Fluorous biphasic hydrolytic kinetic resolution of terminal epoxides

Article abstract of DOI:10.1016/j.jfluchem.2003.07.008

Although application of light-fluorous techniques facilitates the isolation of reaction products from the hydrolytic kinetic resolution (HKR) of terminal epoxides catalysed by cobalt complexes of salen ligands, the extension of the original fluorous biphasic approach to this reaction is far from being a trivial exercise. The nature of the counter anion has a dramatic effect on the catalytic activity of heavily fluorinated chiral (salen) cobalt(III) complexes. Excellent enantioselectivities are obtained in the fluorous biphasic HKR of 1,2-hexene oxide when fluorinated anions are introduced (e.e.s up to 99% both for the diol and the epoxide), with C₈F₁₇COO^- affording reaction rates even higher than those observed with non-fluorous systems.

Full text of DOI:10.1016/j.jfluchem.2003.07.008

Journal of Fluorine Chemistry 125 (2004) 175–180
Fluorous biphasic hydrolytic kinetic resolution of terminal epoxides
*
Ian Shepperson, Marco Cavazzini, Gianluca Pozzi , Silvio Quici
CNR-Istituto di Scienze e Tecnologie Molecolari, via Golgi 19, 20133 Milano, Italy
Received 9 January 2003; received in revised form 19 June 2003; accepted 28 July 2003
Abstract
Although application of light-fluorous techniques facilitates the isolation of reaction products from the hydrolytic kinetic resolution (HKR)
of terminal epoxides catalysed by cobalt complexes of salen ligands, the extension of the original fluorous biphasic approach to this reaction is
far from being a trivial exercise. The nature of the counter anion has a dramatic effect on the catalytic activity of heavily fluorinated chiral
(
salen) cobalt(III) complexes. Excellent enantioselectivities are obtained in the fluorous biphasic HKR of 1,2-hexene oxide when fluorinated
–
anions are introduced (e.e.s up to 99% both for the diol and the epoxide), with C
observed with non-fluorous systems.
8
F17COO affording reaction rates even higher than those
#
2003 Elsevier B.V. All rights reserved.
Keywords: Biphasic catalysis; Cobalt; Diols; Epoxides; Fluorinated anions; Salen ligands
1
. Introduction
reaction products showing the usual affinity for organic
solvents [1,2]. The research field opened in 1994 quickly
expanded to include many other separation techniques
based on the phase behaviour of molecules containing at
least one highly fluorinated domain (fluorous molecules)
[3,4]. Nowadays, the expression ‘‘fluorous chemistry’’
encompasses the study of the structure, composition, prop-
erties and reactions of fluorous molecules, molecular frag-
ments, materials and reaction media [5].
Enantioselective reactions were recognised as a potential
target at a very early stage of development of fluorous
chemistry, not only in view of possible separation advan-
tages but also because of the unusual selectivities and
activities sometimes observed for catalytic reactions. These
improvements can be related to the peculiar solvation envir-
onment offered by fluorous solvents with their many unusual
chemical and physical properties. However, modifications of
the catalyst structure, which must be specifically tailored for
use in these reaction media, should be taken in due account.
The importance of the ligand design was first demonstrated
in the case of the FB asymmetric epoxidation of alkenes
catalysed by fluorous chiral (salen) manganese(III) com-
plexes [6–9]. Other fluorous chiral ligands and complexes
have been subsequently reported as well as the use of
fluorous techniques in several asymmetric reactions [10].
In addition, fluorous separation strategies have been usefully
applied to enzyme-catalysed kinetic resolutions of racemic
alcohols through enantioselective acylation or deacylation
reactions [11–14]. Lipase-catalysed (trans) esterification
The continuing drive toward ‘greener’ chemistry has led
to a growing interest in the use of catalytic processes to
enhance the activity and selectivity of organic reactions
and thus increase their efficiency. Particular attention has
been paid to homogeneous catalysis due to the high
selectivities achievable. In particular, catalytically active
chiral enantiopure organometallic complexes can be used
to accomplish asymmetric transformations with high enan-
tioselectivities. The inherent problem with homogeneous
catalysis is the separation of the catalyst, without decom-
position, from the other reaction components once the
process has gone to completion, which is important both
for facilitating the purification of the products and for
recycling of the catalyst. Indeed, contamination of pro-
ducts with transition metals (even in traces) is unaccep-
table for many compounds, such as pharmaceuticals. Also
homogeneous catalysts can be relatively difficult and
expensive to make, as in the case of chiral complexes.
