Construction of pseudo-heterochiral and homochiral Di-μ- oxotitanium(Schiff base) dimers and enantioselective epoxidation using aqueous hydrogen peroxide

Article abstract of DOI:10.1002/anie.200501318

(Chemical Equation Presented) Salen's lot: Meerwein-Ponndorf-Verley reduction of [Ti(salen) (OiPr)₂] (salen = bis(salicylidene) ethylenediaminato) complexes followed by treatment with water leads to the synthesis of pseudo-heterochiral and homochiral di-μ-oxotitanium dimers. These dimers, in particular (aR,S,Δ,aR,S,Δ)-dimer 1, efficiently catalyze the epoxidation of simple olefins using aqueous hydrogen peroxide (see scheme) with ee values up to 99% and turnover numbers up to 4600.

Full text of DOI:10.1002/anie.200501318

Angewandte
Chemie
Asymmetric Catalysis
DOI: 10.1002/anie.200501318
Construction of Pseudo-Heterochiral and
Homochiral Di-m-oxotitanium(Schiff base)
Dimers and Enantioselective Epoxidation Using
Aqueous Hydrogen Peroxide
Kazuhiro Matsumoto, Yuji Sawada, Bunnai Saito,
Ken Sakai, and Tsutomu Katsuki*
Remarkable advances in asymmetric synthesis using optically
active metal complexes as catalysts have been achieved in the
last half century.[1] Monometallic complexes are used as the
catalysts in most of the asymmetric reactions developed so
far. Particularly in the last two decades, it has been revealed
that chiral metallosalen complexes result in diverse and
excellent asymmetric catalysis.[2] Chiral salen ligands can be
synthesized in a single step from chiral diamine and chiral
and/or appropriately substituted salicylaldehydes and form
complexes with a variety of metal ions. Moreover, metal-
losalen complexes are conformationally flexible as a result of
the presence of two methylene carbon atoms and they can
adopt three different configurations (trans, cis-a, and cis-b).
Metallosalen complexes usually adopt trans configurations,
but they can be readily transformed in the presence of a
bidentate ligand into the corresponding cis-b complexes.[2,3]
Thus, various chiral metallosalen complexes have been
synthesized and used as catalysts for a wide range of
asymmetric reactions. Most of the metallosalen complexes
used are monomeric, but for some reactions, depending on
their mechanisms, dimeric or oligomeric metallosalen com-
plexes have proven to be superior catalysts to the corre-
sponding monomeric ones.[4] Nevertheless, the use of such
dimeric or oligomeric metallosalen complexes as catalysts has
been limited mainly because their synthesis involves laborious
routes, except for metallosalen complexes prepared by self-
assembly. Thus, di-m-oxotitanium(salen) complexes that can
be prepared spontaneously by treatment of monomeric
[Ti(salen)] complexes with water,[5] attracted our attention.[6]
Furthermore, in contrast to the usual monomeric [Ti(salen)]
complexes that adopt trans configurations, each {Ti(salen)}
unit of the di-m-oxo complexes take a cis-b configuration. In
contrast to the trans isomer, the cis-b isomer is chiral and
exists in enantiomeric forms (D or L; Figure 1). Thus, there
are six possible isomers for the di-m-oxotitanium(salen) dimer
[*] K. Matsumoto, Y. Sawada, B. Saito, K. Sakai, T. Katsuki
Department of Chemistry
Faculty of Science
Graduate School
Kyushu University
33, Hakozaki, Higashi-ku, Fukuoka 812-8581 (Japan)
Fax : (+81)92-642-2607
E-mail: katsuscc@mbox.nc.kyushu-u.ac.jp
K. Matsumoto, Y. Sawada, B. Saito, T. Katsuki
CREST
Japan Science and Technology Agency (JST) (Japan)
Angew. Chem. Int. Ed. 2005, 44, 4935 –4939
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4935

Communications
towards the heterochiral dimer as a consequence of a
favorable interaction between the enantiomeric monomers.
We also found that, in analogy with di-m-oxo dimer 1,
complex 3 that was prepared from [TiCl2C] also adopted a
homochiral aR,R,D,aR,R,D configuration (Scheme 1) and it
Figure 1. Enantiomeric isomers for the monomeric cis-b-titanium-
(salen) complex.
