Enantioselective titanium-catalyzed sulfides oxidation: Novel ligands provide significantly improved catalyst life

Full text of DOI:10.1021/jo960359z

J . Org. Chem. 1996, 61, 5175-5177
5175
En a n tioselective Tita n iu m -Ca ta lyzed
Su lfid es Oxid a tion : Novel Liga n d s P r ovid e
Sign ifica n tly Im p r oved Ca ta lyst Life
Fulvio Di Furia, Giulia Licini,* Giorgio Modena, and
Riccardo Motterle
We now report on the synthesis of this new class of
chiral titanium peroxo complexes and on their reactivity
in the catalytic enantioselective oxidation of organic
sulfides.11,12
Universita` degli Studi di Padova, Dipartimento di Chimica
Organica, Centro Meccanismi Reazioni Organiche del CNR,
Via Marzolo 1, I-35131 Padova, Italy
William A. Nugent
Resu lts a n d Discu ssion
The Du Pont Company, Central Research and Development,
P.O. Box 80328, Wilmington, Delaware 19880-0328
The complex 2a formed in situ by addition of titanium
tetraisopropoxide to (S,S,S)-triisopropanolamine 1a , af-
fords, upon addition of tert-butyl hydroperoxide, the
corresponding peroxo complex 3a , as indicated by 1H
NMR experiments (Scheme 1).
Received February 21, 1996
Catalytic enantioselective oxidations are among the
most interesting transformations in asymmetric synthe-
sis, as indicated by the large amount of research carried
out in this field.1 Although some very satisfactory results
have already been obtained, more work is needed before
procedures of general applicability and a complete un-
derstanding of species involved in the oxidative processes
become available. Excellent results have been obtained
in allylic alcohols2 and sulfides oxidation3-5 with chiral
titanium peroxo species bearing C2 symmetric diols as
ligands (e.g. tartaric esters). A drawback of such Ti(IV)/
(+)-diethyltartrate/alkyl hydroperoxide reagents is that
they have a small turnover number.6 In addition, the
structure of the real oxidants and of their precursors is
not well defined. In fact, owing to the presence of
different species and to the complexity of the equilibrium
processes occurring in solution, these systems have not
yet been completely characterized.1,7
The peroxotitanium complex 3a is in equilibrium with
its precursor 2a . The equilibrium constant has been
determined to be ) 3.5 at 22 °C in deuterochloroform.
1
The H NMR spectrum of the peroxo species 3a , as well
as that of the precursor 2a , shows a single set of signals
at δ ) 1.11, 2.82-2.91, and 5.55 ppm for the methyl,
methylene, and methine groups, respectively. The peroxo
species thus obtained is capable of oxidizing sulfides to
sulfoxides. Reactivity and enantioselectivity have been
studied by using p-tolyl methyl sulfide 4a as a model
substrate. The influence of parameters such as the
metal/oxidant ratio and the nature of the ligand and of
the hydroperoxide together with the effect of the tem-
perature and of the solvent have also been tested.
The catalytic nature of the system is demonstrated
by the experiments reported in Table 1 with ligand
(R,R,R)-1b.13
Aiming at overcoming these problems, we decided to
investigate chiral peroxotitanium complexes containing
other ligands. In particular we focused our attention on
C3 symmetric, chiral trialkanolamines 1 ligands.8 Zir-
conium complexes of the (S,S,S)-triisopropanolamine 1a
have been successfully used in the enantioselective ring
opening of meso epoxides with silyl azides (ee up to 93%).9
Enantiopure homochiral trialkanolamines should be
particularly suitable ligands for titanium since they are
tetradentate (thus giving rise to very robust complexes)
and highly symmetric ligands, and they should afford
monomeric peroxotitanium species.10
The data of Table 1 clearly show that the system
operates under truly catalytic conditions (0.01 equiv of
chiral catalyst) as far as the enantioselections are
concerned. At any rate, in order to speed up the reac-
tions, 0.1 equiv of catalyst with respect to the oxidant
have been routinely used (see Experimental Section).
The effect of the nature of the ligand and of the
hydroperoxide has been examined. The results are
collected in Table 2.
Both the nature of the ligand and of the hydroperoxide
are relevant parameters. (R,R,R)-Tris(2-hydroxy-2-phen-
ylethyl)amine (1b) and cumyl and trityl hydroperoxide
(entries 5 and 6) afford the best enantioselections, even
though these are still low. The sulfoxides obtained have
an absolute configuration which is the opposite of that
of the ligand. In all the reactions a sizeable amount of
sulfone is formed (see also data in Table 1). This feature
will be discussed in some more detail later. Out of trend
are the results obtained with tert-butyl ligand 1c that,
with the three different hydroperoxides, affords rather
(1) Ojima, I. Catalytic Asymmetric Synthesis; VCH Publishers,
Inc.: New York, 1993. Noyori, R. Asymmetric Catalysis in Organic
Synthesis; J ohn Wiley & Sons, Inc.: New York, 1994.
