A heterogeneous cis-dihydroxylation catalyst with stable, site-isolated osmium-diolate reaction centers

Article abstract of DOI:10.1002/1521-3773(20010202)40:3<586::AID-ANIE586>3.0.CO;2-L

Involatile OsO₄! A tetrasubstituted olefin is immobilized on SiO₂ and reacts with OsO₄ to form a stable osmate (IV) ester (see scheme), which is a leak-proof heterogeneous catalyst for the cis-dihydroxylation of olefins.

Full text of DOI:10.1002/1521-3773(20010202)40:3<586::AID-ANIE586>3.0.CO;2-L

COMMUNICATIONS
(Table 1). The coordinates of Arg62(n), Arg62(l), Ala-Pro, Wa, Wb, Wc, and
Wd were fixed during occupancy refinement. Arg62(n) represents the Arg62
conformation without Ala-Pro bound and Arg62(l) represents the Arg62
A Heterogeneous cis-Dihydroxylation Catalyst
with Stable, Site-Isolated Osmium ± Diolate
Reaction Centers**
conformation with Ala-Pro bound to ceCyp3. The B-factors of Arg62(n)
,
Wa, Wb, Wc, and Wd were fixed to the same B-factor values as those of the
native structure. Individual atomic B-factors for Arg62(l) and Ala-Pro were
refined together with occupancy (Figure 4). The restraints applied in the
occupancy refinement are summarized below:
An Severeyns, Dirk E. De Vos, Lucien Fiermans,
Francis Verpoort, Piet J. Grobet, and Pierre A. Jacobs*
QArg62(n) ‡ QArg62(l) ˆ 1
QArg62(l) ˆ Ql
QArg62(n) ˆ QWa ˆ QWb ˆ QWc ˆ QWd
where
Osmium tetroxide is by far the most versatile catalyst for
cis-dihydroxylation (DH) of double bonds.[1, 2] When homo-
geneous catalysts are used, free OsO4 is always present in
some step of the catalytic cycle, and the high toxicity and
volatility of OsO4 have hitherto obstructed industrial appli-
cation. Previous attempts to immobilize OsO4 used polymers,
for example, with coordination of OsO4 on polyvinylpyri-
dine.[3, 4] However, hydrolysis of the intermediate OsVI diolate
complex requires that Os is detached from the polymeric
Lewis base,[5] and this implies an inherent liability to Os
leaching. Similarly, reports on immobilized alkaloids for
asymmetric DH mention that Os leaching necessitates Os
supplementation in subsequent runs.[6] In another attempt,
OsO4 was entrapped in polystyrene microspheres, but the
mechanism by which OsO4 is retained within the polymer is
not understood.[7] Herein we report a solid with OsVIII type
reactivity, and with a persistent bond between Os and the
support. Rigorous heterogeneity tests and reactions with
12 olefins substantiate the value of the new Os catalyst.
Our approach is rooted in the mechanism of the cis-
dihydroxylation, which comprises two stages: 1) attack of the
OsVIII cis-dioxo complex on the olefin (osmylation), 2) reox-
idation of OsVI to OsVIII and hydrolytic release of the diol. Two
points are particularly relevant. First, if the hydrolytic
conditions are not too drastic, tetrasubstituted olefins are not
converted into cis-diols.[8, 9] These olefins are smoothly
osmylated to an osmate(vi) ester, but the rate of subsequent
hydrolysis is zero (0% yield for a tetrasubstituted olefin vs.
83% for a trisubstituted olefin, ref. [8]). Second, an OsVI
monodiolate complex can be reoxidized to cis-dioxo OsVIII
without release of the diol; subsequent addition of a second
olefin results in an Os bisdiolate complex.[10] These two
properties make it possible to immobilize a catalytically active
Os compound by the addition of OsO4 to a tetrasubstituted
olefin that is covalently linked to a silica support (1a,
Scheme 1). The tetrasubstituted diolate ester (1b) which is
QArg62(n): is the occupancy of the Arg62 conformation with no Ala-Pro
binding, QArg62(l): is the occupancy of the Arg62 conformation with Ala-Pro
binding, and Ql: is the occupancy of Ala-Pro.