(two enantiomeric pairs and two meso isomers).[5b] However,
Belokonꢀ, North, et al. have reported that treatment of a
trans-[TiCl2(salen)] ([TiCl2A]) complex with water in the
presence of amine spontaneously gives homochiral dimer 1 as
the sole product [(RD,RD)-syn-1] (Figure 2a).[5] This result
indicated that the salen ligand A which included an (R,R)-
cyclohexanediamine moiety as a constituent part forced the
{Ti(salen)} unit to adopt the D configuration.
Scheme 1. Structural change of di-m-oxotitanium(salen) complex 3
bearing chiral salen ligand C as a result of the solvent or reagent.
served as an excellent catalyst for asymmetric sulfoxidation in
the presence of urea·hydrogen peroxide (UHP).[9] However,
complex 3 did not catalyze epoxidation. On the other hand, it
was found that the asymmetric sulfoxidation with 3 showed a
strong positive nonlinear effect and that di-m-oxo-3 and
monomeric trans-4 were readily interconverted by changing
the solvent. This result indicated that the mixing of aR,R,D
and the aS,S,L isomers made a heterochiral (aR,R,D,aS,S,L)
dimer preferentially.[10] Taking into account Tuchimotoꢀs and
our own results, we expected that the pseudo-heterochiral
(aR,R,D,aR,R,L)-6, in which each titanium ion carries the
same but conformationally enantiomeric ligand, would be
formed if the salen ligand became more flexible and if the
aR,R,L configuration that is otherwise unstable is sufficiently
stabilized by the interaction with the aR,R,D ligand. Further-
more, the unprecedented pseudo-heterochiral dimer might
show unique asymmetric catalysis. Herein, we describe the
construction of stable dimeric titanium/tetradentate Schiff
base complexes that provide unique reaction sites as shown by
in situ intramolecular Meerwein–Ponndorf–Verley (MPV)
reduction in combination with self-assembly of the resulting
titanium/tetradentate Schiff base complexes. We also describe
their use as catalysts in the presence of hydrogen peroxide for
asymmetric epoxidation.
Figure 2. a) Structure of 1. b) Structure of 2.
In contrast to the report by Belokonꢀ, North, et al.,
Tsuchimoto has recently reported that di-m-oxo dimer 2
bearing achiral ligand B is heterochiral: each {Ti(salen)} unit
has opposite chirality (Figure 2b) and X-ray diffraction
analysis has shown 2 to have a (D,L)-anti configuration.[7]
This structure indicates that the interaction between the D
and the L ligands is more favorable than the interaction
between two D (or two L) ligands. This situation is
reminiscent of asymmetric amplification (positive nonlinear
effect) observed in asymmetric addition of diethylzinc to
aldehydes using a chiral amino alcohol as the chiral auxiliary:
the enantiomeric excess of the chiral auxiliary correlates
nonlinearly with the enantioselectivity of the reaction, that is,
carrying out the reaction even with a chiral auxiliary of low
enantiomeric excess results in enantioselectivity similar to
that obtained with the enantiopure auxiliary.[8] This unusual
phenomenon has been rationally explained by self-recogni-
tion of the chirality of the complex upon its dimerization: the
equilibrium between enantiomeric monomers, as well as
homochiral and heterochiral dimers is weighted heavily
We tried reducing one of the two imine bonds of a salen
ligand to make it more flexible without losing its high
asymmetry-inducing ability. Since titanium isopropoxide is
known to undergo MPV reduction, we expected that the
[Ti(salen)(OiPr)2] complex might undergo intramolecular
MPV reduction of one or two of the imine bonds. Unfortu-
nately, no reaction was observed for [TiA(OiPr)2] (R = tBu)
that was prepared in situ and then left to stand. However, the
4936 www.angewandte.org
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2005, 44, 4935 –4939

Angewandte
Chemie
corresponding [TiC(OiPr)2] complex prepared in situ from
ligand C and Ti(OiPr)4 in dichloromethane was found to
undergo the desired MPV reduction at room temperature,
and the subsequent water treatment gave a new di-m-oxo
complex 6 that could be crystallized from heptane and
dichloromethane (Scheme 2). X-ray diffraction analysis
unambiguously demonstrated that the configuration of 6
was pseudo-heterochiral [(aR,R,D,aR,R,L)-anti].[11] Further-
(entry 1).[14] Epoxidation using 7 was performed in the
presence of 1.01 equivalents of aqueous (30%) hydrogen
peroxide at room temperature[15,16] and it was found that the
epoxidation proceeded with excellent enantioselectivity as
well as high yield (entry 2). The turnover number of 7 was
4600 when hydrogen peroxide was added slowly over a period
of 8 h (entry 4).