This has led to the development of several catalyst immo-
bilisation and separation techniques, including fluorous
biphasic (FB) systems in which suitable homogeneous
catalysts can effectively operate in perfluorocarbons and
then be easily recovered by simple phase separation from
*
Corresponding author. Tel.: þ39-02-503-14163;
fax: þ39-02-503-14159.
E-mail address: gianluca.pozzi@istm.cnr.it (G. Pozzi).
0022-1139/$ – see front matter # 2003 Elsevier B.V. All rights reserved.
doi:10.1016/j.jfluchem.2003.07.008

176
I. Shepperson et al. / Journal of Fluorine Chemistry 125 (2004) 175–180
heavily fluorinated salen ligands with the aim of filling
this gap.
H
H
N
N
Co
R2
O
O
R2
2
. Results and discussion
R1
R1
R1
R2
Several salen ligands with a high fluorine content
F > 60%) and bearing at least four C C perfluoroalkyl
substituents (R ) have been previously synthesised in our
(
7
– 8
F
(
R,R)-1
R,R)-2
R,R)-3
R,R)-4
But
But
laboratory [6–9,19]. These heavily fluorinated chiral
ligands can be divided into two subsets, characterised
by the presence of electron-withdrawing RF or bulky
But
(
C8F17
(
C8F
17
C8F
17
0
t-butyl substituents in the 3,3 -positions, respectively. In
But
both cases, the corresponding manganese(III) complexes
showed pronounced fluorophilicity and were applied in FB
enantioselective reactions such as the epoxidation of
alkenes and the oxidation of sulfides. It was thus assumed
that cationic cobalt(III) complexes of heavily fluorina-
ted salen ligands, prepared by aerobic oxidation of the
corresponding neutral cobalt(II) species prior to use, could
have been useful catalysts to investigate the HKR of
terminal epoxides under FB conditions. In order to prove
this point, a specimen for each subset was chosen. Cobal-
t(II) complexes (R,R)-3 and (R,R)-4 (Chart 1) were thus
formed by adding Co(OAc) Á4H O dissolved in ethanol to
(
C8F17
C8F17
Chart 1
reactions in perfluorocarbon-hydrocarbon solvent mixtures
enabled direct partitioning of the resolution process products
by liquid–liquid separation [15].
Synthetic fluorous catalysts are also able to promote
kinetic resolution processes, as shown in the case of the
hydrolytic kinetic resolution (HKR) of terminal epoxides
catalysed by cobalt(III) complexes of chiral light-fluorous
salen ligands [16]. Terminal epoxides are readily acces-
sible in racemic form, and HKR of these substrates
provides a convenient access to both highly enantioen-
riched terminal epoxides and 1,2-diols [17]. The chiral
2
2
the proper fluorous chiral salen ligand dissolved in
degassed toluene ((R,R)-3) or aaa-trifluorotoluene (TFT,
(R,R)-4). In analogy with light-fluorous chiral (salen)
cobalt(II) complexes, (R,R)-3 was isolated after crystal-
lisation from EtOH in 93% yield [16]. The cobalt(III)
complexes obtained from aerobic oxidation of (R,R)-3 in
the presence of CH COOH or C F COOH were found to
(
1
salen) cobalt(II) complex (R,R)-1 (Chart 1), introduced in
997 by Jacobsen [18], is a highly efficient and enantio-
3
8 17
selective pre-catalyst for this reaction and the structure of
the chiral fluorous cobalt complexes we developed in our
previous work (e.g. (R,R)-2, Scheme 1) was kept as close
as possible to this model compound [16]. Reactions were
carried out in homogenous systems, where the epoxide
acted as the solvent, to give the diol and the epoxide in
high yields (almost 50% each) and enantioselectivities (up
to 99%) after standard or fluorous workup [16]. However,
the goal of studying the influence of fluorous reaction
media on the HKR of terminal epoxides could not be
achieved, because attempts at using (R,R)-2 and similar
light-fluorous cobalt complexes according to the classical
FB approach were foiled by the lack of sufficient fluorine,
preventing preferential partition into fluorous solvents.
We have now turned to using cobalt(III) complexes of
be insoluble in common organic solvents, except Et O
2
were they partly dissolved and soluble in perfluorocarbons
at room temperature. Unfortunately, they were also soluble
in terminal epoxides such as 1-hexene oxide and therefore
not suitable for use as catalysts under FB conditions.
