Effect of α-fluorination on asymmetric epoxidation of trans-olefins using α-fluorinated cyclohexanone dioxiranes

Article abstract of DOI:10.1016/j.jfluchem.2004.04.007

Epoxidations of trans-β-methylstyrene, trans-stilbene and trans-methyl p-methoxycinnamate using chiral dioxiranes derived from both enantiopure diastereomers of α-fluoro cyclohexanones, (2 S, 5 R )-3a-6a and (2 R,5 R)-3e-6e are studied and compared. From ab initio calculations at the HF/6-31G* level of conformational inter-conversion for (2 S,5 R)-D5a and (2 R,5 R)-D5e dioxiranes it was found that, due to the α-fluorine atom, conformer K1 is more stable in the case of (2 S, 5 R)-D5a while conformer K2 is more stable in the case of (2 R, 5 R)-D5e. However, in both cases, the more stable conformers, K1 and K2, undergo rapid inter-conversion. Therefore, based on slow epoxidation reactions and rapid ring inversion of six-membered ring dioxiranes the Curtin-Hammett principle holds. Conformation K2 with axial fluorine having been found to be more reactive, the inversion of configuration observed for the epoxides obtained with ketones 3e-6e (compared with ketones 3a-6a) could be rationalized from competitive reactions of K2 and K1 conformations leading to simultaneous production of both (-) and (+) epoxides in the case of ketones 3e-6e.

Full text of DOI:10.1016/j.jfluchem.2004.04.007

Journal of Fluorine Chemistry 125 (2004) 1371–1377
Effect of a-fluorination on asymmetric epoxidation of trans-olefins
using a-fluorinated cyclohexanone dioxiranes
a,*
Arlette Solladie-Cavallo , Loıc Jierry , Hassan Norouzi-Arasi , Daryush Tahmassebi
a
a,b
c
´
Laboratoire de Stereochimie Organometallique associe au CNRS, ECPM/Universite L. Pasteur, 25 rue Becquerel, 67087 Strasbourg, France
¨
a
´ ´
´
´
´
bDepartment of Chemistry, Tehran North Campus, Islamic Azad University, Tehran, Iran
cDepartment of Chemistry, Brandon University, Brandon, Man., Canada R7A 6A9
Received 18 February 2004; received in revised form 19 April 2004; accepted 20 April 2004
Available online 5 June 2004
Abstract
Epoxidations of trans-b-methylstyrene, trans-stilbene and trans-methyl p-methoxycinnamate using chiral dioxiranes derived from both
enantiopure diastereomers of a-fluoro cyclohexanones, (2S, 5R)-3a–6a and (2R, 5R)-3e–6e are studied and compared. From ab initio
calculations at the HF/6–31Gꢀ level of conformational inter-conversion for (2S, 5R)-D5a and (2R, 5R)-D5e dioxiranes it was found that, due to
the a-fluorine atom, conformer K1 is more stable in the case of (2S, 5R)-D5a while conformer K2 is more stable in the case of (2R, 5R)-D5e.
However, in both cases, the more stable conformers, K1 and K2, undergo rapid inter-conversion. Therefore, based on slow epoxidation
reactions and rapid ring inversion of six-membered ring dioxiranes the Curtin–Hammett principle holds. Conformation K2 with axial fluorine
having been found to be more reactive, the inversion of configuration observed for the epoxides obtained with ketones 3e–6e (compared with
ketones 3a–6a) could be rationalized from competitive reactions of K2 and K1 conformations leading to simultaneous production of both (ꢁ)
and (þ) epoxides in the case of ketones 3e–6e.
# 2004 Elsevier B.V. All rights reserved.
Keywords: a-Fluorinated cyclohexanones; Fluorine effect; a-Fluorinated dioxiranes; Epoxidation of trans-olefins
1. Introduction
tion of trans-methyl p-methoxycinnamate and a comparison
of the enantioselectivities1 (ee% and configurations)
observed during epoxidations of the three trans-olefins using
these ketones of known configuration which differ only from
the orientation of the a-fluorine atom, either equatorial (3e–
6e) or axial (3a–6a).
Asymmetric epoxidation of trans-olefins by oxone cata-
lysed with chiral ketones [1] which proceed through in situ
formation of the corresponding chiral dioxiranes, has been
significantly improved since the first attempts by Curci in
1984 [2,3]. A significant activating effect by a-fluorine
substitution has been observed [2–5] and then extensively
studied [6–15]. We have recently shown that, during epox-
idation of methyl p-methoxycinnamate (Scheme 1), enan-
tiopure tri-substituted cyclohexanones 1–3 were more
efficient, leading to higher ee% and higher yields, when
the halogen was axial (1a–3a) than when the halogen was
equatorial (1e–3e) [7].
2. Experimental
2.1. General
Chemical shifts of 1H and 13C NMR are given in ‘ppm’ (d)
downfield from TMS as an internal standard. TLC was
performed on Merck glass plates with silica gel 60 F254.