Epoxidation of other conjugated olefins also proceeded
smoothly with high enantioselectivity (entries 9–13). It is
noteworthy that epoxidation of (Z)-5-phenylpent-2-en-3-yne
(11) was stereospecific and gave the corresponding cis-
epoxide as a single product, albeit with a slightly inferior
enantioselectivity of 88% ee (entry 11). This result suggested
that oxygen transfer proceeds in a concerted rather than a
stepwise manner. This proposal was also supported by the fact
that epoxidation of styrene (12) was highly enantioselective
(entry 12). Moreover, the epoxidation of the nonconjugated
olefin 1-octene (14) proceeded with slightly reduced but good
enantioselectivity, albeit somewhat slowly (entry 14).
Although the reaction mechanism of the present epox-
idation is unclear, a peroxotitanium species 15 or 16 is
considered to act as the active species because the epoxida-
tion is stereospecific (Figure 3).[17] Furthermore, we speculate
that the peroxotitanium species may be activated by an
intramolecular hydrogen bond with the amine proton. This
speculation might explain why complex 3 that does not
possess such an NH group cannot catalyze the epoxidation.
The oxidation of methyl phenyl sulfide was also examined
in dichloromethane with 6 as the catalyst in the presence of
UHP. The reaction smoothly proceeded but the enantiose-
lectivity was moderate (67% ee), although the oxidation with
3 under the same conditions showed a high enantioselectivity
of 95% ee. It is, however, noteworthy that the sense of
asymmetric induction by 6 was opposite to that by 3, thus
suggesting that the structure of the active species derived
from 6 was different from 5.
À
more, the C N bond lengths indicated that one of the two
imine bonds of each salen ligand C was reduced to a single
bond and the phenolic oxygen atom ortho to the reduced
imine group occupied the apical position.[12] Encouraged by
this result, we also synthesized the (aRS,aRS)-di-m-oxotita-
nium complex 7, bearing the corresponding half-reduced
salen ligand F, according to the method described for the
preparation of 6. It is, however, noteworthy that complex 7
was found to adopt a homochiral configuration (aR,S,-
D,aR,S,D) by X-ray diffraction analysis (Scheme 2).[11] These
results suggested that the configuration of a di-m-oxotitanium
complex bearing a half-reduced salen ligand is determined by
equilibrium involving ligand chirality, its structural flexibility,
and weak intra- and interligand interactions such as CH–p
interactions. To our delight, complexes 6 and 7 were much
more stable than complex 3: although complex 3 immediately
dissociated into the corresponding monomeric {Ti(salen)}
species 4 in methanol at room temperature (Scheme 1),
complex 6 was stable in [D4]MeOH for at least 5 h and
complex 7 was stable even under aqueous epoxidation
conditions for at least 24 h. On the basis of these findings,
we examined epoxidation using hydrogen peroxide in the
presence of complex 6 or 7.[13]
The catalytic activities of 6 and 7 for epoxidation were
examined with 1,2-dihydronaphthalene (8) as the test mate-
rial (Table 1, entries 1–8). Epoxidation using 6 as the catalyst
in the presence of the urea·hydrogen peroxide adduct
proceeded with good enantioselectivity, but was slow and
the turnover number (TON) of 6 was modest (TON = 14)
Scheme 2. Syntheses of di-m-oxotitanium complexes 6 and 7.