More interestingly, at room temperature the cobalt(II)
complex (R,R)-4 was insoluble in all the solvents we
tested, including perfluorooctane. It was, however, soluble
in boiling TFT from which it was recrystallised in 72%
yield and again it was under these conditions that it was
converted to the catalytically active cobalt(III) species
used in HKR experiments by stirring in air in the presence
þ
À
of two molar equivalents of suitable promoters Y X .
Beside CH COOH, which is generally used to promote
the oxidation of non-fluorous (salen) cobalt(II) complexes
17,20], fluorinated acids and salts were used (vide infra).
3
[
Indeed, it was found previously that a fluorinated counter-
À
O
R
ion such as C F COO strongly enhances the solubility
17
8
of cationic fluorous (salen) manganese(III) complexes in
perfluorocarbons [8]. More recently, weakly co-ordinating
fluorous tetraaryl borate anions have been synthesised
with the final goal of increasing the solubility of cationic
transition-metal catalysts in alkanes, supercritical CO2
[21] and, possibly, perfluorocarbons [22]. The various
O
a) (R,R)-4, QX, air
+
R
b) H O (0.5 equiv), RT
2
OH
OH
R
Scheme 1.

I. Shepperson et al. / Journal of Fluorine Chemistry 125 (2004) 175–180
177
Table 1
Table 2
FB hydrolytic kinetic resolution (HKR) of 1-hexene oxide catalysed by
a
HKR of terminal epoxides catalysed by (R,R)-4 previously oxidised to the
a
(
R,R)-4 previously oxidised to the Co(III) state in the presence of YX
Co(III) state in the presence of C
8
F
17COOH
Entry YX
Time Conversion Diol e.e.
b
Epoxide e.e.
d
(%)
Entry
R
Time (h)
Diol
Epoxide
Yield
c
(h)
(%)
(%)
Yield
b
(%)
e.e.
(%)
e.e.
(%)
e
f
c
b
d
1
2
3
4
5
6
7
8
CH
CH
3
COOH
COOH
5
50
98
98
(%)
3
24
2
0
50
48
50
50
43
19
–
–
e
–
1
2
3
4
5
CH
CH
CH
3
3
3
24
2
49
49
47
42
27
92
ꢀ99
ꢀ99
97
97
C
C
8
F
17COOH
17SO
COONa
COH
ꢀ99
ꢀ99
ꢀ99
ꢀ99
ꢀ99
ꢀ99
ꢀ99
ꢀ99
ꢀ99
ꢀ99
74
(CH
(CH
2
)
)
3
5
46
48
43
35
ꢀ99
ꢀ99
ꢀ99
31
8
F
3
K
3
2
15
3
CF
(CF
NaBF
(CF SO
3
5
ClCH
2
3
)
3
5
C
6
H
5
24
91
4
5
a
3
2
)
2
NLi 24
23
Reaction conditions: racemic epoxide ¼ 25 mmol; H2O ¼ 17:5 mmol;
catalyst ¼ 0:05 mmol; perfluorooctane ¼ 3 ml; molar ratio epoxide/
catalyst ¼ 500; T ¼ 25 8C. See also Section 4.
a
Reaction conditions: racemic epoxide ¼ 25 mmol (3 ml); H2O ¼
1
7:5 mmol (0.32 ml); catalyst ¼ 0:05 mmol; perfluorooctane ¼ 3 ml;
b
Products isolated by fractional distillation. See Section 4.
molar ratio epoxide/catalyst ¼ 500; T ¼ 25 for values in 8C. See also
c
Determined by capillary GC analysis of the isolated product on a HP
Section 4.
b
ChiralÀ20% permethylated b-cyclodextrin column.
Determined by GC analysis of the reaction mixture. See Section 4.
Determined by capillary GC analysis of the isolated product on a HP
d
c
Optical purity of the isolated epoxide. See Section 4.
e
Not determined.
ChiralÀ20% permethylated b-cyclodextrin column.
d
Optical purity of the isolated epoxide. See Section 4.
e
Homogeneous conditions, see Ref. [17]: catalyst ¼ (R,R)-1; molar
ratio epoxide/catalyst ¼ 240.
where a moderate reduction of the catalytic activity
had been observed for [(salen) cobalt(III)] (CH COO)
f
Based on the e.e. of the recovered epoxide and diol. See Ref. [20].