PPIase assay:[13] a-chymotrypsin selectively hydrolyzes the C-terminal p-
nitroanilide bond of the substrate in the trans X-Pro conformer only. This
hydrolysis releases the chromophore 4-nitroaniline, the accumulation of
which is recorded by measuring the absorbance at 400 nm as a function of
time. Substrate (N-succinyl-Ala-Ala-Pro-Phe-p-nitroanilide, Bachem AG)
was dissolved in LiCl/2,2,2-trifluoroethanol (LiCL/TFE) to give a stock
solution of 100 mm. The experiment took place at 48C. Constant temper-
ature was maintained within the cuvette by a Peltier (PTP-1) temperature
control unit. A Perkin ± Elmer UV/Vis Lambda 20 spectophotometer was
used.
Proteins: ceCyp3 solution was freshly prepared before the experiment from
frozen stock solution, at the appropriate concentration, by dilution in
buffer 50 mm 2-[4-[2hydroxyethyl)-1-piperazinyl]ethanesulfonic acid
(HEPES), 100 mm NaCl, pH 8.0 (buffer A).
a-chymotrypsin (Sigma): In a typical experiment 10 mL of 20 nm ceCyp3
was made up to 870 mL with buffer A and the appropriate volume of Ala-
Pro in a 1-mL cuvette. The cuvette was then preincubated for 30 min on ice.
Immediately before the assay, 100 mL of chymotrypsin solution
1
(50 mgmL in 10 mm HCl) was added, followed by 30 mL of a 3.7 mm
stock solution of Suc-Ala-Ala-Pro-PNA in LiCl (470 mm)/TFE. The
reaction progress was monitored by the absorbance change at 400 nm that
accompanies the hydrolysis of the amide bond and the release of
4-nitroaniline.
Received: September 8, 2000 [Z15779]
[1] B. M. Baker, K. P. Murphy, Methods Enzymol. 1998, 295, 294 ± 315.
[2] B. K. Shoichet, A. R. Leach, I. D. Kuntz, Proteins 1999, 34, 4 ± 16.
[3] G. M. Verkhivker, P. A. Rejto, D. Bouzida, S. Arthurs, A. B. Colson,
S. T. Freer, D. K. Gehlhaar, V. Larson, B. A. Luty, T. Marrone, P. W.
Rose, J. Mol. Recognit. 1999, 12, 371 ± 389.
[4] M. Schapira, M. Totrov, R. Abagyan, J. Mol. Recognit. 1999, 12, 177 ±
190.
[5] M. D. Eldridge, C. W. Murray, T. R. Auton, G. V. Paolini, R. P. Mee, J.
Comput. Aided Mol. Des. 1997, 11, 425 ± 445.
[6] H. J. Bohm, J. Comput. Aided Mol. Des. 1998, 12, 309 ± 323.
[7] P. Taylor, H. Husi, G. Kontopidis, M. D. Walkinshaw, Prog. Biophys.
Mol. Biol. 1997, 67, 155 ± 181.
[8] J. Kallen, M. D. Walkinshaw, FEBS Lett. 1992, 300, 286 ± 290.
[9] J. Kallen, V. Mikol, P. Taylor, M. D. Walkinshaw, J. Mol. Biol. 1998,
283, 435 ± 449.
[10] Y. D. Zhao, H. M. Ke, Biochemistry 1996, 35, 7362 ± 7368.
[11] J. Dornan, A. P. Page, P. Taylor, S. Y. Wu, A. D. Winter, H. Husi, M. D.
Walkinshaw, J. Biol. Chem. 1999, 274, 34877 ± 34883.
[12] D. Kern, G. Kern, G. Scherer, G. Fischer, T. Drakenberg, Biochemistry
1995, 34, 13594 ± 13602.