Silica gel for column chromatography (Merck) was used for
the chromatographic purifications. (þ)-Dihydrocarvone,
We present here results showing that this trend is encoun-
tered also using a-fluorinated ketones 3–6 during epoxida-
tion of other olefins (such as trans-b-methylstyrene and
trans-stilbene) as well as using ketones 4–6 during epoxida-
1 Being influenced by decomposition of the ketone (through Baeyer–
Villiger), availability/solubility and rate of stirring, the percentages of
conversion do not constitute a satisfying measure of the efficiency of a
ketone and, as a consequence, of fluorine effects.
* Corresponding author. Tel.: þ33-3-90-24-27-72;
fax: þ33-3-90-24-27-06.
´
E-mail address: cavallo@chimie.u-strasbg.fr (A. Solladie-Cavallo).
0022-1139/$ – see front matter # 2004 Elsevier B.V. All rights reserved.
doi:10.1016/j.jfluchem.2004.04.007

´
A. Solladie-Cavallo et al. / Journal of Fluorine Chemistry 125 (2004) 1371–1377
1372
F
F
Cl
Cl
Cl
Cl
O
O
O
O
S
S
R
R
R
R
O
O
R
R
S
R
R
R
Cl equatorial Cl axial
Cl equatorial Cl axial
F equatorial F axial
conv. 14%
ee 2% (-)
conv. 24% conv. 20%
conv. 73% conv. 43%
ee 22% (-) ee 6% (+)
conv. 99%
ee 40% (-)
ee 2% (-)
ee 5% (-)
1e
1a
2e
2a
3e
3a
Scheme 1. Results of epoxidation of trans-methyl p-methoxy cinnamate [7].
purchased from Fluka, was a 77% (2R, 5R) and 20% (2S, 5R)
mixture, having the enantiomeric purity of the starting (ꢁ)-
carvone, 99% R at C5. All commercially available reagents
were used without further purification.
J ¼ 4.5 Hz, 1H, 6aII, 43%), 1.79 (m, 2H, 6aI and 6aII
overlapped), 1.63 (qd, J ¼ J ¼ J ¼ 12.5 Hz, J ¼ 3.5 Hz, 1H,
6aII, 43%), 1.54 (qd, J ¼ J ¼ J ¼ 13.0 Hz, J ¼ 4.0 Hz, 1H,
6aI, 57%), 1.29 (d, J ¼ 22.5 Hz, 3H, 6aI, 57%), 1.54 (d, J ¼
22.0 Hz, 3H, 6aII, 43%), 1.26 (m, 2H, 6aII, 43%), 1.17 (m,
2H, 6aI, 57%), 1.00 (m, 1H, 6aII, 43%), 0.90 (m, 1H, 6aI,
57%), 0.86 (s, 6H, 6aI and 6aII overlapped); 13C NMR
(100 MHz C6D6) signals of both diastereomers not assigned
d 205.3 (d, J ¼ 25 Hz), 205.1 (d, J ¼ 25 Hz), 95.6 (d, J ¼
172 Hz), 95.5 (d, J ¼ 173 Hz), 57.6, 57.5, 52.2, 52.5, 45.9,
45.5, 40.8 (d, J ¼ 36 Hz), 40.7 (d, J ¼ 36 Hz), 37.9 (d, J ¼
6 Hz), 37.8 (d, J ¼ 6 Hz), 22.8 (d, J ¼ 2 Hz), 22.7 (d, J ¼
2 Hz), 20.3 (d, J ¼ 24 Hz), 20.1 (d, J ¼ 24 Hz), 18.1, 17.6.
MS (m/z %) 178 (2), 153 (4), 130 (4), 109 (19), 99 (2), 75
(19), 48 (100). Anal. calculated for C10H15FO2 C, 64.5; H,
8.1. Found C, 64.4; H, 8.2.
2.2. Computational details
Semi-empirical calculations were carried out using PM3
method with the MOPAC 6.0 program package [26a,b]. The
geometries of the transition states for conformational inter-
conversion of the equilibrium structures were obtained using
the optimized geometries of the equilibrium structures
according to procedure of Dewar et al. [26c] The PM3
results were then used as input for the ab initio molecular
orbital calculations, which were carried out using the Gaus-
sian [25]. Geometries for all structures were fully optimized
by means of Berny analytical gradients optimization routine
[27a,b]. The restricted Hartree–Fock calculations with the
split-valence 6–31Gꢀ basis set which include a set of d-type
polarization functions on all non-hydrogen atoms were used
in these calculations [27c]. Vibrational frequencies were
calculated all minimum energies and transition states, which
were confirmed to have zero and one imaginary frequency,
respectively. The frequencies were scaled by a factor of 0.91
[27d] and used to compute the zero-point vibrational ener-
gies.