(2) Gao, Y.; Hanson, R. M.; Klunder, J . M.; Ko, S. Y.; Masamune,
H.; Sharpless, K. B. J . Am. Chem. Soc. 1987, 109, 5765.
(3) Di Furia, F.; Modena, G.; Seraglia, R. Synthesis 1984, 325.
(4) (a) Pitchen, P.; Dunach, E.; Deshmukh, M. N.; Kagan, H. B. J .
Am. Chem. Soc. 1984, 106, 8188. (b) Zhao, S. H.; Samuel, O.; Kagan,
H. B. Tetrahedron 1987, 43, 5132. Brunel, J . M.; Diter, P.; Duetsch,
M.; Kagan, H. B. J . Org. Chem. 1995, 60, 8086.
(5) Komatsu, N.; Hashizume, M.; Sugita, T.; Uemura, S. J . Org.
Chem. 1993, 58, 4529.
(11) Part of this work has been presented at the COFEM Conference
“Giornate di Chimica Fisica Organica e Meccanicistica”, Perugia, Italy,
September 24-27, 1995.
(6) Lower Ti(IV)/hydroperoxide ratios (1/20 for Sharpless reagent2
and 1/5 for the Kagan one4b) could be employed by carrying out the
oxidations in the presence of activated molecular sieves.
(7) Conte, V.; Di Furia, F.; Licini, G.; Modena, G.; Sbampato, G. In
Dioxygen Activation and Homogeneous Catalytic Oxidation; Simandi,
L. I., Ed.; Elsevier Science Publishers B.V.: Amsterdam, The Neth-
erlands, 1991; p 385.
(12) After our manuscript submission, the X-ray crystal structure
of the achiral dimer ((η2-tert-butylperoxo)titanatrane)2 has been re-
ported. Such a peroxo complex oxidizes methyl benzyl sulfide to the
corresponding sulfoxide (CH2Cl2, 0 °C, 91% yield): Boche, G.; Mo¨bus,
K.; Harms, K.; Marsh J . Am. Chem. Soc. 1996, 118, 2770.
(13) The reactions have been performed at -20 °C in 1,2-dichloro-
ethane, with a substrate concentration ) 0.2 M and cumyl hydroper-
oxide as oxidant. The preformed catalyst was obtained by mixing
stoichiometric quantities of Ti(i-PrO)4 and (+)-1b in 1,2-dichloroethane
at rt and removal of the solvent under vacuum.
(8) Nugent, W. A.; Harlow, R. L. J . Am. Chem. Soc. 1994, 116, 6142.
(9) Nugent, W. A. J . Am. Chem. Soc. 1992, 114, 2768.
(10) 1H and 13C NMR spectra and X-ray crystallographic analysis
show that, unlike zirconium derivatives, titanium complexes of trial-
kanolamine 1a are C3 symmetric monomers.8,9
S0022-3263(96)00359-3 CCC: $12.00 © 1996 American Chemical Society

5176 J . Org. Chem., Vol. 61, No. 15, 1996
Notes
Ta ble 2. Effect of Liga n d s 1 a n d of th e Hyd r op er oxid e
on th e Asym m etr ic Oxid a tion of Meth yl p-Tolyl
Su lfid e (4a )
Sch em e 1
time yield
(h)
5/6
ee 5 absol
no.
ligand
R′
t-Bu
PhCMe2
Ph3C
(%)a
(%)a (%)b configc
1
2
3
4
5
6
7
8
9
(S,S,S)-1a
(S,S,S)-1a
(S,S,S)-1a
(R,R,R)-1b t-Bu
(R,R,R)-1b PhCMe2
(R,R,R)-1b Ph3C
(R,R,R)-1c t-Bu
(R,R,R)-1c PhCMe2
(R,R,R)-1c Ph3C
240
82 86:14
11
19
13
29
36
35
3
R
R
R
S
S
S
R
S
S
75 100 76:24
1320 100 76:24
45 100 89:11
40 100 84:16
256
115
Ta ble 1. Effect of th e Ti(IV)/Cu m yl Hyd r op er oxid e Ra tio
on th e Asym m etr ic Oxid a tion of Meth yl p-Tolyl Su lfid e
(4a ) Usin g th e P r efor m ed Ca ta lyst 2b12
98 87:13
91 90:10
64 100 89:11
1800 75 96:4
29
10
a
Based on the oxidant and determined by gas chromatographic
analysis. Determined by 1H NMR in the presence of (R)-(-)-1-
(9-anthryl)-2,2,2-trifluoroethanol. c Determined by comparison with
the [R]D reported in the literature.4a,14
b
time
(h)
yield
(%)a
5a /6a
ee 5a
absol
Ti(IV)/oxidant
(%)a
(%)b
configc
1:2
16
16
16
56
56
100
100
100
100
84
85:15
84:16
87:13
84:16
86:14
87:13
39
38
38
40
40
29
S
S
S
S
S
S
1:10
1:25
1:50
1:100
1:1000
162
64
a
Based on the oxidant and determined by gas chromatographic
analysis. Determined by 1H NMR in the presence of (R)-(-)-1-
b
(9-anthryl)-2,2,2-trifluoroethanol. c Determined by comparison with
the [R]D reported in the literature.4a,14
different enantioselections and the opposite absolute
configuration of the sulfoxide with tert-butyl hydroper-
oxide (entry 7).