Angew. Chem. Int. Ed. 2005, 44, 4935 –4939
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
4937

Communications
Table 1: Asymmetric epoxidation using aqueous hydrogen peroxide as
In conclusion, we were able to synthesize stable di-m-oxo
dimers—not only a homochiral 7 but also an unprecedented
pseudo-heterochiral di-m-oxo dimer 6—by using an in situ
intramolecular Meerwein–Ponndorf–Verley reduction in
combination with self-assembly of the resulting titanium/
tetradentate Schiff base complexes. Furthermore, we were
able to show that these stable di-m-oxo dimers (6 and 7)
catalyzed epoxidation with hydrogen peroxide; in particular,
complex 7 efficiently catalyzed a highly enantioselective
epoxidation using aqueous (30%) hydrogen peroxide as the
oxidant. Other applications of the current oxidation are under
investigation.
the terminal oxidant.[a]
Entry Substrate Solvent
t
Yield [%][b] ee [%]
Config.[c]
1
2
3
4
5
8
8
8
8
8
CH2Cl2
CH2Cl2
CH2Cl2
CH2Cl2
toluene
24
12
72
48
18
14
>99
>99
92
>99
>99
97
83[d,e] 1S,2R
>99[d]
1R,2S
>99[d,f] 1R,2S
>99[d,g] 1R,2S
>99[d]
>99[d]
97[d]
1R,2S
1R,2S
1R,2S
6
7
8
8
ethyl acetate 18
THF
85
8
8
MeOH
NR
9
9
ethyl acetate 24
ethyl acetate 24
ethyl acetate 24
87
85
64
90
75
70
99
1R,2S
5R,6S
2S,3R
R
1S,2R
S
98[h]
88[i]
93[j]
95[k]
82[l,m]
Experimental Section
10
11
12
13
14
10
11
12
13
14
7: Ti(OiPr)4 (2equiv) was added to a solution of salen ligand
D
(1 equiv) in dry dichloromethane in a nitrogen atmosphere and the
resultant solution was stirred at room temperature. After 3 days,
water (4 equiv) was added and the reaction mixture was stirred for
2h. The resulting yellow precipitate was collected by filtration and
recrystallized from diethyl ether and dichloromethane to give
crystalline complex 7 in 60% yield.
General procedure for epoxidation: Titanium complex 7 (1.8 mg,
1 mmol) and olefin (0.1 mmol) were dissolved in an appropriate
solvent (1.0 mL) in a nitrogen atmosphere. After addition of 30%
aqueous hydrogen peroxide (0.101 mmol), the resultant mixture was
stirred at room temperature for the time indicated in Table 1. The
solvent was removed in vacuo and the residue was purified by
chromatography on silica gel (pentane/Et2O 40:1) to give the
corresponding epoxide. The ee values were determined by HPLC
CH2Cl2
CH2Cl2
CH2Cl2
24
24
48
[a] Reactions were carried out at room temperature with a molar ratio of
substrate/catalyst 7/aq H2O2 =1:0.01:1.01, unless otherwise stated.
[b] Determined by 1H NMR (400 MHz) spectroscopic analysis. [c] Deter-
mined by comparison of the elution order with that of the authentic
sample in HPLC analysis and/or comparison of the chiroptical data with
the literature value. [d] Determined by HPLC analysis on a chiral
stationary phase (Daicel Chiralcel OB-H; hexane/iPrOH 99:1). [e] Reac-
tions were carried out at room temperature with a molar ratio of
substrate/catalyst 6/UHP=1:0.01:1. [f] 0.1 mol% of 7 was used.
[g] 0.02 mol% of 7 was used and hydrogen peroxide was added over
the period of 8 h. [h] Determined by HPLC analysis on a chiral stationary
phase (Daicel Chiralpak AS-H; hexane/iPrOH 99.9:0.1). [i] Determined
by HPLC analysis on a chiral stationary phase (Daicel Chiralcel OD-H;
hexane/iPrOH=99:1). [j] Determined by HPLC analysis on a chiral
stationary phase (Daicel Chiralcel OD-H; hexane:iPrOH 99.9:0.1).
[k] Determined by HPLC analysis on a chiral stationary phase (Daicel
Chiralcel OB-H; hexane/iPrOH=99:1). [l] 3 mol% of 7 was used.