3
compared to the corresponding [(salen) cobalt(III)](C8F17-
COO) complexes [16].
cationic cobalt(III) complexes formed, which differed in
À
Under optimised FB conditions (entry 3), the maximal
conversion level of 50% was reached in 2 h and excellent
e.e.s (up to 99%) were observed for (S)-1,2-hexane diol and
the residual (R)-1-hexene oxide. These results compare
favourably to those obtained for the HKR of 1-hexene oxide
carried out under standard homogeneous conditions, cata-
lysed by the cobalt(III) complex obtained by oxidation of
(R,R)-1 in the presence of AcOH (entry 1) [20]. In contrast to
what was generally observed in the case of FB oxidation
reactions catalysed by fluorous (salen) manganese(III)
[7,19], enhanced catalytic activities are not associated with
lower levels of enantioselectivity.
Other terminal epoxides were subjected to HKR under FB
conditions in the presence of the most active cobalt(III)
complex and results are summarised in Table 2. Reaction
products were isolated in good yields by fractional distilla-
tion of the upper organic layer, and the configurations of the
enantiomerically enriched epoxides and diols (established
by comparison with authentic samples) always corre-
sponded to those reported in Scheme 1.
The catalytic FB system based on the fluorous complex
(R,R)-4 usually gave clean reactions with no detectable
traces of by-products, affording highly selective kinetic
resolutions over reasonable reaction times even at a low
catalyst loading. Also in the case of epichlorohydrin (entry
4), the reaction proceeded quickly without the addition of
any organic co-solvent as required with light-fluorous cat-
alysts [16].
We finally addressed the recyclability of the fluorous
catalyst using 1-hexene oxide as a model substrate (Table 3).
Independently of the identity of the counterion, in the first
run partial precipitation of a brown-orange slurry from the
the nature of the counterion X , were isolated by remov-
ing the solvent under reduced pressure. Independently on
À
the identity of X , they were all soluble at room tem-
perature in TFT, perfluorooctane and partially soluble in
diethyl ether, very slightly soluble in dichloromethane
and insoluble in all other solvents, including terminal
À2
epoxides, at the concentration (1:67 Â 10 M) employed
for the HKR of the model compound 1-hexene oxide
(
Table 1).
FB reactions (Scheme 1) were carried out at room tem-
perature without addition of any organic co-solvent. A sub-
stoichiometric amount of water (0.7 equivalents with respect
to the epoxide) was added to the biphasic mixture formed by
1
-hexene oxide and the solution of the fluorous catalyst in
perfluorooctane. The progressive consumption of the epox-
ide upon vigorous stirring of the FB system under aerobic
conditions was followed by GC analysis of the upper organic
layer.
À
The nature of the counterion X determined the cata-
lytic activity of the cobalt(III) complexes tested. The
resolution process came to completion after only 2 h in
À
the case of C F COO (Table 1, entry 3), other fluori-
8
17
nated anions requiring longer reaction times to attain
either complete (entries 4–6) or partial (entries 7,8) con-
version of the S-epoxide: all these complexes gave (S)-1,2-
hexanediol as the only detectable product. It is important
to note that the epoxide ring-opening reaction did not
occur when the fluorous cobalt(III) catalyst was generated
in the presence of CH COOH (entry 2). FB conditions
3
accentuate a behaviour already observed for homogeneous
HKR reactions catalysed by light-fluorous complexes

178
I. Shepperson et al. / Journal of Fluorine Chemistry 125 (2004) 175–180
OH
fluorous layer was observed just after the diol started to form
À
O
a) (R,R)-4, YX, air
À
O
(
or for X ¼ CH COO , on addition of H O), as already
C H
C H
4 9
Ph
3
2
b) PhOH
4
9
found in homogeneous reactions [16]. After reaction com-
pletion the precipitate, a mixture of cobalt(III) and cobalt(II)
complex (R,R)-4 as shown by UV/Vis and IR spectroscopic
data, was filtered off, dissolved in boiling TFT and reox-
idised in air under the same conditions used for fresh (R,R)-
5
Scheme 2.
(entry 3). This procedure gave identical results to those
obtained upon separation of the precipitated solid by simple
filtration.
4
. The regenerated cobalt(III) complexes were dissolved in
perfluorooctane and added to racemic 1-hexene oxide. A
drop in catalytic activity was consistently observed in this
second run with all the anions tested but the enantioselec-
tivity was not affected and only (S)-1,2-hexanediol was
formed (entries 1,4–6). This and the mild conditions under
which the HKR is performed suggested that decomposition
of the fluorous chiral ligand were not primarily involved.
