Synthesis of optically active α-hydroxy carbonyl compounds by the catalytic, enantioselective oxidation if silyl enol ethers and ketene acetals with (salen)manganese(III) complexes

Article abstract of DOI:10.1021/ja9726668

A set of silyl enol ethers and ketene acetals 1a-h with α- and/or β-phenyl as well as alkyl substituents of different steric bulk has been submitted to the enantioselective catalytic oxidation by chiral (salen)Mn(III) complexes 3. Highest conversion and best enantioselectivities have been obtained with bleach rather than iodosobenzene as oxygen source for the active oxo-metal species. With regard to substrate structure ee values up to 89% have been achieved for enol ethers with short and unbranched alkyl substituents at the siloxy position. While β-phenyl groups are beneficial for enantiofacial control, phenyl substituents α to the siloxy functionality result in lower ee values, while the diphenyl-substituted derivative 1d displays the lowest stereoselectivity. The fact that β- versus α-phenyl substituents exhibit not only differences in the magnitude but also in the sense (opposite absolute product configuration) of the stereoselectivity may be utilized as a valuable mechanistic probe to assess steric and electronic effects in the substrate and the catalyst as a function of the type and pattern of substitution. Our results display that sterid interactions between the substrate and the oxo-metal complex are mainly responsible for the observed stereochemical preferences. Indeed, significantly increased enantioselectivities are achieved even for the remote siloxy group when bulkier derivatives are employed. In contrast, primarily electronic effects operate in the (salen)Mn(III) catalyst 3 since electron-donating groups in the 5,5' positions of the salicylaldelhyde ligand afford higher ee values in this catalytic oxidation. The skewed side-on approach (trajectory) of the substrate onto the oxo-metal catalyst is favored, the metallaoxetane mechanism adequately accounts for the observed enantioselectivities. Herewith a synthetically valuable method for the preparation of optically active α-hydroxy carbonyl products 2 has been made available through the catalytic, enantioselective oxidation of the silyl enol ethers 1 by (salen)Mn(III) complexes.

Full text of DOI:10.1021/ja9726668

708
J. Am. Chem. Soc. 1998, 120, 708-714
Synthesis of Optically Active R-Hydroxy Carbonyl Compounds by
the Catalytic, Enantioselective Oxidation of Silyl Enol Ethers and
Ketene Acetals with (Salen)manganese(III) Complexes
Waldemar Adam, Rainer T. Fell,* Veit R. Stegmann, and Chantu R. Saha-Mo1ller
Contribution from the Institute of Organic Chemistry, UniVersity of Wu¨rzburg,
Am Hubland, D-97074 Wu¨rzburg, Germany
ReceiVed August 4, 1997
Abstract: A set of silyl enol ethers and ketene acetals 1a-h with R- and/or â-phenyl as well as alkyl substituents
of different steric bulk has been submitted to the enantioselective catalytic oxidation by chiral (salen)MnIII
complexes 3. Highest conversions and best enantioselectivities have been obtained with bleach rather than
iodosobenzene as oxygen source for the active oxo-metal species. With regard to substrate structure, ee
values up to 89% have been achieved for enol ethers with short and unbranched alkyl substituents at the
siloxy position. While â-phenyl groups are beneficial for enantiofacial control, phenyl substituents R to the
siloxy functionality result in lower ee values, while the diphenyl-substituted derivative 1d displays the lowest
stereoselectivity. The fact that â- versus R-phenyl substituents exhibit not only differences in the magnitude
but also in the sense (opposite absolute product configuration) of the stereoselectivity may be utilized as a
valuable mechanistic probe to assess steric and electronic effects in the substrate and the catalyst as a function
of the type and pattern of substitution. Our results display that steric interactions between the substrate and
the oxo-metal complex are mainly responsible for the observed stereochemical preferences. Indeed,
significantly increased enantioselectivities are achieved even for the remote siloxy group when bulkier derivatives
are employed. In contrast, primarily electronic effects operate in the (salen)MnIII catalyst 3 since electron-
donating groups in the 5,5′ positions of the salicylaldehyde ligand afford higher ee values in this catalytic
oxidation. The skewed side-on approach (trajectory b) of the substrate onto the oxo-metal catalyst is favored,
the metallaoxetane mechanism adequately accounts for the observed enantioselectivities. Herewith a
synthetically valuable method for the preparation of optically active R-hydroxy carbonyl products 2 has been
made available through the catalytic, enantioselective oxidation of the silyl enol ethers 1 by (salen)MnIII
complexes.
Introduction
optically active oxidants, e.g., oxaziridines.3,6 Highly enantio-
selective catalytic methods are known such as the oxidation of
enol ethers by the osmium-catalyzed asymmetric dihydroxyla-
tion,7 the kinetic resolution of R-hydroperoxy esters with
horseradish peroxidase8 and that of R-hydroxy esters and ketones
with lipases.9
A recent major advance in catalytic enantioselective oxida-
tions has been the epoxidation of prochiral, unfunctionalyzed
olefins.10 For this purpose (salen)MnIII complexes, which are
accessible from readily available precursors,11 have been
The optically active R-hydroxy ester and R-hydroxy ketone
functionalities are widespread in natural products and have
during the last years been frequently used as convenient building
blocks in organic synthesis.1 For example, the enantiopure
R-hydroxy ketone 2a is an intermediate product of the “Knoll
process”, the major industrial route to the sympathomimetica
ephedrine and pseudoephedrine with an adrenalin-like activity.2
Unquestionably, efficient methods for the construction of
enantiomerically pure or at least enriched R-hydroxy carbonyl
compounds are in demand.3 For their preparation, the meth-
odology of electrophilic hydroxylation of enolates, in which the
optically active organic or organometallic auxiliary is covalently
bound to the enol unit, has mainly been employed.4,5 Alter-
natively, prochiral enolates have been directly oxidized by
(6) (a) Davis, F. A.; Weismiller, M. C. J. Org. Chem. 1990, 55, 3715-
3717. (b) Davis, F. A.; Weismiller, M. C.; Lal, G. S; Chen, B.-C.;
Przeslawski, R. M. Tetrahedron Lett. 1989, 30, 1613-1616.