1
on a chiral stationary phase or by H NMR analysis using [Eu(hfc)3]
(hfc = 3-(heptafluoropropylhydroxymethylene)-d-camphorate)
under the conditions described in the footnotes to Table 1.
Received: April 15, 2005
Published online: July 6, 2005
1
[m] Determined by H NMR spectroscopic analysis using a chiral shift
Keywords: asymmetric catalysis · enantioselectivity ·
epoxidation · hydrogen peroxide · titanium
reagent [Eu(hfc)3].
.
[1] a) Comprehensive Asymmetric Catalysis (Eds.: E. N. Jacobsen,
A. Pfaltz, H. Yamamoto), Springer, Berlin, 1999; b) Catalytic
Asymmetric Synthesis, 2nd ed. (Ed.: I. Ojima), Wiley-VCH, New
York, 2000; c) Transition Metals For Organic Synthesis: Building
Blocks And Fine Chemicals (Eds.: M. Beller, C. Bolm), Wiley-
VCH, Weinheim, 2004.
[2] a) L. Canali, D. C. Sherrington, Chem. Soc. Rev.1999, 28, 85 – 93;
b) T. Katsuki, Synlett 2003, 281 – 297; c) T. Katsuki, Chem. Soc.
Rev. 2004, 33, 437 – 444; d) J. F. Larrow, E. N. Jacobsen, Top.
Organomet. Chem. 2004, 6, 123 – 152; e) P. G. Cozzi, Chem. Soc.
Rev. 2004, 33, 410 – 421.
[3] A cis-a isomer is generally less stable than the corresponding
cis-b isomer: see ref. [2c].
[4] a) J. M. Ready, E. N. Jacobsen, J. Am. Chem. Soc. 2001, 123,
2687 – 2688; b) Z. Luo, Q. Liu, L. Gong, X. Cui, A. Mi, Y. Jian,
Angew. Chem. 2002, 114, 4714 – 4717; Angew. Chem. Int. Ed.
2002, 41, 4532– 4535.
[5] a) Y. Belokonꢀ, S. Caveda-Cepas, B. Green, N. Ikonnikov, V.
Khrustalev, V. Larichev, M. Moscalenko, M. North, C. Orizu, V.
Tararov, M. Tasinazzo, G. Timofeeva, L. Yashkina, J. Am. Chem.
Soc. 1999, 121, 3968 – 3973; b) V. Tararov, V. Larichev, M.
Moscalenko, L. Yashkina, V. Khrustalev, M. Antipin, A. Borner,
Y. Belokonꢀ, Enantiomer 2000, 5, 169 – 173.
Figure 3. Possible peroxo intermediates for the epoxidation.
4938 www.angewandte.org
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2005, 44, 4935 –4939

Angewandte
Chemie
[6] Although the construction of a polymeric metallosalen complex
by self-assembly has been reported, the catalysis of the
polymeric complex is essentially the same as that of the parent
monomeric complex: G. A. Morris, S. T. Nguyen, J. T. Hupp, J.
Mol. Catal. A 2001, 174, 15 – 20.
[7] M. Tsuchimoto, Bull. Chem. Soc. Jpn. 2001, 74, 2101 – 2105.
[8] a) C. Puchot, O. Samuel, E. Dunach, S. Zhao, C. Agami, H. B.
Kagan, J. Am. Chem. Soc. 1986, 108, 2353 – 2357; b) N. Oguni, Y.
Matsuda, T. Kaneko, J. Am. Chem. Soc. 1988, 110, 7877 – 7878;
c) M. Kitamura, S. Okada, S. Suga, R. Noyori, J. Am. Chem. Soc.
1989, 111, 4028 – 4036; d) C. Girard, H. B. Kagan, Angew. Chem.
1998, 110, 3088 – 3127; Angew. Chem. Int. Ed. 1998, 37, 2 92 3 –
2959.
[9] B. Saito, T. Katsuki, Tetrahedron Lett. 2001, 42, 3873 – 3876.
[10] B. Saito, T. Katsuki, Tetrahedron Lett. 2001, 42, 8333 – 8336.