[13] J. L. Kofron, P. Kuzmic, V. Kishore, E. Colonbonilla, D. H. Rich,
Biochemistry 1991, 30, 6127 ± 6134.
[14] G. M. Sheldrick, SHELX97, University of Göttingen, 1997 .
[15] J. Liu, C. T. Walsh, Proc. Natl. Acad. Sci. USA 1990, 87, 4028 ± 4032.
[16] T. A. Jones, J. Y. Zou, S. W. Cowan, M. Kjeldgaard, Acta Crystallogr.
Sect. A 1991, 47, 110 ± 119.
[17] Crystallographic data for the structures reported in this paper have
been deposited with the Protein Data Bank as supplementary
publication no. PDB-1E8K (see http://www.rcsb.org/pdb/index.html).
[*] Prof. Dr. Ir. P. A. Jacobs, Ir. A. Severeyns, Prof. Dr. Ir. D. E. De Vos,
Prof. Dr. P. J. Grobet
Center for Surface Chemistry and Catalysis
Katholieke Universiteit Leuven
Kardinaal Mercierlaan 92, 3001 Heverlee (Belgium)
Fax : (‡32)16-32-1998
E-mail: pierre.jacobs@agr.kuleuven.ac.be
Prof. Dr. L. Fiermans
Laboratory for Crystallography and Study of Solid Matter
Universiteit Gent (Belgium)
Prof. Dr. F. Verpoort
Deptment of Inorganic and Physical Chemistry
Universiteit Gent (Belgium)
[**] This work was supported by the Belgian Federal Government in the
frame of an Interuniversitary Attraction Pole on Supramolecular
Catalysis. We are indebted to FWO for fellowships (A.S., D.D.V.) and
for a research grant (D.D.V., F.V., P.A.J.).
586
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001
1433-7851/01/4003-0586 $ 17.50+.50/0
Angew. Chem. Int. Ed. 2001, 40, No. 3

COMMUNICATIONS
Scheme 1. Immobilization of Os in a tertiary diolate complex, and proposed catalytic cycle for cis-dihydroxylation ([O] ˆ N-methylmorpholine N-oxide).
formed at one side of the Os atom is stable, and keeps the
catalyst fixed to the support material. The catalytic reaction
can then take place at the free coordination sites of Os.
For the preparation of a silica-anchored tetrasubstituted
olefin, silica is first functionalized with 3-aminopropyltrime-
thoxysilane (Scheme 1).[11] Next 3,4-dimethylcyclohex-3-enyl-
carbonyl chloride is added, which reacts with the grafted
amino groups to form an amide. The 3,4-dimethylcyclohex-3-
enylcarbonyl chloride is prepared by the Diels ± Alder reac-
tion of 2,3-dimethyl-1,3-butadiene and ethyl acrylate, and
conversion of the ester into the acid chloride.[12] Next, OsO4
adds to the double bond in the functionalized support 1a
(!1b). To avoid handling the poisonous OsO4, hexavalent
K2OsO2(OH)4 is used as an Os source, and oxidized in situ to
OsO4 with N-methylmorpholine N-oxide (NMO) in tert-butyl
alcohol:dichloromethane (2:1).[13] Excess OsO4 is removed
from 1b by threefold washing with the same solvent mixture.