(7) Hashiyama, T.; Morikawa, K.; Sharpless, K. B. J. Org. Chem. 1992,
57, 5067-5068.
(8) Adam, W.; Fell, R. T.; Hoch, U.; Saha-Mo¨ller, C. R.; Schreier P.
Tetrahedron: Asymmetry 1995, 6, 1047-1050.
(9) (a) Adam, W.; D´ıaz, M. T.; Fell, R. T.; Saha-Mo¨ller, C. R.
Tetrahedron: Asymmetry 1996, 7, 2207-2210. (b) Hsu, S.-H.; Wu, S.-S.;
Wang, Y.-F.; Wong, C.-H. Tetrahedron Lett. 1990, 31, 6403-6406.
(10) (a) Jacobsen, E. N. In Catalytic Asymmetric Synthesis; Ojima, I.,
Ed.; VCH: New York, 1993; Chapter 4.2. (b) Jacobsen, E. N. In
ComprehensiVe Organometallic Chemistry II; Abel, E. W., Stone, F. G.
A., Wilkinson, G., Hegedus, L. S., Eds.; Pregamon: New York, 1995; Vol.
12; Chapter 11.1. (c) Katsuki, T. J. Mol. Catal. A, Chem. 1996, 113, 87-
107. (d) Katsuki, T. Coord. Chem. ReV. 1995, 140, 189-214. (e) Linker,
T. Angew. Chem., Int. Ed. Engl. 1997, 36, 2060-2062.
(1) (a) Coppola, G. M.; Schuster, H. F. R-Hydroxy Acids in Enantiose-
lectiVe Synthesis; Wiley-VCH: Weinheim, 1997. (b) Hanessian, S. Total
Synthesis of Natural Products: The Chiron Approach; Pergamon: New
York, 1983; Chapter 2.
(2) (a) Poelert, M. A.; Hulshof, L. A.; Kellogg, R. M. Recl. TraV. Chim.
Pays-Bas 1994, 113, 355-364. (b) Brauch, F.; Wernicke, F.; Knoll A.-G.
Ger. Offen DE. 3,230, 333 1982; Chem. Abstr. 1984, 100, P6059f.
(3) Davis, F. A.; Chen, B.-C. Chem. ReV. (Washington, D.C.) 1992, 92,
919-934.
(4) Enders, D.; Bhushan, V. Tetrahedron Lett. 1988, 29, 2437-2440.
(5) Adam, W.; Prechtl F. Chem. Ber. 1994, 127, 667-671.
(11) Larrow, J. F.; Jacobsen, E. N.; Gao, Y.; Hong, Y.; Nie, X.; Zepp,
C. M. J. Org. Chem. 1994, 59, 1939-1942.
S0002-7863(97)02666-8 CCC: $15.00 © 1998 American Chemical Society
Published on Web 01/16/1998

Optically ActiVe R-Hydroxy Carbonyl Compounds
J. Am. Chem. Soc., Vol. 120, No. 4, 1998 709
developed as highly enantioselective catalysts not only for the
epoxidation of aryl-conjugated, cis-disubstituted olefins, but also
for the epoxidation of tri- and tetrasubstituded derivatives.12 An
increase of the enantioselectivities for these substrates has been
achieved by modulating steric10 and electronic properties13 of
the catalyst through varying the substitution pattern at the
salicylidene moiety and the chiral diamine part.
Scheme 1. Asymmetric Oxidation of the Silyl Enol Ethers
and Silyl Ketene Acetals 1 by (salen)Mn(III) Catalysts 3
In this context, recently we have reported in a preliminary
contribution that also silyl enol ethers are good prochiral
substrates for such oxidations to afford optically active R-hy-
droxy carbonyl compounds.14 Similarly, the enantioselective
oxidation of alkyl and acetyl enol ethers with (salen)MnIII
complexes to R-hydroxy acetals has been conducted.15 Previ-
ously an optically active pyrrolidine-based (salen)MnIII complex
was employed as catalyst for the enantioselective oxidation of
silyl enol ethers with iodosobenzene (ee 14-62%).16
For the present study, the silyl enol ethers 1 and the (salen)-
MnIII complexes 3 were chosen to assess the structural and
electronic factors that influence the enantioselectivity of this
metal-catalyzed oxidation, in the hope to increase its stereo-
control by proper tuning of the substrate and catalyst. The ideal
products. It was expected that the direction of the enantiofacial
attack should decisively depend on the steric nature of this third
substituent.
The silyl ketene acetals 1g-h were chosen to probe the
electronic effect of a methoxy or a thioethyl group on the
enantioselectivity compared to an alkyl or aryl R2 group. To
test the influence of the steric nature of the siloxy protecting
group on the enantioselectivity, besides the trimethylsilyl
derivatives TMS-1, also the sterically more demanding tert-
butyldimethylsilyl-substituted TBDMS-1′ substrates were se-
lected for this catalytic oxidation. The electronically modified
salen complexes 3b,c with electron-donating substituents in the
5,5′ positions should provide valuable mechanistic information
on the electronic role that the manganese catalyst 3 plays in
this enantioselective catalytic oxyfunctionalizaion.