F igu r e 1. Dependence of the concentration of products [(2)
p-tolyl methyl sulfide; (9) p-tolyl methyl sulfoxide; (() p-tolyl
methyl sulfone] (%) and of (b) the ee (%) on the time (h) in
the oxidation of p-tolyl methyl sulfide with cumyl hydroper-
oxide catalyzed by (R,R,R)-1b [Ti(i-PrO)N(CH2CHPhO)3] )
10-2 M in DCE at -20 °C.
Other parameters such as the temperature and the
solvent were examined. A decrease of the temperature
from 20 to 0 and -20 °C results in an enhancement of
the enantioselection (ee ) 28, 31, and 36%, respectively).
The nature of the solvent appears to have a minor effect
on the enantioselectivity. At any rate, chlorinated sol-
vents seem to be particularly suitable for these oxida-
tions. The information so far obtained is that the most
effective system involves (R,R,R)-tris(2-hydroxy-2-phen-
ylethyl)amine (1b) as ligand and cumyl hydroperoxide
as oxidant. A preliminary kinetic study on such a system
has been carried out (Figure 1).
3a gives the S sulfoxide with an ee ) 33% (50% conver-
sion). An interesting aspect, which is rather unusual in
the chemistry of the titanium peroxo complexes,16 is that
the rates of the two subsequent oxidations, sulfide to
sulfoxide and sulfoxide to sulfone, are comparable. This
behavior deserves a more detailed investigation.
Different aryl alkyl sulfides have also been employed
(Table 3). The experiments have been performed at 0
°C in order to accelerate the reactions.
It is observed that p-tolyl methyl sulfone is formed from
the very beginning of the reaction. Therefore, two
different asymmetric processes are involved. One is the
asymmetric oxidation to sulfoxide, the other is the kinetic
resolution via oxidation to sulfone. Fortunately, both
processes work in the same direction. In fact, the S
sulfoxide is formed preferentially (ee ) 29% when almost
no sulfone is present), and the oxidation to the sulfone
of the R enantiomer is faster than that of the S one.15
The kinetic resolution of racemic methyl p-tolyl sulfoxide
The results reported in Table 3 show that both the
nature of the alkyl moiety and the electronic effect of the
aryl substituents play a relevant role in the enantiose-
lections. The aryl branched alkyl sulfides 4g,h give the
best results, ee ) 60 ÷ 84% (R ) tert-butyl and benzyl,
last three entries). Lower temperatures and an increase
(15) In sulfoxide asymmetric synthesis both enantioselective oxida-
tion and kinetic resolution have been contemporarily observed and also
in other chiral metal-catalyzed processes. Cf. ref 5. Komatsu, N.;
Hashizume, M.; Sugita, T.; Uemura, S. J . Org. Chem. 1993, 58, 7624;
Nagata, T.; Imagawa, K.; Yamada, T.; Mukaiyama, T. Bull. Chem. Soc.
J pn. 1995, 68, 3241.
(14) Sugimoto, T.; Kokubo, T.; Miyazaki, J .; Tanimoto, S.; Okano,
M. J . Chem. Soc., Chem. Commun. 1979, 402. Sakuraba, H.; Natori,
K.; Tanaka, Y. J . Org. Chem. 1991, 56, 4124.
(16) Bonchio, M.; Campestrini, S.; Conte, V.; Di Furia, F.; Moro, S.
Tetrahedron 1995, 51, 12363.

Notes
Ta ble 3. Asym m etr ic Oxid a tion of Ar yl Meth yl
J . Org. Chem., Vol. 61, No. 15, 1996 5177
589 nm corresponding to the sodium D line. Radial chromatog-
raphy was performed on a Chromatotron 7924 T (Harrison
Research) over silica gel 60 (Merck, TLC PF 254). Gas chro-
matographic analyses were performed on a Varian 3700 GC
equipped with a 0.5 m × 2 mm glass column packed with 3%
FFAP on Chromosorb W AW DMCS (80-100 mesh), eicosane
as external standard.