2.4. (2R,5R,7S/R)-2-fluoro-2-methyl-5-(2-methyl-
oxiranyl)-cyclohexanone 6e
Synthesized from 13e following identical procedure as for
6a and has been obtained in 95% yield as a 55/45 mixture of
.
diastereomer 6eI and 6eII; IR: 1732 cmꢁ1 1H NMR
(400 MHz C6D6) d 2.25 (2.dm, J ¼ 12.5 Hz, 1H, 6eI þ
6eII), 2.12 (d, J ¼ 5 Hz, 1H, 6eI 55%), 2.06 (d, J ¼ 5 Hz,
1H, 6eII 45%), 2.0 (d, J ¼ 5 Hz, 1H, 6eII 45%), 1.95 (d, J ¼
5 Hz, 1H, 6eI 55%), 1.90–1.50 (m, 3H, 6eI þ 6eII), 1.45–
1.27 (m, 3H, 6eI þ 6eII), 1.15 (d, J ¼ 22 Hz, 3H, 6eI 55%),
1.08 (d, J ¼ 22 Hz, 3H, 6eII 45%), 0.83 (s, 3H, 6eI 55%),
0.81 (s, 3H, 6eII 45%). 13C NMR (100 MHz C6D6) signals
of both diastereomers not assigned d 204.5 (d, J ¼ 17 Hz),
204.1 (d, J ¼ 17 Hz), 96 (d, J ¼ 185 Hz), 95.9 (d, J ¼
182 Hz), 57.5, 52.2, 51.3, 47.7, 43.8, 42.6, 41.2, 37.4 (d, J ¼
22 Hz), 37.0 (d, J ¼ 22 Hz), 24.6, 24.5, 22.1 (d, J ¼ 22 Hz),
21.9 (d, J ¼ 22 Hz), 19.2, 18.4. Anal. calculated for
C10H15FO2 C, 64.50; H, 8.12. Found C, 64.38; H, 8.21.
In both cases diastereomers I and II could not be separated
and mixtures were used for the epoxidation reactions.
For description of ketones 3–5 and 13 cf. ref. [16a,b].
2.3. (2S,5R,7S/R)-2-fluoro-2-methyl-5-(2-methyl-oxiranyl)-
cyclohexanone 6a
To a solution of 13a (144 mg, 0.847 mmol) in acetone
(2 mL), under stirring, at 0 8C, 0.08 M dimethyldioxirane
solution in acetone (20 mL, 16 mmol) was added. Upon
stirring to completion, as judged by TLC (about 1 h 30 min),
the reaction mixture was concentrated. Water was added and
the crude extracted with CH2Cl2. The organic phase was
dried over Na2SO4, filtered and concentrated, to give 6a as a
colorless oil (149 mg, 95%); 57/43 mixture of diastereomer
6aI and 6aII; IR: 1727 cmꢁ1. 1H NMR (400 MHz C6D6) d
2.57 (td, J ¼ J ¼ 12.5 Hz, J ¼ 6.0 Hz, 1H, 6aI, 57%), 2.51
(td, J ¼ J ¼ 13 Hz, J ¼ 6.0 Hz, 1H, 6aII, 43%), 2.34 (bd, J ¼
12.5 Hz, 1H, 6aI, 57%), 2.20 (bd, J ¼ 13.0 Hz, 1H, 6aII,
43%), 2.11 (d, J ¼ 5.0 Hz, 1H, 6aI, 57%), 2.07 (d ¼ 4.5 Hz,
1H, 6aII, 43%), 2.02 (d, J ¼ 5.0 Hz, 1H, 6aI, 57%), 2.01 (d,
3. Results and discussion
Ketones 3–6 have been synthesized in three steps from
(R)-(þ)-dihydrocarvone, Scheme 2, and the diastereomers
separated by chromatography [16]. Some isomerisation

´
A. Solladie-Cavallo et al. / Journal of Fluorine Chemistry 125 (2004) 1371–1377
1373
F
F
O
O
5e 95%
(+5% 5a)
5
5
6eI/6eII
95%
F
*
O
F
iii
O
5a 90%
(+10% 5e)
5
iii
iv
F
F
F
O
O
F
5
O
i; ii
+
5
F
F
iv
6aI/6aII
O
13a
O
13e
49%
5
O
95%
+
5
5
51%
5
*
vi; i; ii
O
F
F
3a
3e
v
O
O
O
+
5
5
5
i; ii
Cl
Cl
Cl
4e : 57%
4a : 43%
Scheme 2. (i) TMSCI, NaI, NEt3 in pentane/MeCN (ii) selectfluor in DMF (iii) HF-pyridine (30% pyr./70% HF) in THF, 0 8C, in a polystyrene flask (iv)
dimethyl dioxirane/acetone (v) HCl gas in Et2O, 0 8C and (vi) Pd/C, H2 (1 bar) in EtOH, rt.
occurs using HF-pyridine, therefore even starting from
diastereomerically pure 13a and/or 13e, 5a and 5e must
be purified.
gel, the ee’s were determined by 1H NMR using Eu(hfc)3 in
CDCl3 and, all three trans-epoxides being known, the
absolute configurations were determined using the sign of
the optical rotations ([a]D measured under identical condi-
tions: same solvent, same concentration). None of the
ketones studied here underwent Baeyer–Villiger oxidation.
