Groves, J . T. In The Enzymes; Sigman, D. S.,
Ed.; Academic Press: San Diego, 1992; Vol. XX, Chapter 9.
34) Ohtake, H.; Higuchi, T.; Hirobe, M. Heterocycles 1995, 40, 867.
(
(

Metalloporphyrin-Catalyzed Epoxidation of Olefins
J . Org. Chem., Vol. 62, No. 16, 1997 5521
Su m m a r y a n d Con clu sion s
the X-ray crystal structures of the first two derivatives were
The porphyrin 2-H and its iron
2
complex 2-Fe(Cl) were obtained by procedures similar to that
described previously.1
2,13
Several variables that affect the enantioselective ep-
oxidation of unfunctionalized olefins by metalloporphyrin
catalysis were investigated in the present studies. Of the
three metal complexes of one particular porphyrin, much
better results were obtained with iron and ruthenium
than with manganese. The poor ee’s obtained with the
manganese catalyst, even in the presence of amines,
suggest oxomanganese(IV) rather than oxomanganese-
of porphyrin 1-H
2
and its iron complex. The exact procedures
will be published separately in a forthcoming publication.
Preparation of the iron(III) and manganese(III) porphyrins
from 1-H
Hom och ir a l Ch lor oir on (III) P or p h yr in 1-F e(Cl). Por-
phyrin 1-H (30 mg, 24 µmol) was dissolved in DMF (4 mL)
containing 2,6-lutidine (24 µL) and heated to reflux under Ar.
A solution of anhydrous FeCl (13 mg, 0.1 mmol) in DMF (4
mL) was added in one portion, and heating was continued for
h. The cooled reaction mixture was diluted with CH Cl
2
was achieved by the following procedures.
2
2
(
V) as the oxygen-transfer intermediate. For the ruthe-
nium and iron catalysts, the superior enantioselectivity
in aromatic compared to both more and less polar
nonaromatic solvents suggest that specific association of
aromatic molecules to the metalloporphyrin affects its
solution structure. Strong evidence for the involvement
of active oxidants that are more selective than trans-
dioxoruthenium(VI) porphyrin is provided by the signifi-
cant effect of primary oxidants on the ruthenium por-
phyrin-catalyzed reactions. Although variations of the
porphyrin’s superstructure were not the major research
goal of the present studies, seemingly remote structural
changes had major effects. Preliminary results with the
iron complex of the modified porphyrin under the opti-
mized reaction conditions found in this study resulted
in epoxidation of styrene derivatives to their epoxides
with enantiomeric excesses identical to the best ever
reported, together with an unprecedented large catalytic
turnover number. The high yield condensation of the
porphyrin precursor with the chiral moieties, which are
readily available by simple modification of tartaric acid,
together with the selective formation of only one isomer,
ensures that a large number of similar complexes can
be prepared. Optimization of the catalytic process, which
already has the advantage of mild and simple working
conditions and high catalytic efficiency, with a large
number of similar porphyrin complexes is currently
under investigation.
2
2
2
,
washed with 5% HCl (twice) and water, and dried by solid
NaCl. After evaporation of the solvent, one fast-moving
fraction was obtained by column chromatography on basic
alumina (EtOAc/CH
lization from CHCl
product (20 mg, 62%): FAB MS m/z: 1300.5 ([M - H] , 100);
) 0.32 (alumina, CH1 Cl /EtOAc 2:3); UV-vis (CH Cl
nm) 444 (Soret), 368; H NMR (CDCl ) δ 76.4 (s, pyrrole-H).
Hom och ir a l Ch lor om a n ga n ese(III) P or p h yr in 1-Mn -
Cl). Porphyrin 1-H (15 mg, 12 µmol) was dissolved in DMF
15 mL) containing three drops of 2,6-lutidine, and MnBr (52
2
Cl
2
, 3:2). Final purification by recrystal-
3
/heptane resulted in the dark pink solid
+
R
f
2
2
2
2
, λmax,
3
(
2
(
2
mg, 0.24 mmol) was added in one portion. The reaction
mixture was heated overnight at 100 °C, during which time
the changes in the UV-vis spectrum from λmax ) 440 to λmax
)
486 nm were complete. The cold reaction mixture was
diluted with CH Cl
2
2
(30 mL), washed with brine (2 × 30 mL),
and dried with solid NaCl. After evaporation of the solvent,
column chromatography on basic alumina was used to separate
traces of free base (CH
manganese porphyrin (MeOH/EtOAc 1:1). Final purification
by recrystallization from CHCl /heptane resulted in the green
2 2 3
Cl /EtOAc/Et N 50:49:1) from the
3
+
solid product (14.6 mg, 90%): FAB MS m/z 1299.7 ([M - H] ,
1
λ
f
00); R ) 0.42 (alumina, MeOH/EtOAc 1:1); UV-vis (benzene,
max, nm) 384, 412, 494 (Soret), 606, 648.