Cobalt leaching was also excluded as the cause of the loss of
catalytic activity because the addition of Co(OAc) Á4H O to
The loss of catalytic activity upon recycling of the fluor-
ous complex seems to be independent on the nature of
the nucleophile used in the asymmetric ring-opening
reaction, as shown by preliminary experiments carried out
with phenol (Scheme 2, Table 3, entries 7,8). Although this
latter is a less active nucleophile than water [23], (S)-1-
phenoxy-2-hexanol 5 was obtained in high e.e. and 53%
yield based on phenol in the FB biphasic reaction carried out
with fresh catalyst (entry 7). However, recycled catalyst
gave only traces of product after prolonged reaction times
(entry 8).
2
2
the recovered solid under the conditions affording (R,R)-4
from the corresponding ligand, followed by the oxidative
step, did not improve the results obtained in the second run
(
entry 2).
In a further attempt to avoid any loss of material and to
reduce the possible presence of water in the recovered cat-
alyst, the lower fluorous layer containing the solid was
evaporated to dryness under reduced pressure and the residue
dissolved in boiling TFT treated with C F COOH in air
3. Concluding remarks
We have demonstrated that the catalytic HKR of terminal
epoxides is feasible under FB conditions provided that
proper highly-fluorinated ligands are selected. Moreover,
the identity of the counterion for the (salen) cobalt(III)
complex was found to be crucial for the activity of the
catalytic system, with fluorinated anions, in particular
8
17
Table 3
a
FB kinetic resolution of 1-hexene oxide catalysed by recovered (R,R)-4
b
Entry YX
À
Time Conversion e.e. Ring
c
e.e. Epoxide
e
(%)
C F COO , giving the best results. Recycling of the
17
8
(
h)
(%)
opening
heavily fluorinated chiral catalyst proved to be less effective
than that of similar light-fluorous complexes. Further study
on this particular aspect of the FB system is currently
underway.
d
product (%)
f
1
2
3
4
5
6
7
8
C
C
C
C
8
8
8
8
F
F
F
F
17COOH
17COOH
17COOH
5
5
5
37
34
38
25
23
18
ꢀ99
ꢀ99
ꢀ99
ꢀ99
ꢀ99
ꢀ99
59
52
60
33
30
22
–
g
h
f
17SO
3
K
24
24
f
NaBF
(CF SO
4
f
)
2
NLi 24
4. Experimental
3
2
i
l
m
C
C
8
F
17COOH
17COOH
18
76
53
l
97
f
m
96
8
F
3
–
4.1. General experimental procedures
a
Reaction conditions and results obtained with the corresponding fresh
complexes: see Table 1.
b
Solvents were purified by standard methods and dried
Entries 1–6¼ hydrolytic kinetic resolution (Scheme 1); entries 7–
if necessary, except for n-perfluorooctane which was used
as received. All commercially available reagents were used
as received except for phenol which was distilled under
vacuum (100 8C, 2 Torr). Melting points (uncorrected)
were determined with a B u¨ chi SMP-20 capillary melting
point apparatus. Optical rotations were measured using a
Perkin–Elmer 241 polarimeter. FT-IR spectra were recor-
ded as KBr disks on a Perkin–Elmer 1725 X spectrophot-
ometer. UV/Vis spectra were measured using a Lambda 6
Perkin–Elmer spectrometer. GC analyses were performed
on a Hewlett–Packard 5890 instrument (column: 30 m Â
0:32 mm HP-5 5% phenyl methyl siloxane or 30 m Â
0:25 mm HP chiral-20% permethylated b-cyclodextrin
column.). HPLC analysis was performed using an Agilent
8
see Section 4).
¼ asymmetric ring-opening with phenol (Scheme 2, reaction conditions:
c
Based on the e.e. of the recovered epoxide and ring opening product.
See Ref. [20].
d
Determined by capillary GC analysis of the isolated product on a HP
ChiralÀ20% permethylated b-cyclodextrin column.
e
Optical purity of the isolated epoxide. See Section 4.
f
Recycling experiment carried out in the presence of regenerated
catalyst recovered by filtration of the fluorous layer.
g
Co(OAc)
2
2
Á4H O (0.2 mmol) was added to the catalyst recovered by
filtration of the fluorous layer before regeneration of the active species.
h
Recycling experiment carried out in the presence of regenerated
catalyst recovered by evaporation of the fluorous layer.
i
Fresh catalyst.
l
Isolated yield of 5 based on phenol.
m
Determined by HPLC analysis of isolated 5 on Chiralcel OD column.