We present herewith the full details of the (salen)Mn-
catalyzed oxidation of phenyl-substituted silyl enol ethers and
ketene acetals 1 to the respective optically active R-hydroxy
carbonyl products 2. Our results demonstrate that the substitu-
tion pattern in the substrates 1 significantly influences the
enantioselectivity and the absolute configuration of the products
2 in this novel metal-catalyzed R-hydroxylation. However,
through electronic tuning of the (salen)MnIII catalyst 3 only
moderately enhanced stereocontrol may be achieved.
Results
Syntheses. The synthesis of the (salen)MnIII complexes 3b,c
was conducted according to reported procedures.11,19
The silyl enol ethers and ketene acetals 1a-h were prepared
by following literature methods, in which the corresponding
ketone or ester was deprotonated with LDA, NaH, or NaN-
(SiMe3)2. The reaction conditions were optimized to get mainly
enolates which on treatment with chlorosilane afforded the
Z-configured silyl enol ethers 1. The E/Z ratios were determined
by analysis of the characteristic olefinic proton resonances in
the 1H NMR spectra of the silylated products.20 Racemic
samples of R-hydroxy ketones 2a-f were obtained by dimethyl-
dioxirane oxidation of the corresponding silyl enol ether
1a-f.21
Enantioselectivity Studies. The catalytic oxidations were
carried out at 0 °C with NaOCl (7.5 equiv) in a phosphate buffer
(pH 11.3) or PhIO (1.5 equiv) as oxygen atom source, with 30
mol % 4-phenylpyridine N-oxide (PPNO) as additive and 7 mol
% of the appropriate (salen)Mn complex 3 as catalyst.10,19 The
optically active R-hydroxy carbonyl compounds were released
from the resulting R-siloxy/R-hydroxy product mixture by
treatment with acidified (HCl) or pure methanol (Scheme 1).21
In general, the substrate conversion was very fast and usually
substrate17 would be the cis-disubstituted silyl enol ether derived
from phenyl acetaldehyde (R1 ) Ph and R2 ) H), but the
resulting R-hydroxy aldehyde is too labile for handling.18
Therefore, it was necessary to place an alkyl or a phenyl group
in the R position to the siloxy group as third substituent, which
would afford the more persistent R-hydroxy ketones 2 as
(12) (a) Brandes, B. D; Jacobsen, E. N. J. Org. Chem. 1994, 59, 4378-
4380. (b) Fukuda, T.; Irie, R.; Katsuki, T. Synlett 1995, 197-198. (c)
Brandes, B. D; Jacobsen, E. N. Tetrahedron Lett. 1995, 36, 5123-5126.
(13) Jacobsen, E. N.; Zhang, W.; Gu¨ler, M. L. J. Am. Chem. Soc. 1991,
113, 6703-6704.
(14) Adam, W.; Fell, R. T.; Mock-Knoblauch, C.; Saha-Mo¨ller, C. R.
Tetraha edron Lett. 1996, 37, 6531-6534.
(19) (a) Deng, L.; Jacobsen, E N. J. Org. Chem. 1992, 57, 4320-4323.
(b) Pospisil, P. J.; Carsten, D. H.; Jacobsen, E. N. Chem. Eur. J. 1996, 2,
974-980. (c)Finney, N. S.; Pospisil, P. J.; Chang, S.; Palucki, M.; Konsler,
R. G.; Hansen, K. B.; Jacobsen, E. N. Angew. Chem., Int. Ed. Engl. 1997,
36, 1720-1723.
(20) Davis, F. A.; Sheppard, A. C.; Chen, B.-C.; Haque, M. S. J. Am.
Chem. Soc. 1990, 112, 2, 6679-6690.
(21) (a) Adam, W.; Hadjiarapoglou, L.; Wang, X. Tetrahedron Lett. 1991,
32, 1295-1298. (b) Rubottom, G. M.; Vazquez, M. A.; Pelegrina, D. R.
Tetrahedron Lett. 1974, 4319-4322.
(15) Fukuda, T.; Katsuki, T. Tetrahedron Lett. 1996, 37, 4389-4392.
(16) Reddy, D. R.; Thornton, E. R. J. Chem. Soc., Chem. Commun. 1992,
172-173.
(17) Jacobsen, E. N.; Zhang W.; Muci, A. R.; Ecker, J. R; Deng, L. J.
Am. Chem. Soc. 1991, 113, 7063-7064.
(18) (a) Weygand, F.; Bestmann, H. J.; Ziemann, H.; Klieger, E. Chem.
Ber. 1958, 91, 1043-1049. (b) Griffiths, D. W.; Gutsche, C. D. J. Org.
Chem. 1971, 36, 2184-2186.

710 J. Am. Chem. Soc., Vol. 120, No. 4, 1998
Adam et al.