Physicochemical observations confirm that Os is immobi-
lized in surface-linked diolate complexes. In solution chem-
istry, reaction of OsO4 with an olefin gives rise to dark brown
OsVI complexes. In the reaction of the solid support 1a with
OsO4, the solid (1b) turns dark brown, while no color
develops in solution. Diffuse reflectance spectroscopy meas-
urements of the solid show intense absorptions at wavelengths
shorter than 700 nm. More detailed information is obtained
from solid-state 13C NMR spectrometry (Figure 1). The amide
signal at dˆ177 (for 1a) confirms the successful attachment of
the acyl group to the surface. The spectrum of the Os-free
1b confirms that Os is bound by esters of tertiary diols. The
immobilized Os is easily observed with X-Ray photoelectron
spectroscopy (XPS). Os 4f7/2 lines at 53.6 eV and 51.2 eV
demonstrate that osmium is in the ‡vi and ‡iv state. Based
on reported values, this is clear evidence that the octavalent
OsO4 is reduced in the reaction with the covalently linked
double bond.[14]
Catalysis with the new solid Os materials reveals that the
activity strongly depends on the concentration of the Os ester
groups on the surface: the highest activity is observed for the
lowest concentration of Os-binding groups! (Table 1, en-
tries 1 ± 3). The Os loading is easily controlled by performing
the surface functionalization with mixtures of silylating agents
ˆ
material (1a) shows a C C signal at d ˆ 124, characteristic for
the tetrasubstituted olefin. When OsO4 is added (!1b), this
signal is replaced by a new signal at dˆ92. The tertiary alcohol
groups in Os(vi) diolate complexes (prepared from OsO4 and
3,4-dimethylcyclohex-3-enylcarboxylic acid) have a resonance
signal in solution NMR spectroscopy between d ˆ 90 and 95.
Thus, the intense peak at d ˆ 92 in the solid-state spectrum of
Figure 1. Solid state 13C MAS NMR spectra of 1a (top), with immobilized
tetrasubstituted olefin, and 1b (bottom), that is, 1a after addition of OsO4
to the double bond (high power proton decoupling; 458 pulses with 10 s
recycle time; spinning rate 10 kHz).
Angew. Chem. Int. Ed. 2001, 40, No. 3
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001
1433-7851/01/4003-0587 $ 17.50+.50/0
587

COMMUNICATIONS
Table 1. Heterogeneous cis-dihydroxylation of 1-hexene: effect of the
dilution of active sites with propyltrimethoxysilane (PrTMS);[a] comparison
with homogeneous dihydroxylation with or without 2,3,4-trimethyl-2-
pentene.
Table 2. Cis-dihydroxylation of olefins with NMO and the heterogeneous
3b catalyst.[a]
Entry Olefin
t [h]
Con-
Selecti-
version [%] vity [%]
Entry
Catalyst
PrTMS:NH2PrTMS
t [h]
Conversion [%]
1
1-pentene
1-hexene
1-heptene
cyclopentene
cyclohexene
cis-2-hexene
trans-2-hexene
styrene
48
48
48
48
48
48
48
48
48
48
150
24
48
10
83
99
96
83
99
99
98
99
72
50
99
65
85
21
21
99
98
99
98
99
99
98
96
99
99
99
99
99
1
2
3
4
5
1b
2b
3b
0:1
1:1
9:1
±
24
24
24
10
10
1
5
66
98
0
2[b]
3
4
5
6
7
8
9
10
[b]
OsO4
[c]
OsO4
±
[a] 100 mg supported catalyst, 1-hexene (1.6 mmol), NMO (1.6 mmol),
solvent (3 mL), and RT unless otherwise stated. [b] NMO:1-hexene:Os ˆ
400:400:1 [c] NMO:1-hexene:2,3,4-trimethyl-2-pentene:Os ˆ 400:400:10:
1; 1-hexene is added 4 h after the other reagents.