Used in sub-stoichiometric amounts (0.3 equiv.) they were
recovered almost quantitatively after reaction (through chro-
matography) and no lactones were detected by NMR
(400 MHz) of crude products in none of the reactions.
The results of the asymmetric epoxidation of trans-b-
methylstyrene 7, trans-stilbene 8 and trans-methyl p-meth-
oxy cinnamate 9 (Scheme 3), with ketones 3e–6e (having the
fluorine atom equatorial in C1) are gathered in Table 1.
Epoxidation of trans-b-methylstyrene 7 with the diastereo-
meric ketones 3a–6a (having the fluorine atom axial in C1),
epoxidation of trans-methyl p-methoxy cinnamate 9 with
ketones 4a and 6a as well as epoxidation of trans-stilbene 8
with ketone 6a are gathered in Table 2.
The 1D and 2D NMR data [17] (in benzene-d6) of ketones
3e–6e fit with the (2R, 5R)-configuration and conformation
C1 having both the fluorine atom and the iso-propyl-type
group equatorial as the most populated, while in the case of
ketones 3a–6a the data fit with the (2S, 5R)-configuration
and conformation C1 having the methyl and the iso-propyl-
type groups equatorial but the fluorine axial as the most
populated (Fig. 1). For convenience the ketones will thus be
called F-equatorial and/or F-axial (instead of (2R, 5R) and/or
(2S, 5R), respectively).
The epoxidation reactions have been conducted as pre-
viously described [7,9] either in DME/H2O or in dioxane/
H2O as solvent. The percentage conversions were deter-
1
mined by a combination of masses and H NMR spectro-
scopy (400 MHz) of the solvent-free crude products. The
epoxides were isolated by flash chromatography on silica
The level of enantioselectivity obtained during epoxida-
tion of trans-olefins 7–9 with ketones 3e–6e (having iden-
tical absolute configurations, Fig. 1) is very low to nil
(Table 1, line 3 and note a).
F
F
F
F
O
O
O
O
Me
5e
Me
4e
Me
3e
Me
6e
Inversions of the enantioselectivity, compared with epox-
idations using the ketones with an axial fluorine, are
O
F
Cl
H
(+)-(2R,5R)
(+)-(2R,5R)
(+)-(2R,5R)
(2R,5R,7R/S)
R1
R
R
Me
Me
Me
Me
O
O
O
O
O
H
2
R1
H
2
R1
3
+
ketone (0.3equiv.)
oxone
3
H
H
O
R
F
F
F
F
O
F
Cl
H
7: R = H, R1 = Me
8: R = H, R1 = Ph
10: (S,S)-(-)
4a
5a
3a
6a
(2S,5R,7R/S)
(R,R)-(+)
(R,R)-(+)
(2S,3R)-(+)
(-)-(2S,5R)
(-)-(2S,5R)
(S,S)-(-)
(2R,3S)-(-)
(+)-(2S,5R)
11:
9: R = OMe, R1 = CO2Me
12:
Fig. 1. Absolute configuration and major conformation C1 (determined by
1H NMR in benzene-d6) of ketones 3e–6e and 3a–6a.
Scheme 3. Epoxidations and absolute configuration of epoxides.

´
A. Solladie-Cavallo et al. / Journal of Fluorine Chemistry 125 (2004) 1371–1377
1374
Table 1
Epoxidation of trans-olefins 7–9 using ketones (2R, 5R)-3e–6e having an equatorial fluorine in the most populated conformer C1
b-Methylstyrene 7
Stilbene 8
p-Methoxy cinnamate 9a
Ketone
3e
4e
85
20
(ꢁ)
SS
A
5e
43
0
6e
25
2
3e
4e
22
23
(ꢁ)
SS
B
5e
6e
25
2
4e
5
5e
6e
20
8
Percentage conversion
Percentage ee
Major enantiopure
Configuration
Conditionsc
100
26
69
18
(þ)
RR
B
70
20
(þ)
RR
B
17
10
(þ)
SR
A
ndb
(þ)
RR
A
–
(þ)
RR
A
(þ)
RR
B
–
(þ)
SR
B
–
–
A
A
a Ketone 3e: 43% yield, 6% ee (þ)-(SR), condition B; cf. ref. [7].
b nd: not determined.
c A: DME/H2O, 0 8C, 1.4 equiv. oxone, 6 h; B: dioxane/H2O, rt, 3 equiv. oxone, 8 h.