Table 1. Enantioselective Oxidation of Silyl Enol Ethers 1a-f and
Silyl Ketene Acetals 1g,g′,h by the (S,S)-(salen)MnIII Catalyst 3a
with NaOCla
In the set of chosen Z-configured trimethylsilyl enol ethers
1a-f, the best enantioselectivities in the oxidation with catalyst
(S,S)-3a and bleach were obtained for the silyl enol ethers 1a,b,
which bear one phenyl group in the â position to the siloxy
group (R1 ) Ph) and a less sterically demanding alkyl group
(R2 ) Me, Et; 1a,b) in the R position to the siloxy group (Table
1, entries 1,3). The ethyl-substituted derivative 1b resulted in
an enantiomeric excess of 87%, which is a slightly better value
than for the methyl-substituted substrate 1a (ee 79%).
An increase of the steric demand of the R2 substituent by
using the bulky tert-butyl group (1c) led to a drastic decrease
of the enantioselectivity (Table 1, entry 4). With the diphenyl-
substituted substrate 1d (R1 ) R2 ) Ph) the catalytic oxidation
gave the benzoin (2d) with the lowest (12%) ee value in Table
1 (entry 5). Contrary to the S configuration for R-hydroxy
ketones 2a-c with the catalyst (S,S)-3a, the R configuration
prevails in the case of the product 2d.
The substrates 1e,f with R1 an alkyl and R2 a phenyl group,
afforded R-hydroxy ketones 2e,f in enantiomeric excess up to
60% for the R enantiomer (Table 1, entries 6,8). Comparison
of the absolute configurations of the products 2e,f with the
already discussed ones 2a-d, reveals that the â-phenylated (R1
) Ph) substrates 1a-c afford the (S)-configured R-hydroxy
ketones 2a-c as the major enantiomer (Table 1, entries 1, 3,
and 4). In contrast, the R-phenylated (R2 ) Ph) substrates 1d-f
preferentially lead to the R enantiomer of the R-hydroxy ketones
2d-f (Table 1, entries 5, 6, and 8).
entry
substrate
convb,c,d (%)
ee (%)d,e
2
configf 2
1
2
3
4
5
6
7
8
9
1a
91 (81)
59
82 (88)
96
88
95
95
84
64
70
79 (56)
86
87 (60)
35
12
42
81
60
22
S (+)
S (+)
S (+)
S (+)
R (-)
R (+)
R (+)
R (+)
S (+)
S (+)
S (+)
1a′
1b
1c
1d
1e
1e′
1f
1gg
1g′
1hh
10
11
57
-- (35)
--i (18)
a 7 mol % catalyst (S,S)-3a, 7.5 equiv of NaOCl as 0.5 M solution
in phosphate buffer (pH 11.3), 0.3 equiv of PPNO (4-phenylpyridine
N-oxide), in CH2Cl2. b Determined by HPLC analysis (RP-18, 64:34:2
MeOH/H2O/CH3CN, flow rate 1.0 mL/min or Chiralcel OD, 9:1
n-hexane/2-propanol, flow rate 0.6 mL/min). c Yield of isolated product
was 60-98% relative to the conversion of 1 to R-hydroxy carbonyl
products 2. d Values in parentheses are for PhIO (1.5 equiv). e Deter-
mined by HPLC analysis (Chiralcel OD, 9:1 n-hexane/2-propanol, flow
rate 0.6 mL/min), error limits ∼<5% of the stated values. f Configu-
rations assigned according to literature (refs 20 and 22-25), for 2c in
analogy to the elution order of the enantiomers 2a and 2b. g 70:30 E/Z.
h
i
∼50:50 E/Z. Neither product nor starting material could be reisolated.
high. The results of the oxidations with the standard catalyst
(S,S)-3a are summarized in Table 1.
The oxidation of the silyl enol ether 1a with catalyst 3a and
NaOCl proceeded to 88% conversion (HPLC analysis) in less
than 1 h, but the conversion could not be increased significantly
(91%, 24 h) by allowing longer reaction times (Table 1, entry
1). The reason is that under the aqueous reaction conditions,
the hydrolysis of the silyl enol ethers and ketene acetals
competes with the catalytic oxidation to afford the nonhydroxy-
lated ketone or ester, the starting materials for the silylated
substrates 1. To achieve the highest possible conversion and
to facilitate the separation of the decomposed catalyst, the
reaction time was extended in most cases up to 24 h.
Fortunately, control experiments established that under these
reaction conditions the optically active R-hydroxy products 2
were not racemized.
Experiments with PhIO as oxygen atom source to perform
catalytic oxidations under water-free conditions to avoid the
competing hydrolysis of the silyl enol ethers showed that
conversions could also not be raised significantly. In fact, the
conversion even dropped under these conditions, but worse, the
observed enantiomeric excess was on the average ∼25% lower
than for NaOCl in aqueous media (Table 1, entries 1 and 3).
This could be due to the fact that with PhIO also oxo MnIV
species besides oxo MnV are generated. Both species were
detected and isolated in the oxidation of manganese porphyrin
complexes with iodosobenzene and the oxo MnIV species is
known to decrease the diastereoselectivity in the epoxidation
of cis-â-methylstyrene.26 This feature could also be responsible
for the decrease of the enantioselectivity in the here described
oxidations with iodosobenzene.
Catalytic oxidation of silyl enol ethers 1a′,e′, which are
functionalized with the bulkier TBDMS instead of the TMS
group, was performed to compare the influence on the enantio-
selectivity of the steric bulk of the siloxy group. For the
TBDMS-substituted substrate 1e′ an increase of the ee values
to 81% was observed (Table 1, entry 7), compared to only 42%
for the TMS-substituted substrate 1e (Table 1, entry 6). The
same trend was obtained in the oxidation of the substrates 1a′
versus 1a, but the increase was smaller yet still significant (Table
1, entries 1 and 2). The absolute configurations of the products
2a,e were the same for the TMS and the TBDMS groups.