Ca ta lytic Oxid a tion P r oced u r es. a . With Iod osylben -
zen e. The reactions were performed at 25 °C by adding 100
µmol of iodosylbenzene in one portion to well-stirred 1 mL
solutions of 1 mmol of olefin and 1 µmol of catalyst. Reactions
were stopped after 1 h by freezing the reaction mixture using
2
liquid N . The reaction products were separated from the
Exp er im en ta l Section
P h ysica l Meth od s. The H NMR spectra were recorded
catalyst and any unreacted iodosylbenzene by bulb-to-bulb
vacuum distillation prior to gas chromatographic analysis.
Both the chemical yieldssreported relative to the reduced
oxidant, iodobenzenesand the ee’s were determined by GC,
using a Cyclodex-B capillary column. For determination of
the effect of catalyst concentration (1.0, 0.1, and 0.01 mol %)
on the process, stock solutions of 1-Fe(Cl) were utilized under
otherwise identical reaction conditions. The reaction with 0.01
mol % catalyst was also examined in two additional ways. The
results in Figure 2 were obtained by removing aliquots from
the reaction mixture at different time intervals, followed by
freezing, bulb-to-bulb vacuum distillation, and gas chromato-
graphic analysis. The yields in these reactions are reported
relative to nitrobenzene, present as internal standard in the
reaction mixtures. For even more reliable turnover numbers,
the epoxidation reactions of styrene by iodosylbenzene in the
presence of 0.01 mol % 1-Fe(Cl) and 2-Fe(Cl) were performed
1
on a Brucker AM 200, operating at 200 MHz. Chemical shifts
are reported in ppm relative to residual hydrogens in the
deuterated solvents: 7.20, 7.24, and 5.32 ppm for benzene,
chloroform, and dichloromethane, respectively. An HP 8452A
diode array spectrophotometer was used to record the elec-
tronic spectra. Gas chromatographic analysis was performed
on a HP-5890 GC with a J &W chiral cyclodex-B capillary
column and FID detector, linked to the HP Chem-Station (HP-
3
365). The ee’s were reproducible within (2% for multiple
experiments.
Ma ter ia ls. Dichloromethane (Lab-Scan, HPLC grade) was
dried by distillation over CaH
free) was further purified by repeated washing with concen-
trated H SO until the aqueous layer was colorless, followed
by washing with water and aqueous NaOH. The benzene was
2
. Benzene (RDH, thiophene
2
4
dried by CaCl
grade) was dried by distillation over CaH
The deuterated solvents C , CDCl
2
and distilled over CaH
2
.
DMF (Merck, GC
at reduced pressure.
Cl (Aldrich
in toluene-d
8
with nitrobenzene (50 µmol, 5.15 µL) as internal
1
2
standard. The chemical yields were determined by H NMR
without any workup procedure, while the ee’s were determined
after the usual workup procedure.
6
D
6
3
, and CD
2
2
products) were used as received. The olefins were purchased
from Aldrich (cis-â-methylstyrene from K&K, ICN Biomedi-
cals) and filtered through a plug of basic alumina prior to their
use to remove stabilizers. (R)-(+)-styrene oxide, which was
purchased from Aldrich, was utilized for determination of (R)-
b. With P yr id in e N-Oxid es. The reactions were per-
formed at 25 °C by adding 330 µmol of the appropriate pyridine
N-oxide in one portion to well-stirred 2 mL benzene solutions
of 330 µmol of olefin and 1 µmol of catalyst, followed by the
same workup procedure. The chemical yields are reported
relative to iodobenzene, added after the workup procedure.
(
+)-styrene oxide as the major enantiomer in all reactions. The
other epoxides were obtained from the corresponding olefins
by standard m-CPBA oxidations. All epoxides, except trans-
â-methylstyrene oxide, were resolved on the J &W chiral
cyclodex-B capillary column, but the absolute configuration of
the major enantiomer in the reactions of the substituted
styrenes (Table 5) was not determined.
Ack n ow led gm en t. This research was supported by
The Israel Science Foundation, administered by The
Israel Academy of Sciences and Humanities, and The
Fund for the Promotion of Research at the Technion.
P r ep a r a tion of Ca ta lysts. The preparation of porphyrin
1
-H
2
, its ruthenium complexes 1-Ru(CO) and 1-Ru(O)
2
, and
J O970463W