B. Saito, T. Katsuki / Tetrahedron Letters 42 (2001) 8333–8336
8335
This strongly suggests that the square planar structure
of the salen ligand is readily changeable to the cis-b
structure when a bidentate ligand is coordinated to the
titanium ion.
into a monomeric Ti(salen) complex 6 and reacts with
hydrogen peroxide, a bidentate ligand, to give the
corresponding peroxo species 4 which undergoes sul-
foxidation (Scheme 4).
These results also suggested that the sulfoxidation
using Ti(salen) complex 2 as the catalyst would show
non-linear relationship between the enantiomeric
excesses of the catalyst and of sulfoxide. Thus, we
examined the sulfoxidation of methyl phenyl sulfide
with di-m-oxo complexes 2 of different enantiomeric
excesses, which were prepared by mixing (R,R)- and
To prove this consideration, we treated complex (R)-6
with aqueous H O2 (2 equiv.) in CDCl –MeOH-d
4
2
3
1
(1:1) and performed H NMR measurement of the
4,5
1
mixture. In the H NMR spectrum of (R)-6 which is
C -symmetric, protons (Ha and Ha%) at the carbons
2
bearing imino group and protons (Hb and Hb%) at the
imino carbons show the identical chemical shifts,
respectively (Table 1). On the other hand, protons
(Ha and Ha%) and (Hb and Hb%) of complex (R,R)-2
(
S,S)-di-m-oxo complex 2 (3.6 mg [(R,R)+(S,S)], 2.0
mmol) in methanol (750 ml). Although the mixing
caused turbidity, the mixture was used for the sulfoxi-
dation without removing the precipitate. As shown in
Fig. 1, a positive non-linear effect was actually
observed. We next isolated the precipitate and mea-
which are not C -symmetric show different chemical
2
1
shifts. The H NMR spectrum of the mixture (6+
H O ) showed the signals similar to those of complex
2
2
2. This strongly suggested that the complex generated
by mixing complex 6 and H O is not C -symmetric
1
sured its H NMR spectrum that showed 44 proton
2
2
2
signals similar to the spectrum of (R,R)-di-m-oxo
complex 2. However, the chemical shifts of the pro-
tons of the precipitate were slightly different from
those of (R,R)-di-m-oxo complex 2, suggesting that
the precipitate was not an (R,R)- but a racemic
and probably has cis-b structure. These results
strongly supported the above consideration that com-
plex 6 reacts with hydrogen peroxide to give the corre-
sponding peroxo species 4.
(
R,S)-di-m-oxo complex. Indeed, a chloroform solu-
tion of the precipitate showed very small optical rota-
2
4
tion: [h] =−0.06 (c 0.038, CHCl ), while the specific
H2O2
D
3
MeOH
optical rotation of (R,R)-di-m-oxo complex 2 was
O
2
4
[
h] =−372 (c 0.030, CHCl ). FABMS analysis of the
N
O
N
O
D
3
Ti
di-µ-oxo Ti(salen) 2
6
precipitate showed the parent peak (m/z=1776.5662)
indicating that it was an (R,S)-di-m-oxo complex.
O
4
The above results meant that (R,R)- and (S,S)-di-m-
oxo complexes 2 are readily dissociated in methanol
into the corresponding monomeric isomers [(R)-6 and
R-S*O-R', H2O R-S-R', MeOH
(
S)-6] which are equilibrated with the racemic (R,S)-
Scheme 4.
di-m-oxo complex and, due to the low solubility and
probably the high stability of the (R,S)-di-m-oxo com-
plex, the equilibrium is shifted to the (R,S)-di-m-oxo
complex and the concentration of (S)-6 in the
methanolic solution is reduced, showing the positive
non-linear effect as observed (Scheme 3).
Acknowledgements
From these results, we considered that (R,R)-di-m-oxo
complex 2 dissolved in methanol rapidly dissociates
We are grateful to Ms. Mie Iriguchi, this department,
for the measurements of MS spectroscopy.
Table 1. NMR spectra of complexes (2, 6, and 4): chemical shifts of Ha, Ha%, Hb and Hb%
H H
b
H 'Hb'
a
a
N
N
Ti
Complex
Solvent
CDCl3
Chemical shift (ppm)
Ha
Ha%
Hb
Hb%
2
6
3.18
3.22
3.29
4.20
3.22
4.11
8.43
8.65
8.55
8.68
8.65
8.66
CD OD
3
6
+H O2 (4)
CDCl /CD OD
2
3
3
Saito, Bunnai