Table 2
Epoxidation of trans-olefins 7–9 using ketones (2S, 5R)-3a–6a having an axial fluorine in the most populated conformer C1
b-Methylstyrene 7
Stilbene 8a
p-Methoxy cinnamate 9b
Ketone
3a
76
46
(ꢁ)
SS
A
4a
71
72
(ꢁ)
SS
A
5a
6a
6a
100
88
4a
88
58
(ꢁ)
RS
A
6a
Percentage conversion
Percentage ee
Major enantiopure
Configuration
Conditionc
100
80
100
74
100
66
(ꢁ)
SS
(ꢁ)
SS
(ꢁ)
SS
(ꢁ)
RS
B
A
A
B
a Ketone 3a: 78% yield, 60% ee (ꢁ)-(SS), condition B; ketone 4a: 80% yield, 86% ee (ꢁ)-(SS), condition B; ketone 5a: 90% yield, 90% ee (ꢁ)-(SS),
condition B; cf. ref. [9].
b Ketone 3a: 97% yield, 40% ee (ꢁ)-(RS), condition B; ketone 5a: 74% yield, 60% ee (ꢁ)-(RS), condition A; cf. ref. [9].
c A: DME/H2O, 0 8C, 1.4 equiv. oxone, 6 h; B: dioxane/H2O, rt, 3 equiv. oxone, 8 h.
observed (compare lines 4 and 5 in Table 1 with the same
lines and notes a and b in Table 2) and it appears that only
ketone 4 undergoes no inversion on passing from 4a to 4e.
In the case of ketones 3a–6a (having identical absolute
configurations, Fig. 1) the enantioselectivities are signifi-
cantly higher (compare Table 2 line 3 and notes a and b with
Table 1 line 3) and all types of ketones behave in the same
way (Table 2, lines 4, 5 and notes a and b) all providing
identical enantiomers: epoxides (ꢁ)-(SS), (ꢁ)-(SS) and (ꢁ)-
(RS), respectively for olefins 7–9.
p-methoxy cinnamate 9 with ketones 4a and 6a (leading to
(ꢁ)-(2R, 3S) cinnamate oxide) as well as epoxidation of
trans-stilbene 8 with ketone 6a (leading to (ꢁ)-(S, S)
stilbene oxide), are also correctly rationalized through this
(Me)Ph
Me(Ph)
Me(Ph)
(Me)Ph
(-)-(S,S)
(+)-(R,R)
O
O
Me(Ph)
(Me)Ph
(Me)Ph
Me(Ph)
While lower enantioselectivities (decrease in ee%) could
be due to contribution of direct epoxidation by oxone which
provides racemic epoxide, [18] inversions of enantioselec-
tivity can not be ascribed to direct epoxidation by oxone.
The stereochemical outcomes of epoxidations of trans-
stilbene with ketones 3a–5a (leading to (ꢁ)-(S, S) stilbene
oxide) and p-methoxy cinnamate with ketones 3a and 4a
(leading to (ꢁ)-(2R, 3S) cinnamate oxide) have already been
rationalized [7,9] by considering the model shown on
Scheme 4: conformation K1 of the dioxirane (predicted
to be the most populated because of the large i-Pr group
equatorial); spiro geometry; [19–22] equatorial approaches
of the olefin toward the dioxirane (with some contribution of
the axial approaches, although less favored [9]) and
n(F).p(Ar) or n(F).CH3 repulsion in approach E-I (making
the corresponding transition state more energetic than the
one corresponding to approach E-II).
Axial- approaches
A-I
A-II
O
O
O
O
Me
Me
X
X
K1
K1
F
F
E-I
E-II
Me(Ph)
Equatorial- approaches
(Me)Ph
(Me)Ph
Me(Ph)
O
(+)-(R,R)
(Me)Ph
O
(-)-(S,S)
)Ph
Me(Ph)
(M
e
Me(Ph)
E-II >>E-I(F/Ph n.π repulsion or F/CH repulsion) ;
3
A-I ~A-II when X=H ; A-I ~or slightly>A-II when X = Cl, F, Epox.
Scheme 4. Epoxidation of b-methylstyrene 7 by dioxiranes derived from
ketones (2S, 5R)-3a–6a having an axial fluorine in the expected most
populated conformation K1.
The epoxidations of b-methylstyrene with ketones 3a–6a
(leading to (ꢁ)-(S, S) epoxide), epoxidation of trans-methyl

´
A. Solladie-Cavallo et al. / Journal of Fluorine Chemistry 125 (2004) 1371–1377
1375
Approach on conformation K1 of dioxirane Approach on conformation K2 of dioxirane
X
X
O
O
O
O
O
O
O
O
F
F
X
X
K2
K1
K1
K2
Me
Me
Me
Me
F
F
E-I
E-II
Ph
E-I
E-II
Ph
Equatorial- approaches
Equatorial- approaches
Ph
Ph
Ph
Ph
Ph
Ph
O
O
O
O
(-)-(S,S)
(+)-(R,R)
(-)-(S,S)
(+)-(R,R)
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
(a)
(b)
E-II >> E-I : Me/Ph repulsion in E-I
E-II << E-I : F/Ph repulsion in E-II
Scheme 5. Epoxidation of trans-stilbene 8 with ketones (2R, 5R)-3e–6e, comparison between conformations K1 and K2.
model using the K1 conformation for the (2S, 5R)-dioxir-
anes (Scheme 4).
of ab initio calculations at the HF/6–31Gꢀ level [25]3 for
structure optimization [26,27] and conformational inter-
conversion pathways of (2S, 5R)-D5a and (2R, 5R)-D5e
are shown in Figs. 2 and 3.