I. Shepperson et al. / Journal of Fluorine Chemistry 125 (2004) 175–180
179
1
100 HPLC with a Chiralcel OD column. Elemental ana-
1/1). Optical purity of the isolated epoxide was deter-
mined by comparison of its optical rotation with literature
data.
lysis: Departmental Service of Microanalysis (University of
Milano, C, H, N).
0
4
heptadecafluorooctyl)phenyl]salicylidene}-1,
.2. (R,R)-(À)-N,N -Bis{3-tert-butyl-5-[3,5-bis(n-
4.4. (S)-1-phenoxy-2-hexanol 5
2
-cyclohexanediamino-cobalt(II) ((R,R)-4)
The fluorous Co(II) complex (R,R)-4 (116 mg,
.05 mmol) was oxidised with perfluorononanoic acid
0
(46.4 mg, 0.1 mmol) as before. This was dissolved in
The parent fluorous salen ligand (1.3 g, 0.58 mmol),
synthesised according to a literature procedure [9], was
dissolved in degassed trifluorotoluene (10 ml) and refluxed.
Co(OAc) Á4H O (0.21 g, 1.14 mmol) suspended in ethanol
perfluorooctane (3 ml) and (ꢁ)-1-hexene oxide (0.6 ml,
˚
5 mmol) and molecular sieves 3 A (10 mg) were added.
This mixture was stirred for 5 min, then phenol (212 mg,
2.25 mmol) was added and the reaction was stirred for 18 h.
Acetonitrile (5 ml) was added and the precipitated catalyst
was filtered off and reoxidised in air for recycling experi-
ments. The fluorous layer was separated and extracted three
times with acetonitrile. The enantioenriched epoxide and the
solvent were removed by distillation of the organic phase at
atmospheric pressure. Kugelrohr distillation of the residue
(100–103 8C, 2 Torr, then 115–120 8C, 2 Torr) provided the
title compound (230 mg, 53% based on phenol) as a colour-
less oil, in 97% e.e. by chiral HPLC (Chiralcel OD, 5%
EtOH in hexane). Physical data in agreement with those
reported in the literature [23].
2
2
(
10 ml) was added dropwise to the salen solution under
nitrogen. The reaction was stirred at reflux for 4 h and then
allowed to cool and the solvents were evaporated off. The
crude product was purified by washing with ethanol and
recrystallisation from trifluorotoluene to give the pure pro-
2
0
duct (0.98 g, 72%) as an orange solid: mp > 275 8C; ½aŠ -
D
6
1
10 (c ¼ 0:02, CCl FCF Cl); IR (KBr, selected); n
À1
2
2
À5
603.6 m and 1541.2 (w) cm ; UV/Vis (1 Â 10 M,
ðEÞ ¼ 268 (48510), 318(33170), 373
CCl FCF Cl): l
2
2
max
(19230) and 420 (15660) nm. Analysis; calculated for
C H CoF N O : C 37.3, H 1.7, N 1.2; found: C 36.1,
H 1.7, N 1.2.
7
2
40
68 2 2
4
.3. HKR of terminal epoxides: typical procedure
Acknowledgements
The fluorous Co(II) complex (R,R)-4 (116 mg,
.05 mmol) and perfluorononanoic acid (46.4 g, 0.1 mmol)
0
The financial contribution of the European Commission
through the Human Potential Programme (Contract no.
HPRN-CT-2000-00002, Development of Fluorous Phase
technology for Oxidation processes) is gratefully acknowl-
edged.
were refluxed in trifluorotoluene in a 25 ml flask with a
condenser open to the air for 2 h. After oxidation the solvent
was evaporated of and the residue was dried under vacuum.
The crude cobalt(III) complex (brown solid; IR (KBr,
selected); n 1690.7 m, 1645.9 m, 1609.2 m and 1537.8 w
À1
À5
cm ; UV/Vis (1 Â 10 M, CCl FCF Cl): l
ðEÞ ¼ 272
2
2
max
References
(
57830), 321 (53700), 402 (9060) and 420 (15660) nm) was
then dissolved in perfluorooctane (3 ml) and the racemic
epoxide (25 mmol) was added. The biphasic mixture was
stirred for 5 min and then water was added (0.32 ml,
[
1] I.T. Horv a´ th, J. R a´ bai, Science 266 (1994) 72–75.