The use of silyl ketene acetals 1g,g′,h as substrates for the
catalytic oxidation with MnIII complex (S,S)-3a showed that the
S configuration predominates in the resulting R-hydroxy esters
2g,h (Table 1, entries 9-11), the same as for the silyl enol ethers
1a-c (Table 1, entries 1-4). The change of the substituent R2
from Me (1a) to MeO (1g) or EtS (1h) had no influence on the
absolute configuration of the R-hydroxylated products 2, but
on the enantioselectivity of the reaction. The ee values for 1g,h
were only ∼20%, but the starting materials consisted of Z/E
diastereomeric mixtures of 1g,h (Table 1, entries 9 and 11).
Indeed, the diastereomerically pure silyl ketene acetal z-1g′
yielded a higher ee value of 57% (Table 1, entry 10). For the
hydrolytically labile phenylketene (S)-ethyl O-trimethylsilyl
acetal 1h, only with PhIO oxidation to the R-hydroxy thioester
2h was observed. With aqueous bleach merely decomposition
products were detected (Table 1, entry 11).
The results on the asymmetric oxidations with the modified
catalysts (R,R)-3b,c, which possess electron-donating substit-
uents at the 5,5′ positions, are given in Table 2 and compared
to those with catalyst (R,R)-3a. Conversions to the hydroxylated
products 2 were for most oxidations with catalysts (R,R)-3b,c
(Table 2) as high as for the catalyst (S,S)-3a (Table 1). The
conversions were usually in the range of 80-100% with
insignificant differences between the catalysts 3a-c, but the
enantioselectivities differed significantly. For silyl enol ether
1a an increase in enantiomeric excess from 79% for catalyst
(22) Fuganti, C.; Grasselli, P.; Poli, G.; Servi, S.; Zorzella; A. J. Chem.
Soc., Chem. Commun. 1988, 1619-1621.
(23) Davis, F. A.; Weismiller, M. C.; Murphy, C. K.; Reddy, R. T.; Chen,
B.-C. J. Org. Chem. 1992, 57, 7274-7285.
(24) Bonner, W. A. J. Am. Chem. Soc. 1951, 73, 3126-3132.
(25) (a) Aggarwal, V. K.; Thomas, A.; Franklin, R. J. J. Chem. Soc.,
Chem. Commun. 1994, 1653-1654. (b) Aggarwal, V. K.; Thomas, A.;
Schade, S. Tetrahedron 1997, 53, 16213-16228.
(26) Arasasingham, R. D.; He, G.-X.; Bruice, T. C. J. Am. Chem. Soc.
1993, 115, 7985-7991.

Optically ActiVe R-Hydroxy Carbonyl Compounds
J. Am. Chem. Soc., Vol. 120, No. 4, 1998 711
Table 2. Influence of 5,5′ Substitution in the (R,R)-(salen)MnIII Catalysts 3a-c on the Enantioselectivity in the Oxidation of Silyl Enol Ethers
and Silyl Ketene Acetals 1a
(R,R)-3a
convb,c (%)
(R,R)-3b
convb,c (%)
(R,R)-3c
convb,c (%)
entry
substrate
ee (%)d 2
ee (%)d 2
ee (%)d 2
configuration of 2e
1
2
3
4
5
6
7
8
1a
1b
1c
1d
1e
1e′
1f
91
82
95
98
95
95
83
70
79
86
34
11
42
81
60
53
99
72
49
98
88
99
88
85
89
87
37
28
74
84
75
68
98
92
95
99
85
99
91
70
87
83
51
18
74
81
77
53
R (-)
R (-)
R (-)
S (+)
S (-)
S (-)
S (-)
R (-)
1g′
a 7.5 equiv of NaOCl as 0.5 M solution in phosphate buffer (pH 11.3), 0.3 equiv of PPNO (4-phenylpyridine N-oxide), in CH2Cl2. b Determined
by HPLC analysis (RP-18, 64:34:2 MeOH/H2O/CH3CN, flow rate 1.0 mL/min or Chiralcel OD, 9:1 n-hexane/2-propanol, flow rate 0.6 mL/min).
c Yield of isolated product was 60-98% relative to conversion of 1 to R-hydroxycarbonyl products 2. d Determined by HPLC analysis (Chiralcel
OD, 9:1 n-hexane/2-propanol, flow rate 0.6 mL/min), error limits ∼<5% of the stated values. e Oxidation with (R,R)-3, configurations assigned
according to literature (refs 20 and 22-25), for 2c in analogy to the elution order of the enantiomers of 2a and 2b.
high enantioselectivities observed in the epoxidation of cis-di-
and trisubstituted olefins with (salen)Mn catalysts 3. These
models shall be applied to the silyl enol ethers 1 in terms of
the orientation of the incoming olefin (its plane and that of the
salen ligand perpendicular) in a side-on approach onto the metal
oxo ligand along the direction of the chiral diamine bridge from
the back (trajectory a, Figure 2)17,19b-c and the skewed side-on
attack along the metal CdN ligand of the diamine bridge
(trajectory b, Figure 2).27 The top-on approach (the olefin and
salen-ligand planes parallel) has been proposed only for
tetrasubstituted olefins because of steric reasons12c and should
not be relevant for the trisubstituted olefins employed herein.12a,b
Along the trajectories a and b we shall explicitely consider the
likely transition states for the oxygen transfer from the oxo-
metal complex to the olefin to form the possible radical19 or
metallaoxetane intermediates.28
Figure 1. Absolute product configurations in the oxy functionalizations
of the silyl enol ethers and ketene acetals 1a-h with catalyst (S,S)-3a.