The conformation K1 of (2S, 5R)-D5a is by 7.05 kcal molꢁ1
more stable than the second energy-minimum conformation
K2. The calculated strain energy barriers for the inter-conver-
sion of conformations K1 and K2 to twist-boat are 14.68 and
15.38 kcal molꢁ1, respectively (Fig. 2) corresponding to rapid
processes in consistency with literature results [23].
However, contrary to what is observed (low to nil and in
many cases inverted enantioselectivities), the use of this
model in the case of stilbene and (2R, 5R)-dioxiranes (derived
from ketones 3e–6e) leads to the prediction that equatorial
approach on conformer K1 will also provide epoxide (ꢁ)-(S,
S) as the major isomer (Scheme 5a), and that comparable (if
not as high) enantioselectivities should be expected (approach
E-II being favored over approach E-I due to steric CH3/Ph
hindrance in E-I). At this point it must be noted that an
epoxide of opposite configuration (þ)-(R, R) is expected to be
formed from the equatorial approach of stilbene on conformer
K2 (of (2R, 5R)-dioxiranes) (Scheme 5b).
In spite of the satisfying consistency between the model
and the experimental data in the case of (2S, 5R)-dioxiranes,
the model suffers from an important weakness: the necessary
hypothesis that K1 is the most populated conformation as
well as the most reactive.
The degenerate inter-conversion of twist-boat family of
(2S, 5R)-D5a with itself via the boat intermediate requires
2.35 kcal molꢁ1
.
On the other hand, the most stable conformation of (2R,
5R)-D5e is K2 (isopropyl and F group are axial) and the
second energy minimum conformation K1 is 0.54 kcal molꢁ1
above that of K2 (Fig. 3). The calculated strain energy barrier
for the inter-conversion of K2 to TB2 is 9 kcal molꢁ1, which
is higher than for inter-conversion into K1. Also the twist-
boat family of (2R, 5R)-D5e are not degenerate. This corre-
sponds to higher electrostatic repulsion between F and gauche
oxygen (which are held closer in the TB2 than in the TB1).
The dihedral angle fFCCO1 in the TB1 is ꢁ94.38, while the
same dihedral angle is ꢁ58.08 in the TB2 conformer.
Although these calculations have been done for D5a and
D5e, one can reasonably postulate that the trend will be the
same for D3a–e, D4a–e and D6a–e with K1 more populated
for D3a, D4a, D6a and K2 significantly populated for D3e,
D4e, D6e.
Ring inversion in six-membered ring (equilibration
between K1 and K2 conformations) is known to be a rapid
process [23] and the epoxidation reactions slow (about 6 h2
to be complete), the Curtin–Hammett principle holds and
conformation K2 could well contribute to the reaction if
reactive enough.
Interestingly, it has already been shown by Armstrong
et al. [22] using quantum mechanical methods, that a-fluoro
dioxirane is more reactive when the fluorine is axial than
when it is equatorial (Eact(TS with F-equatorial) ꢁ Eact(TS
with F-axial) ¼ þ2.5 kcal molꢁ1).
Thus, from this modeling, conformation K2 (axial fluor-
ine) of (2R, 5R)-D3e–6e (derived from ketones 3e–6e) is
significantly populated (with [K2] ꢂ [K1]) and expected to
be more reactive, [22] one can thus expect K2 to contribute
significantly to the reaction. As conformation K2 leads to
the epoxide of opposite configuration (compared with K1) a
These dioxiranes are generated in situ and cannot be
isolated, therefore an NMR study of their conformation in
solution is not possible; we have thus performed a modeling
of both dioxiranes derived from ketones 5a and 5e, respec-
tively (for modelization of other rings cf. [24]). The results
3 It must be noted that the most stable conformation K1 of (2S, 5R)-D5a
was calculated to be 2.61 kcal/mol more stable than the most stable
conformation K2 of (2R, 5R)-D5e.
2 Reactions are monitored by TLC using silicagel 60 F254 plates from
Merck (eluent: hexana/Et2O ¼ 9/1).

´
A. Solladie-Cavallo et al. / Journal of Fluorine Chemistry 125 (2004) 1371–1377
1376
15.38
14.68
11.89
9.54
9.54
O
7.05
F
F
F
F
F
O
O
O
F
TB1
TB2
O
O
K2
0.00
F
O
O
Dioxirane D5a
K1
F
Fig. 2. Ring inversion in dioxiranes: calculated strain energy (kcal molꢁ1) profile for (2S, 5R)-D5a.