2] I.T. Horv a´ th, Acc. Chem. Rev. 31 (1998) 641–650, and references
therein.
[
1
7.5 mmol). The reaction was stirred at room temperature
[
3] D.P. Curran, Fluorous techniques for the synthesis of organic
molecules: a unified strategy for reaction and separation, in: F. Vogtle,
J.F. Stoddart, M. Shibasaki (Eds.), Stimulating Concepts in Chemistry,
Wiley-VCH, Weinheim, 2000, pp. 25–37.
and the HKR was monitored by GC analysis of aliquots
of the organic layer taken at different times (column ¼
HP-5). Conversion was evaluated by comparison of the
relative area of the epoxide and diol peaks with a calibra-
tion curve built with mixtures of the two products in
known amounts. After the reaction had gone to completion,
[4] D.P. Curran, Synlett (2001) 1488–1496, and references therein.
[5] J.A. Gladysz, D.P. Curran, Tetrahedron 58 (2002) 3823–3825.
[6] G. Pozzi, F. Cinato, F. Montanari, S. Quici, Chem. Commun. (1998)
877–878.
[
[
[
7] G. Pozzi, M. Cavazzini, F. Cinato, F. Montanari, S. Quici, Eur. J.
Org. Chem. (1999) 1947–1955.
CH CN (3 ml) was added and the precipitated cobalt(II)
3
complex was filtered off. The organic layer was separated
and the fluorous layer was further extracted with CH CN
8] M. Cavazzini, A. Manfredi, F. Montanari, S. Quici, G. Pozzi, Chem.
Commun. (2000) 2171–2172.
3
(
3 ml). The organic layers were combined, the solvent
9] M. Cavazzini, A. Manfredi, F. Montanari, S. Quici, G. Pozzi, Eur. J.
Org. (2001) 4639–4649.
and remaining epoxide were removed under vacuum to
leave the pure diol. The epoxide if required could be isolated
by careful fractional distillation of the organic phase using
a Kugelrohr apparatus [16], except in the case of styrene
oxide and 1-phenyl-1,2-ethane diol, which were separated
by column chromatography on silica gel (AcOEt/hexane
[
[
[
10] G. Pozzi, M. Cavazzini, S. Quici, D. Maillard, D. Sinou, J. Mol.
Catal. A 182–183 (2002) 455–461 and references therein.
11] B. Hungerhoff, H. Sonnenschein, F. Theil, Angew. Chem. Int. Ed.
Engl. 40 (2001) 2492–2494.
12] B. Hungerhoff, H. Sonnenschein, F. Theil, J. Org. Chem. 67 (2002)
1781–1785.

180
I. Shepperson et al. / Journal of Fluorine Chemistry 125 (2004) 175–180
[
13] S.M. Swaleh, B. Hungerhoff, H. Sonnenschein, F. Theil, Tetrahedron
8 (2002) 4085–4089.
14] Z. Luo, S.M. Swaleh, F. Theil, D.P. Curran, Org. Lett. 4 (2002)
585–2587.
[19] M. Cavazzini, G. Pozzi, S. Quici, I. Shepperson, J. Mol. Cat. A 204–
205 (2003) 433–441.
5
[
[20] D.A. Annis, E.N. Jacobsen, J. Am. Chem. Soc. 121 (1999) 4147–
4154.
2
[
[
15] P. Beier, D. O’Hagan, Chem. Commun. (2002) 1680–1681.
[21] J. van den Broeke, M. Lutz, H. Kooijman, A.L. Spek, B.-J. Deelman,
G. van Koten, Organometallics 20 (2001) 2114–2117.
[22] J. van den Broeke, B.-J. Deelman, G. van Koten, Tetrahedron Lett.
42 (2001) 8085–8087.
16] M. Cavazzini, S. Quici, G. Pozzi, Tetrahedron 58 (2002) 3943–
3949.
[
[
17] E.N. Jacobsen, E.N. Acc. Chem. Res. 33 (2000) 421–431.
18] M. Tokunaga, J.F. Larrow, F. Kakiuchi, E.N. Jacobsen Sci. 277
[23] J.M. Ready, E.N. Jacobsen, J. Am. Chem. Soc. 121 (1999) 6086–
6087.
(1997) 936–938.