(S,S)- or (R,R)-3a to 89% for the 5,5′-MeO-substituted derivative
(R,R)-3b was observed (Table 2, entry 1). For this substrate,
the TIPSO-substituted catalyst (R,R)-3c afforded an ee value
of 87%, and constitutes no improvement over catalyst 3b (Table
2, entry 1). In general, the differences for the catalysts 3a-c
are not substantial (∼15% increase in the ee values) but
significant for most substrates in Table 2. An exception is the
silyl enol ether of the propiophenone 1e, for which the ee value
increased by 32% (Table 2, entry 5) for the catalysts 3b,c (74%
ee) in comparison to catalyst 3a (42% ee).
In most cases the catalyst (R,R)-3b with the sterically less
bulky but electron-rich methoxy substituent afforded the best
enantioselectivities in the oxidation of silyl enol ethers 1a,b,d-
e′ and ketene acetal 1g′ (Table 2, entries 1, 2, 4-6, and 8). The
ee values for catalyst 3c with the electron-donating and more
bulky TIPSO group are similar to those for catalyst 3b (Table
2, entries 1, 5, and 7) or fall between the values for the catalysts
3a,b (Table 2, entries 2, 4, 6, and 8). Only for substrate 1c
was the TIPSO-substituted catalyst (R,R)-3c (51% ee) the most
effective (Table 2, entry 3).
Substrate Structure. The silyl enol ethers 1a-c and ketene
acetals 1g,h attack the oxo-metal complex of the catalyst (S,S)-
3a to result the S-configured, R-hydroxylated products 2a-c,g,h,
(Table 1, entries 1-4, and 9-11), as it is shown in Figure 2A
for the substrate 1a as an illustrative example. In the side-on
approach along the trajectory a (Jacobsen′s model),17,19b-c the
two possible enantiofacial orientations of the enol ether are
specified, of which the gray structure would lead to the R- and
the black one to the S-configured products. Both arrangements
are equally favored, since the steric interactions between the
methyl group of the substrate and the axial hydrogens of the
cyclohexanediamine unit of the catalyst should be similar.
Therefore, a very low enantioselectivity is to be expected for
this approach. The facts are, however, as mentioned above,
that for the substrates 1a-c,g,h preferential formation of the S
enantiomers for the oxidation products was observed (ee up to
87%). In contrast, both possible enantiofacial orientations of
the enol ether 1a in a skewed side-on approach along the
trajectory b (Katsuki’s model)12b,27 are sterically differentiable.
To avoid the repulsive steric and electronic (π-π interaction
Mechanistic Discussion
between the arene rings of salen ligand and substrate12b,27
)
interactions, the phenyl group of the substrate should be as far
as possible away from the salicyl moiety.10c,12b,27 Therefore,
the enantiofacial approach which leads to the S-configured
R-hydroxylated products should be favored, in accord with our
experimental data.
The results presented above demonstrate that the choice of
the enol-ether substrates 1a-h has been instructive in exploring
the stereocontrol in the (salen)Mn-catalyzed oxidation of
trisubstituted olefins. Thus, the degree of enantioselectivity and
the absolute product configuration in the manganese-assisted
oxyfunctionalizations of the silyl enol ethers and ketene acetals
1a-h to the corresponding R-hydroxy carbonyl compounds
(Figure 1) are controlled by the substitution pattern of the
substrates 1 and the arene substituents of the (salen)Mn catalysts
3. The trends in these structural factors shall now be analyzed
in terms of the mechanistic models proposed to rationalize the
(27) (a) Hosoya, N.; Hatayama, A.; Yanai, K.; Fujii, H.; Irie, R.; Katsuki,
T. Synlett 1993, 641-645. (b) Hamada, T.; Irie, R.; Katsuki, T. Synlett 1994,
479-481.
(28) (a) Norrby, P.-O.; Linde, C.; A° kermark, B. J. Am. Chem. Soc. 1995,
117, 11035-11036. (b) Linde, C.; Arnold, M.; Norrby, P.-O.; A° kermark,
B. Angew. Chem., Int. Ed. Engl. 1997, 36, 1723-1725. (c) Hamada, T.;
Fukuda, T.; Imanishi, H.; Katsuki, T. Tetrahedron 1996, 52, 515-530.

712 J. Am. Chem. Soc., Vol. 120, No. 4, 1998
Adam et al.
Figure 2. Trajectories a and b and the radical mechanism for the oxidation of the silyl enol ethers 1a (A) and 1e (B) by the oxo-metal complex
of catalyst (S,S)-3a.
Since in the side-on approach along the trajectory b the
substituent R2 of the enol ethers 1a-b,g,g′,h is directed toward
the salicyl moiety of the catalyst, an increase of the steric
demand for the R2 substituent (tBu in substrate 1c) results in
major steric interactions and leads to a much lower enantiose-
lectivity (Table 1, entry 4). When the R2 alkyl group is replaced
by a methoxy or thioethyl substituent (Table 1, entries 9-11),
a lower enantioselectivity is also observed.
of the catalyst during the oxygen transfer along the trajectory
b, the enantioselectivity is substantially lower than for the silyl
enol ethers 1e,f with an aliphatic R1 group (Table 1, entries
5-8). These results once more demonstrate that the skewed
side-on approach along the trajectory b10c,12b,27 is the more
favored pathway than the trajectory a in these oxidations.