Su lfid es 4
Ch em ica ls. tert-Butyl hydroperoxide (Fluka, 80%, 20% di-
tert-butyl peroxide) was purified by distillation under vacuum
(bp 31 °C/16 Torr) and stored at 0 °C. Cumene hydroperoxide
(Fluka) was stored under molecular sieves at 0 °C. Trityl
hydroperoxide was prepared following a literature procedure.17
Titanium(IV)-tetraisopropoxide (Aldrich) was distilled under
vacuum (bp 60-63 °C/0.1 Torr). 1,2-Dichloroethane was washed
three times with 10% concentrated H2SO4 and with several times
until pH ) 7, dried overnight over CaCl2, distilled over P2O5,
and stored over molecular sieves. Sulfides were prepared
accordingly to the literature by alkylation of the corresponding
thiols. Enantiopure trialkanolamines 1a -c were prepared
following the literature procedure.8
yield
(%)a
5/6
ee 5
absol
no.
Ar
p-Tol
2-Napt
p-NO2C6H4
p-MeOC6H4
p-Tol
p-Tol
Ph
Ph
Ph
R
(%)b
(%)c
configd
4a
4b
4c
4d
4e
4f
4g
4g
4h
Me
Me
Me
Me
98
86
88
94
99
99
98
96
62:38
60:40
26:74
68:32
64:36
58:42
60:40
63:37
77:23
45
38
15
41
38
44
60
66e
84
S
S
S
S
S
S
S
S
S
Et
n-Bu
t-Bu
t-Bu
CH2Ph
94
Asym m etr ic Oxid a tion . A typical procedure is the follow-
ing: in a 10 mL volumetric flask Ti(i-PrO)4 (0.015 mL, 0.054
mmol), (-)-(R,R,R)-tris(2-hydroxy-2-phenylethyl)amine (24.5 mg,
0.065 mmol) and the methyl p-tolyl sulfide (2a ) (150 mg, 1.084
mmol) are dissolved in dry 1,2-dichloroethane. After cooling to
-20 °C, cumyl hydroperoxide (0.040 mL, 0.544 mmol) is added
under magnetic stirring. The solution is stirred at -20 °C until
disappearance of the oxidant, warmed at room temperature, and
poured into a 5% sodium metabisulfite aqueous solution. After
filtration through Celite, the mixture is extracted with chloro-
form. The organic layers are washed once with NaOH 5%
aqueous solution and with brine and dried over MgSO4, and the
solvent was removed under vacuum. The products are purified
via radial chromatography over silica gel (petroleum ether/ethyl
acetate). Ee’s were determined directly on the reaction mixture
before purification by 1H NMR in the presence of (R)-(-)-1-(9-
anthryl)-2,2,2-trifluoroethanol (Fluka).
a
Isolated yields based on the oxidant. b Determined by 1H NMR.
c Determined by 1H NMR in the presence of (R)-(-)-1-(9-anthryl)-
d
2,2,2-trifluoroethanol. Determined by comparison with the [R]D
reported in the literature.4a,14 e Reaction performed at -20 °C.
of the kinetic resolution, assuming a behavior similar to
that of methyl p-tolyl sulfide, are expected to further
increase the ee’s. The elongation of the linear chain does
not affect significantly the ee’s (substrates 4a , 4e, and
4f) meanwhile a stronger effect is observed in the
presence of aryl groups with different electronic proper-
ties (substrates 4a -d ). This is again at variance with
the behavior of the Ti(IV)/(+)-DET/TBHP reagents. There-
fore, to some extent, these Ti(IV)/trialkanolamines peroxo
complexes complement the other Ti(IV)-diethyltartrate
based chiral oxidizing reagents. The oxidations by the
latter, in fact, are almost unaffected by the substituents
on the aromatic ring, while any increase or branching of
the aliphatic chain dramatically decreases their enantio-
selectivity in comparison with aryl methyl sulfides.
Ack n ow led gm en t. Grateful thanks are due to Dr.
Cesare Cappellari, for preliminary experiments on this
work. This research was carried out in the frame of
“Progetto Strategico: Tecnologie Chimiche Innovative”
of CNR. Financial support by MURST is also gratefully
acknowledged. We also thank the INTAS Association
for 94/1515 grant.
Exp er im en ta l Section
J O960359Z
Gen er a l Meth od s. 1H NMR spectra were determined on a
Bruker WP200 (200 MHz) instrument. Specific rotations were
obtained with a Perkin-Elmer 241 polarimeter operating at λ )
(17) Bissing, D. E.; Matuszac, C. A.; McEwen, W. E. J . Am. Chem.
Soc. 1964, 86, 3824.