9.00
fluorine equatorial) for dioxirane D5e derived from ketones
5e, while it is K1 which has the fluorine axial which is more
populated than conformation K2 (having the fluorine equa-
torial) for dioxirane D5a derived from ketones 5a. Accord-
ing to these calculations it has also been found that K1/K2
inter-conversions should be rapid processes (in the gas
phase) consistent with literature results (usually determined
in solution) [23]. Moreover, dioxirane conformations having
the fluorine axial (K2 for D5e or K1 for D5a) are expected to
be more reactive according to Armstrong/Houk ab initio
modeling. Based on the hypothesis that all a-fluorinated
dioxiranes (D3, D4 and D6) behave as D5, on slow epox-
idation reactions, rapid ring inversion of six-membered ring
dioxiranes, relative population of conformers K1 and K2
and the Curtin–Hammett principle, the decrease in enantios-
electivity and the inversion of configuration observed for the
epoxides obtained with ketones 3e–6e (compared to ketones
3a–6a) are due to competitive reactions of K1 and K2 in
dioxiranes D3e–D6e and to simultaneous production, in
these cases, of both (ꢁ) and (þ) epoxides with lowering
of enantioselectivity and even inversion of enantioselectivity
according to the extent of contribution of K2 to the reaction.
It thus appears that activation of ketones (as epoxidation
catalysts) by an a-fluorine atom is strongly orientation depen-
dant at least in the cases where flexible dioxiranes are formed.
4.88
4.13
2.93
2.03
F
1
O
F
0.54
F
F
O
O
F
O 1
0.00
TB2
O
F
O
TB1
O
K1
F
F
O
K2
Dioxirane D5e
Fig. 3. Ring inversion in dioxiranes: calculated strian energy (kcal molꢁ1
)
profile for (2R, 5R)-D5e.
lowering of the ee% and even an inversion of the enantios-
electivity (according to the extent of contribution of K2) is
expected, which is indeed observed.
In (2S, 5R)-D3a–6a dioxiranes derived from ketones 3a–
6a conformation K1 (F-axial) is more populated and more
reactive than K2 (F-equatorial) and will contribute almost
exclusively to the reaction providing solely the correspond-
ing enantiomer.
Acknowledgements
`
We are grateful to the Ministere de l’Education et de la
Recherche-France for a Ph.D. grant (MNERT) to L.J.
4. Conclusion
References
According to our calculations, and due to the presence of
the a-fluorine atom, conformation K2 having the fluorine
axial is more populated than conformation K1 (having the
¨
[1] (a) W. Adam, C.R. Saha-Moller, P.A. Ganeshpure, Chem. Rev. 101
(2001) 3499–3548;

´
A. Solladie-Cavallo et al. / Journal of Fluorine Chemistry 125 (2004) 1371–1377
1377
(b) M. Frohn, Y. Shi, Synthesis 14 (2000) 1979–2000;
(c) S.E. Denmark, Z. Wu, Synlett 4 (1999) 847–859.
[2] R. Curci, M. Fiorentino, M.R. Serio, J. Chem. Soc. Chem. Commun.
3 (1984) 155–156.
[22] A. Armstrong, I. Washington, K.N. Houk, J. Am. Chem. Soc. 122
(2000) 6297–6298.
[23] NMR observation of both conformers of substituted cyclohexanes
needs very low temperature (about ꢁ100 8C), in: E.L. Eliel, S.H.
Wilen, L.N. Mander (Eds.), for extended literature cf.: Stereochem-
istry of Organic Compounds, Wiley, 1994, pp. 686–753.
[24] (a) I. Yavari, H. Norouzi-Arasi, H. Fallah-Bagher-Shaidaei, J. Chem.
Res. 10 (1999) 616–617;
[3] (a) R. Mello, M. Fiorentino, C. Fusco, R. Curci, J. Am. Chem. Soc.
111 (1989) 6749–6757;
(b) R. Curci, L. D’Alccolti, M. Fiorentino, A. Rosa, Tetrahedron
Lett. 36 (1995) 5831–5834.
[4] D.S. Brown, B.A. Marples, P. Smitt, L. Walton, Tetrahedron 51
(1995) 3587–3606.
(b) I. Yavari, H. Norouzi-Arasi, D. Nori-Shargh, H. Fallah-Bagher-
Shaidaei, J. Chem. Res. 2 (1999) 154–155;
[5] S.E. Denmark, Z. Wu, M.C. Crudden, H. Matsuhashi, J. Org. Chem.
62 (1997) 8288–8289.
(c) I. Yavari, D. Tahmassebi, F. Nourmohammadian, J. Mol. Struct.
(THEOCHEM) 492 (1999) 209–215;
[6] A. Armstrong, B.R. Hayter, Chem. Commun. 5 (1998) 621–623.
(d) I. Yavari, A. Jabbari, D. Tahmassebi, M.R. Hosseini-Tabatabaei,
J. Mol. Struct. (THEOCHEM) 574 (2001) 1–7;
´
´
[7] A. Solladie-Cavallo, L. Bouerat, Org. Lett. 2 (2000) 3531–3534.
[8] A. Armstrong, B.R. Hayter, W.O. Moss, J.R. Reeves, J.S. Wailes,
Tetrahedron Asymm. 11 (2000) 2057–2061.