[11] CCDC-259826 (6) and CCDC-264266 (7) contain the supple-
mentary crystallographic data for this paper. These data can be
obtained free of charge from the Cambridge Crystallographic
Data Centre via www.ccdc.cam.ac.uk/data_request/cif.
[12] This partially reduced structure was also supported by FABMS
analysis ([C120H92N4O6Ti2]+: m/z = 1780.6): JEOL JMX-SX/SX
102A spectrometer by using m-nitrobenzyl alcohol as the matrix.
[13] For a review of asymmetric epoxidation, see: a) E. N. Jacobsen,
M. H. Wu in Comprehensive Asymmetric Catalysis, Vol. II (Eds.:
E. N. Jacobsen, A. Pfaltz, H. Yamamoto), Springer, Heidelberg,
1999, chap. 21, pp. 649 – 677; b) T. Katsuki in Comprehensive
Coordination Chemistry II, Vol. 9 (Ed.: J. McCleverty), Elsevier
Science, Oxford, 2003, chap. 9.4, pp. 207 – 264; c) T. Katsuki,
K. B. Sharpless, J. Am. Chem. Soc. 1980, 102, 5974 – 5976.
[14] Complex 6 was poorly soluble in methanol relative to complex 3,
and the reaction was carried out in dichloromethane at room
temperature.
[15] For enantioselective epoxidation using hydrogen peroxide, see:
a) M. K. Tse, C. Doebler, S. Bhor, M. Klawonn, W. Maegerlein,
H. Hugl, M. Beller, Angew. Chem. 2004, 116, 5367 – 5372;
Angew. Chem. Int. Ed. 2004, 43, 5255 – 5260; b) R. I. Kureshy,
N. H. Khan, S. H. R. Abdi, S. Singh, I. Ahmed, R. S. Shukla,
R. V. Jasra, J. Catal. 2003, 219, 1 – 7; c) L. Shu, Y. Shi,
Tetrahedron 2001, 57, 5213 – 5218; d) S. Arai, H. Tsuge, T.
Shioiri, Tetrahedron Lett. 1998, 39, 7563 – 7566; e) P. Pietkainen,
Tetrahedron 1998, 54, 4319 – 4326; f) A. Berkessel, M. Frauen-
kron, T. Schwenkreis, A. Steinmetz, G. Baum, D. Fenske, J. Mol.
Catal. A 1996, 113, 321 – 342; g) P. Pietkainen, Tetrahedron Lett.
1994, 35, 941 – 944; h) R. Irie, N. Hosoya, T. Katsuki, Synlett
1994, 255 – 256; i) S. Juliꢁ, J. Masana, J. C. Vega, Angew. Chem.
1980, 92, 968 – 969; Angew. Chem. Int. Ed. Engl. 1980, 19, 92 9 –
931; j) S. Colonna, H. Molinari, S. Banfi, S. Juliꢁ, J. Masana, A.
Alvalez, Tetrahedron 1983, 39, 1635 – 1641.
[16] For examples of nonstereoselective but highly efficient epox-
idation using aqueous hydrogen peroxide as oxidant, see:
a) D. E. De Vos, J. L. Meinershagen, T. Bein, Angew. Chem.
1996, 108, 2355 – 2357; Angew. Chem. Int. Ed. Engl. 1996, 35,
2211 – 2213; b) K. Sato, M. Aoki, M. Ogawa, T. Hashimoto, D.
Panyella, R. Noyori, Bull. Chem. Soc. Jpn. 1997, 70, 905 – 915;
c) J. Rudolph, K. L. Reddy, J. P. Chiang, K. B. Sharpless, J. Am.
Chem. Soc. 1997, 119, 6189 – 6190; d) M. C. White, A. G.
Doyle. E. N. Jacobsen, J. Am. Chem. Soc. 2001, 123, 7194 –
7195; e) K. Komatsu, K. Yonehara, Y. Sumida, K. Yamaguchi,
S. Hikichi, N. Mizuno, Science 2003, 300, 964 – 966.
[17] At the moment, we cannot determine which titanium species,
monomeric or dimeric, participates in activation of hydrogen
peroxide. Study on the active species in this epoxidation is under
way.
Angew. Chem. Int. Ed. 2005, 44, 4935 –4939
ꢀ 2005 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 4939