indene
2-methyl-2-pentene
11
12
13[c]
ethyl trans-cinnamate
ethyl trans-crotonate
1-hexene
(Scheme 1). The highest activity is obtained after silylation
with a 9:1 mixture of propyltrimethoxysilane (PrTMS) and
3-aminopropyltrimethoxysilane (NH2PrTMS). This results in
a material (3b) with the Os centers diluted by propyl groups
(Si:Os ˆ 270, as determined by XPS). In contrast, the material
is practically inactive when the surface coverage with Os-
binding tetrasubstituted alkene groups is raised (for 1b:
Si:Os ˆ 36, XPS). The deactivation of Os by a high surface
concentration of Os-binding alkene groups is explained by
double-sided binding of Os by two adjacent tetrasubstituted
alkenes. This situation is easily mimicked in homogeneous
catalytic experiments, by adding 10 equivalents of the tetra-
substituted 2,3,4-trimethyl-2-pentene per Os to the catalytic
dihydroxylation of 1-hexene with NMO (Table 1, entries 4
and 5). Product formation from 1-hexene is fully blocked by a
small amount of the tetrasubstituted alkene, since the latter
forms bisdiolate complexes that are not hydrolyzed in the
mild conditions of our experiments. A similar situation arises
in the heterogeneous catalyst if the Os-binding tetrasubsti-
tuted alkene groups are too close to each other. In contrast,
appropriate site isolation, as in 3b, ensures that a nonhydro-
lyzable ester is only formed at one side of the Os atom; at the
other side, an olefin such as 1-hexene is dihydroxylated at a
high rate.
Stringent heterogeneity tests were performed with the 3b
catalyst, by splitting the reaction suspension in the dihydrox-
ylation of 1-hexene at a conversion of 21%, and monitoring
the conversion in the suspension and in the clear supernatant.
Zero activity was found in the supernatant, while the reaction
continues in the suspension (21% in the clear solution vs.
60% in the suspension; Table 2, entry 13). This test was
successfully performed for the reaction of 3b with several
olefins, however, it fails for OsO4 bound on polyvinylpyridine;
this is because of the dissociation of the coordinate bond
between Os and the nitrogen base in the hydrolytic release of
the diol.
20, filtrate
20, suspension 60
[a] reaction conditions: 100 mg heterogeneous catalyst (4 Â 10 6 mol Os),
olefin (1.6 mmol), NMO (1.6 mmol), solvent (3 mL), H2O (200 mL), RT.
[b] A second run was performed with the used catalyst from entry 2. After
48 h, conversion and selectivity were again 99 and 98% respectively.
[c] Filtrate test: the reaction mixture is splitted after 10 h; further
conversion in filtrate and suspension is determined 20 h later.
ketol make up less than 1% of the products. Aromatic olefins
such as styrene and indene are suitable substrates as well
(entries 8 ± 9). The reaction proceeds more slowly with a
trisubstituted olefin. This is not unexpected, since in our mild
conditions, the hydrolysis of the trisubstituted diolate is slow
because of steric hindrance. Note that an even greater steric
hindrance is at the basis of the stable association between Os
and the surface-bound tetrasubstituted diolate. Nevertheless,
after a somewhat increased reaction time, 98% yield is
obtained from the trisubstituted 2-methyl-2-pentene. While it
is known that more electron-rich double bonds are osmylated
at a higher rate,[16] the reactivity of OsVIII dioxo species is
sufficient to even react with relatively deactivated double
bonds. Thus, unsaturated esters such as cinnamates or
crotonates are dihydroxylated in high yields (entries 11 ± 12).
As is highlighted by the heterogeneity tests, our Os-
immobilization concept, and the formation of a stable
tetrasubstituted diolate complex, is to date the only solution
to the problem of producing such Os supported heteroge-
neous catalysts. The concept can be expanded to other catalyst
carriers that form nonhydrolyzable bonds with Os in the
conditions of the catalytic dihydroxylation. Site isolation has
been demonstrated to be crucial to obtain an active and truly
heterogeneous Os catalyst.
Experimental Section
With 3b, we succeeded in oxidizing olefins to the corre-
sponding cis-diols with an excellent conversion and selectivity
(Table 2). Monosubstituted, cis and trans disubstituted ali-
phatic olefins and cyclic olefins are stereoselectively con-
verted to cis diols with good conversions and selectivities over
98% (entries 1 ± 7). Note that the excellent chemoselectivity
of the homogeneous reactions with NMO is preserved in the
heterogeneous system;[15] overoxidation products such as the
The support material was commercial SiO2 60 from Fluka, predried at
1508C. Surface functionalization was performed by standard reported
techniques.[11] For the functionalization of the support with Os, a solution
containing OsO4 (4 Â 10 3 mmol) was treated for 4 h with the support
(100 mg, containing 10 2 mmol of double bonds in the case of catalyst 3b).