With the preferred substrate approach (skewed side-on
trajectory b) and the favored orientation (black structures)
defined in Figure 2, it is now instructive to diagnose the possible
transition states for the oxygen transfer. Once the enol-ether
substrate has penetrated far enough into the salen complex to
be sufficiently proximate to the oxo-metal functionality for
bonding with the oxenoid oxygen atom, two mechanistic
alternatives apply for the oxygen transfer. There are the
currently much debated homolytic pathway proposed by
Jacobsen10a,b,19 to afford a radical intermediate and the concerted
route propagated by Katsuki10c,15,28c and Norrby-A° kermark28a,b
to form the metallaoxetane cycloadduct.
Contrary to the substrates 1a-c,g,g′,h with a phenyl group in
the â position (R1 ) Ph) to the siloxy group (Table 1, entries
1-4 and 9-11), substrates 2e-f with an R-phenyl substituent
(R2 ) Ph) display the opposite enantiofacial selectivity in the
resulting R-hydroxy ketones (R)-2e,f (Table 1, entries 6-8).
Again, this enantioselectivity cannot be rationalized in terms
of the trajectory a12a,19b,c because similar steric interactions apply
for both possible enantiofacial approaches, as illustrated for
substrate 1e as model case (Figure 2B). The stereochemical
course in this oxidation follows again the expectations of the
skewed side-on approach for the enolate 1e along trajectory b.
The enantiofacial approach to result in the S-configured product
(gray structure in Figure 2B) will be less favored because of
steric interactions of the phenyl substituent at the R1 position
of the substrate 1e with the tBu group of the salicyl moiety, as
well as repulsive steric and electronic interactions with the aryl
ring of the salicyl ligand.10c,12b,27 As observed, the R-configured
enantiomer (black structure in Figure 2B) prevails in the
oxyfunctionalization of the silyl enol ether 1e with the catalyst
(S,S)-3a.
We shall scrutinize first the Jacobsen proposal to assess how
well it rationalizes the observed enantioselectivities in Table 1.
The preferred radical intermediate for the enol ether 1a with
R1 a phenyl substituent (Figure 2A) should be the one with the
unpaired spin at the benzyl site in view of the more effective
phenyl conjugation. Subsequently the radical intermediate
collapses to an intermediary epoxide and the latter transposes
to the observed R-hydroxy ketone (S)-2a.
For the enol ether 1e with R2 a phenyl group (Figure 2B),
the site of preferred homolytic attack in the Jacobsen mechanism
is clear-cut, namely the oxenoid oxygen atom is fixed at the
methyl-bearing terminal. For the resulting intermediary radical
For the diphenyl-substituted silyl enol ether 1d, which cannot
avoid having one phenyl group near the salicylaldehyde ligand

Optically ActiVe R-Hydroxy Carbonyl Compounds
J. Am. Chem. Soc., Vol. 120, No. 4, 1998 713
Figure 3. Metallaoxetane mechanism for the oxidation of the silyl enol ethers 1a (A) and 1e (B) by the oxo complex of catalyst (S,S)-3a.
the unpaired spin is localized at the R2 site, which is stabilized
by phenyl as well as silyloxy conjugation. Subsequent collapse
to the epoxide and transposition of the latter affords the
R-hydroxy ketone (R)-2e, again as observed (Table 1).
A relevant feature of the Jacobsen stepwise mechanism, in
fact, the prevalent incentive to propose radical intermediates,
is the loss of epoxide diastereoselectivity for phenyl-substituted
cis-olefins due to rotation about the C-C bond before collapse
to the epoxide product. Our enol-ether substrates, unfortu-
nately, do not provide any information on this point because
the stereocenter at the silyloxy-substituted site is lost due to
ketone formation in the final R-hydroxycarbonyl product. To
attribute the lower (up to 80%) enantioselectivity for the enol
ether 1a compared to cis-disubstituted alkenes (up to 98%10) to
C-C bond rotation in the radical intermediate does not apply.
If this were so, for the enol ether 1e higher (>95%) ee values
would have to be expected because the final chirality center is
permanently fixed through C-O bond formation in the radical
intermediate (Figure 2B).
The Jacobsen homolytic pathway, provided a skewed side-
on approach (trajectory b) of the substrate applies, explains quite
adequately the observed enantioselectivities for the trimethylsilyl
derivatives 1e,g; however, it is difficult to understand in terms
of radical intermediates why for the tert-butyldimethylsilyl
derivatives 1e′,g′ the enantioselectivities are dramatically in-
creased by such remote groups. This feature had so far not

714 J. Am. Chem. Soc., Vol. 120, No. 4, 1998
Adam et al.
been examined for trisubstituted olefins, since most employed
substrates contained the double bond to be oxidized in a ring
system (usually decreased ee values were obtained for larger
cycloalkenes) or were phenyl-substituted ethylenes without
variation of the steric bulk of the remote substituents.10,12,15,16
Let us now analyze the observed enantioselectivities in terms
of the alternative Katsuki metallaoxetane mechanism. The key
feature here is that the metallaoxetane is enantioselectively
formed and subsequently diastereoselectively dissociated into
the epoxide product, with or without the intervention of radicals.