(e) I. Yavari, H. Norouzi-Arasi, J. Mol. Struct. (THEOCHEM) 593
(2002) 199–207.
´ ´
[9] A. Solladie-Cavallo, L. Bouerat, L. Jierry, Eur. J. Org. Chem. 23
[25] M.J. Frisch, G.W. Trucks, H.B. Schlegel, G.E. Scuseria, M.A. Robb,
J.R. Cheeseman, J.A. Montgomery, Jr., T. Vreven, K.N. Kudin, J.C.
Burant, J.M. Millam, S.S. Iyengar, J. Tomasi, V. Barone, B.
Mennucci, M. Cossi, G. Scalmani, N. Rega, G.A. Petersson, H.
Nakatsuji, M. Hada, M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa,
M. Ishida, T. Nakajima, Y. Honda, O. Kitao, H. Nakai, M. Klene, X.
Li, J.E. Knox, H.P. Hratchian, J.B. Cross, C. Adamo, J. Jaramillo, R.
Gomperts, R.E. Stratmann, O. Yazyev, A.J. Austin, R. Cammi, C.
Pomelli, J.W. Ochterski, P.Y. Ayala, K. Morokuma, G.A. Voth, P.
Salvador, J.J. Dannenberg, V.G. Zakrzewski, S. Dapprich, A.D.
Daniels, M.C. Strain, O. Farkas, D.K. Malick, A.D. Rabuck, K.
Raghavachari, J.B. Foresman, J.V. Ortiz, Q. Cui, A.G. Baboul, S.
Clifford, J. Cioslowski, B.B. Stefanov, G. Liu, A. Liashenko, P.
Piskorz, I. Komaromi, R.L. Martin, D.J. Fox, T. Keith, M.A. Al-
Laham, C.Y. Peng, A. Nanayakkara, M. Challacombe, P.M.W. Gill,
B. Johnson, W. Chen, M.W. Wong, C. Gonzalez, J.A. Pople,
Gaussian Inc., Pittsburgh PA, 2003.
(2001) 4557–4560.
[10] A. Armstrong, W.O. Moss, J.R. Reeves, Tetrahedron Asymm. 12
(2001) 2779–2781.
[11] A. Armstrong, G. Ahmed, B. Dominguez-Fernandez, B.R. Hayter,
J.S. Wailes, J. Org. Chem. 67 (2002) 8610–8617.
[12] S. Klein, S.M. Roberts, J. Chem. Soc. Perkin Trans. 1 19 (2002)
2686–2688.
[13] C.J. Stearman, V. Behar, Tetrahedron Lett. 43 (2002) 1943–1946.
[14] S.E. Denmark, H. Matsuhashi, J. Org. Chem. 67 (2002) 3479–
3486.
´
[15] A. Solladie-Cavallo, L. Jierry, A. Klein, Comptes Rendus Chimie
´
(Academie des sciences, Paris) 6 (2003) 603–606.
´ ´
[16] (a) A. Solladie-Cavallo, L. Bouerat, Tetrahedron Asymm. 11 (2000)
935–941;
´ ´
(b) A. Solladie-Cavallo, L. Jierry, L. Bouerat, P. Taillasson,
Tetrahedron Asymm. 12 (2001) 883–891.
´ ´
[17] A. Solladie-Cavallo, L. Jierry, L. Bouerat, M. Schmitt, Tetrahedron
[26] (a) J.J.P. Stewart, J. Comput. Chem. 10 (1989) 221–264;
(b) J.J.P. Stewart, J. Comput. Aided Mol. Des. 4 (1990) 1–105;
(c) M.J.S. Dewar, E.F. Healy, J.J.P. Stewart, J. Chem. Soc., Faraday
Trans. 80 (1984) 227–233 (keyword SADDLE).
58 (2002) 4195–4199.
[18] Direct oxidation of olefins by oxone is known to be a slow process,
cf.: J.O. Edwards, R.H. Pater, R. Curci, F. DiFuria, Photochem.
Photobiol. 30 (1979) 63–69 (and included references).
[19] A.L. Baumstark, C.J. McCloskey, Tetrahedron Lett. 28 (1987)
3311–3314.
[27] (a) F. Jensen, Introduction to Computational Chemistry, Wiley, New
York, 1999;
(b) W.J. Hehre, L. Radom, P.R. Schleyer, J.A. Pople, Ab initio
Molecular Orbital Theory, Wiley, New York, 1986;
[20] K.N. Houk, J. Liu, C. DeMello, K.R. Condroski, J. Am. Chem. Soc.
119 (1997) 10147–10152.
(c) P.C. Hariharan, J.A. Pople, Theor. Chim. Acta 28 (1973) 213–222;
(d) R.S. Grev, C.L. Janssen, H.F. Schaefer, Chem. Phys. 95 (1991)
5128–5132.
[21] C. Jenson, J. Liu, K.N. Houk, W.L. Jorgensen, J. Am. Chem. Soc.
119 (1997) 12982–12983.