After all the OsO4 had reacted with the support, the catalyst was
thoroughly washed with tBuOH:CH2Cl2 (2:1) to remove traces of unbound
OsO4.
588
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001
1433-7851/01/4003-0588 $ 17.50+.50/0
Angew. Chem. Int. Ed. 2001, 40, No. 3

COMMUNICATIONS
In a typical catalytic dihydroxylation, the 3b material (100 mg) was added
to a mixture of the olefin (1.6 mmol), NMO (1.6 mmol), and water (200 mL)
in tBuOH:CH2Cl2 (3 mL; 2:1) solvent. The mixture was stirred at room
temperature and regularly analyzed by GC.
importance in the drugs industry, direct asymmetric synthesis
of trifluoromethylated amines is a challenge. Pirkle et al.[4]
and Mosher and Wang[5] prepared 2,2,2-trifluoro-1-phenyl-
ethylamine, and Soloshonok and Ono[6] recently reported an
elegant method for the preparation of perfluorinated amines
by a novel [1,3]-proton shift reaction. However, all of these
methods require fluorinated ketones. Nucleophilic transfer of
ªCF3º to nitrones and imines for direct preparation of
trifluoromethylated amines was recently achieved by Nelson
et al.[7] and Blazejewski et al.,[8] respectively. These methods
suffer from low yield and lack generality. We now report the
first stereoselective synthesis of trifluoromethylated amines
by using TMSCF3 2 (TMS ˆ SiMe3).
Received: August 18, 2000 [Z15661]
[1] a) M. Schröder, Chem. Rev. 1980, 80, 187; b) H. C. Kolb, M. S.
Van Nieuwenhze, K. B. Sharpless, Chem. Rev. 1994, 94, 2483; c) K. A.
Hofmann, Chem. Ber. 1912, 45, 3329; d) N. A. Milas, S. Sussman, J.
Chem. Soc. 1936, 58, 1302.
[2] C. Döbler, G. Mehltretter, M. Beller, Angew. Chem. 1999, 111, 3211;
Angew. Chem. Int. Ed. 1999, 38, 3026.
[3] G. Cainelli, M. Contento, F. Manescalchi, L. Plessi, Synthesis 1989, 45.
[4] W. A. Herrmann, R. M. Kratzer, J. Blümel, H. B. Friedrich, R. W.
Fischer, D. C. Apperley, J. Mink, O. Berkesi, J. Mol. Catal. A 1997, 120,
197.
Our systematic investigation began as an extension of our
earlier work,[9] based on the fact that imines are less electro-
philic than carbonyl compounds, and that O Si bonds are
stronger than N Si bonds. We predicted that strongly electro-
philic imines might be a solution to this problem under
noncatalytic conditions. When N-sulfonylaldimines[10] were
used as imine sources the reaction indeed proceeded smoothly
in the presence of CsF and gave only the trifluoromethylated
adducts in 45 ± 95% yield. Next we turned our attention to
sulfinylimines 1 to make this reaction stereoselective. When
chiral sulfinylimines[11] were subjected to similar reaction
conditions little or no products were obtained. Sulfinylimines
were recovered intact, but TMSCF3 decomposed. We sur-
mised that sulfinylimines are not reactive enough to add
TMSCF3. Use of different aprotic solvents and excess of
reagents was not helpful. When an excess of TMSCF3 was
used, a number of unidentified fluorinated products with little
or none of the expected adduct were obtained. TMS-
Imidazole,[8] was recently reported to facilitate addition of
TMSCF3 to imines. In our case, however, it was ineffective. In
all experiments the starting material was recovered.
Â
[5] E. N. Jacobsen, I. Marko, W. S. Mungall, G. Schröder, K. B. Sharpless,
J. Am. Chem. Soc. 1988, 110, 1968.
[6] C. Bolm, A. Gerlach, Eur. J. Org. Chem. 1998, 21.