This more complex mechanism is illustrated in detail for the
enol ethers 1a (Figure 3A) and 1e (Figure 3B) as model
substrates, viewed along the skewed side-on trajectory b, in
which energy-favored structures are coded in black, the others
in gray. In this [2 + 2] cycloaddition the olefinic substrate
approaches the oxo-metal bond crosswise and once the two
partners are close enough for bonding, the substrate may either
rotate clockwise or counterclockwise to afford the metalla-
oxetane intermediate; the rotation with the lower steric interac-
tions is selected. Thus, for the enol ether 1a (Figure 3A) the
clockwise rotation applies because it is sterically less encum-
bered since the bulky siloxy group avoids bumping into the
arene framework of the salen complex. Subsequently the
metallaoxetane opens up fast to the benzylic radical intermediate,
which then collapses by way of the epoxide to the S-configured,
R-hydroxylated ketone (S)-2a. Alternatively, the intermediary
epoxide may be generated directly from the metallaoxetane
adduct.28b [Note that the radical intermediates in the Jacobsen
(Figure 2A) and the Katsuki (Figure 3A) mechanisms are
identical! Moreover, since the stereochemical information at
the silyloxy-substituted site is lost in this oxidation process, it
is not possible to assess whether rotation about the C-C bond
in the radical intermediate takes place prior to epoxide ring
closure.]
in the enol-ether substrates, the present set of stereochemical
data (Table 1) are more adequately accommodated by the
metallaoxetane model.
Catalyst Structure. The best enantioselectivities for the
substrates 1 pertain to the 5,5′-dimethoxy-substituted (salen)-
MnIII catalyst 3b (Table 2), but the improvement is relatively
small (at best ∼15% in the ee values). These substituent effects
of the catalysts 3 on the enantioselectivity in the oxidation of
the silyl enol ethers and ketene acetals may be explained
primarily in terms of the electronic influence on the reactivity
of the active oxo-metal species rather than steric interactions.
The electron-donating substituents in the catalysts 3b,c decrease
the reactivity of the oxo species and the oxidation of the silyl
enol ethers with these less reactive oxo species involves a more
product-like transition state with more specific nonbonded
interactions and, consequently, enhanced enantioselectivities.13
That merely steric effects between the approaching substrate
and the catalyst are not the decisive factor becomes evident from
the comparison of the 5,5′-di-tert-butyl- (3a) and the 5,5′-
dimethoxy-substituted (3b) catalysts. Although the latter pos-
sesses the less sterically demanding substituents, it controls the
higher enantioselectivity, presumably through electronic effects.
This conclusion is further substantiated by the fact that sterically
shielding features of the triisopropylsiloxy group in the catalyst
3c do not raise the enantioselectivity compared to the methoxy
derivative 3b.
In summary, the stereocontrol in the catalytic, enantioselective
R hydroxylation presented here is imposed by the substitution
pattern and steric bulk of the enol ether substrates 1 and the
electronic nature of the 5,5′ substituents in the salen ligand of
the manganese catalyst 3. Besides the direct access to syntheti-
cally valuable, optically active R-hydroxy carbonyl products 2,
the silyl enol ethers 1 comprise mechanistically informative
substrates to probe steric and electronic effects in the catalytic,
enantioselective oxidation by (salen)MnIII complexes. The
favored attack of the olefinic substrate on the oxo-metal
complex has been identified as the skewed side-on trajectory b
(Figure 2). Of the two currently debated mechanisms for the
oxygen transfer, namely the radical versus the metallaoxetane
models, the latter accounts more adequately for the observed
enantioselectivity.
The favored route in the metallaoxetane model for the enol
ether 1e is displayed by the black structures in Figure 3B. Again
steric and electronic interactions of the phenyl group with the
salicylidene moiety are minimized by clockwise rotation in the
crosswise [2 + 2] encounter complex to produce the metal-
laoxetane. The latter opens up fast to the stabilized radical,
whose collapse to the epoxide and rearrangement leads to the
observed R-hydroxy ketone (R)-2e. [Note that rotation about
the C-C bond in the radical would have no consequence on
the enantioselectivity since the chirality center is permanently
fixed to the oxenoid oxygen atom (Figure 3B)!]
Acknowledgment. This work was supported by the Deutsche
Forschungsgemeinschaft (Sonderforschungsbereich 347 “Sele-
ktive Reaktionen metallaktivierter Moleku¨le”), the Bayerische
Forschungsstiftung (Bayerischer Forschungsverbund Katalyse-
FORKAT) and the Fonds der Chemischen Industrie.
How does the metallaoxetane mechanism cope with the fact
that remote bulky silyl substituents enhance drastically the
enantioselectivity (compare the trimethylsilyl derivatives 1e,g
with the tert-butyldimethylsilyl ones 1e′,g′ in Table 1).
A
Supporting Information Available: Synthetic details and
characteristic spectral data of the (salen)Mn(III) complexes 3b,c,
the silyl enol ethers and silyl ketene acetals 1a-h, the racemic
R-hydroxy ketones and esters 2a-h, the general catalytic
oxidation procedure, analytical protocols for measuring of the
enantiomeric excess and results of frontier orbital calculations
of substrates 1 (23 pages). See any current masthead page for
ordering and Internet access instructions.
buttressing effect in the highly substituted C-Mn bond of the
strained four-membered ring of the metallacycle appears to be
responsible, which promotes its faster ring opening to the radical
intermediate. Thus, metallaoxetane formation becomes less
reversible and the formation of the metallaoxetane from the
disfavored orientation (gray structure) will be supressed; as the
result the overall oxidation proceeds more enantioselectively.
Since the radical mechanism (Figure 2) does not explain the
enhanced enantioselectivity caused by remote bulky silyl groups
JA9726668