[7] S. Nagayama, M. Endo, S. Kobayashi, J. Org. Chem. 1998, 63, 6094.
[8] K. Akashi, R. E. Palermo, K. B. Sharpless, J. Org. Chem. 1978, 43,
2063.
[9] P. G. Andersson, K. B. Sharpless, J. Am. Chem. Soc. 1993, 115, 7047.
Â
[10] J. S. M. Wai, I. Marko, J. S. Svendsen, M. G. Finn, E. N. Jacobsen, K. B.
Sharpless, J. Am. Chem. Soc. 1989, 111, 1123.
[11] J. H. Clark, D. J. Macquarrie, Chem. Commun. 1998, 853.
[12] J. Monnin, Helv. Chim. Acta 1958, 225, 2112.
[13] W. D. Lloyd, B. J. Navarette, M. F. Shaw, Synthesis 1972, 610.
[14] J. F. Moulder, W. F. Stickle, P. E. Sobol, K. D. Bomden, Handbook of
X-ray Photoelectron Spectroscopy, Perkin-Elmer Corp., 1992.
[15] V. Van Rheenen, R. C. Kelly, D. Y. Cha, Tetrahedron Lett. 1976, 1973.
[16] K. B. Sharpless, D. R. Williams, Tetrahedron Lett. 1975, 35, 3045.
The above results indicate that TMSCF3 decomposes prior
to reacting with the starting material. Hence, we thought that
increasing the substrate concentration might be a solution to
this problem. Indeed, when neat TMSCF3 was added to a
concentrated solution of the imines, the desired adduct was
obtained. The mass balance corresponds to recovered starting
material. Attempts to complete the reaction by using excess of
reagent in different solvents was, however, unsuccessful.
Imines were treated with TMSCF3 in the presence of a
stoichiometric amount of CsF to give the corresponding
trifluoromethylated sulfinamides in 50 ± 65% yields of iso-
lated products (Table 1, entries 1 ± 7, values in parentheses).
Imines with acidic a-hydrogen atoms gave lower yields
because of competitive deprotonation. The diastereoselectiv-
ity was not very high.
Stereoselective Nucleophilic
Trifluoromethylation of N-(tert-Butylsulfinyl)-
imines by Using Trimethyl(trifluoromethyl)-
silane**
G. K. Surya Prakash,* Mihirbaran Mandal, and
George A. Olah*
Trifluoromethylated amines are important building blocks
for pharmaceutical research.[1] The CF3 group, because of its
strongly electron withdrawing nature, lowers the basicity of
the amide bond towards nonspecific proteolysis[2] when these
amines are incorporated into peptides, as well as modify the
solubility and desolvation properties.[3] In spite of its prime
During these investigations we thus encountered two
problems: a) Conversion of imines was incomplete even in
the presence of excess TMSCF3 and CsF; b) imines with an a-
hydrogen atom failed to react with TMSCF3 because of the
basic nature of CsF. However, we overcame these problems
by employing a nonmetallic fluoride source. DeShong et al.
reported that tetrabutylammonium difluorotriphenylsilicate
(TBAT),[12] a soluble fluoride source, is very effective for
nucleophilic displacement reactions. We found that TBAT is
also effective in our system. Reaction of N-sulfonylaldimines
[*] Prof. Dr. G. K. S. Prakash, Prof. Dr. G. A. Olah, M. Mandal
Loker Hydrocarbon Institute and Department of Chemistry
University of Southern California
University Park, Los Angeles, CA 90089-1661 (USA)
Fax : (‡1)213-740-6270
E-mail: gprakash@usc.edu
[**] Support of our work by Loker Hydrocarbon Research Institute is
gratefully acknowledged.
Supporting information for this article is available on the WWW under
http://www.angewandte.com or from the author.
Angew. Chem. Int. Ed. 2001, 40, No. 3
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2001
1433-7851/01/4003-0589 $ 17.50+.50/0
589