Regio-, diastereo-, and chemoselectivities in the dioxirane oxidation of acyclic and cyclic allylic alcohols by methyl(trifluoromethyl)dioxirane (TFD): A comparison with dimethyldioxirane

Article abstract of DOI:10.1002/(sici)1099-0690(199802)1998:2<349::aid-ejoc349>3.0.co;2-r

The solvent-dependent shift in the regioselectivity of the geraniol epoxidation by methyl(trifluoromethyl)dioxirane (TFD) reveals that as for the less reactive dimethyldioxirane (DMD). hydrogen bonding stabilizes the transition state of the epoxidation. In protic media, the hydrogen bonding is exerted intermolecularly by the solvent, whereas in unpolar, non-hydrogen-bonding solvents intramolecular assistance through the adjacent hydroxy functionality comes into the play and the attack on the allylic alcohol moiety is favored. For chiral allylic alcohols, additional steric interactions control the π-facial selectivity in the conformationally fixed transition state. Analogous to DMD, the preferred dihedral angle in the hydrogen-bonded transition state of the TFD epoxidation constitutes approximately 130°, but contrary to DMD and for synthetic purposes important, the allylic alcohols and derivatives 1 and 3-5 investigated here are chemoselectively epoxidized by TFD without formation of the corresponding enones.

Full text of DOI:10.1002/(sici)1099-0690(199802)1998:2<349::aid-ejoc349>3.0.co;2-r

FULL PAPER
Regio-, Diastereo-, and Chemoselectivities in the Dioxirane Oxidation of
Acyclic and Cyclic Allylic Alcohols by Methyl(trifluoromethyl)dioxirane
(TFD): A Comparison with Dimethyldioxirane
Waldemar Adam*a, Rodrigo Paredesb, Alexander K. Smerza, and L. Angela Velozab
Institute of Organic Chemistry, University of Würzburg,a
Am Hubland, D-97074 Würzburg, Germany
Fax: ϩ49(0)931/8884756
E-mail: adam@chemie.uni-wuerzburg.de
Department of Chemistry,b
Universidad del Valle, A.A. 25360 Cali, Colombia
Received August 11, 1997
Keywords: Dioxirane / Oxidation / Allylic alcohols / Regioselectivity / Diastereoselectivity / Chemoselectivity /
Epoxides / Enones
The solvent-dependent shift in the regioselectivity of the ge- allylic alcohols, additional steric interactions control the π-
raniol epoxidation by methyl(trifluoromethyl)dioxirane (TFD)
facial selectivity in the conformationally fixed transition
reveals that as for the less reactive dimethyldioxirane (DMD), state. Analogous to DMD, the preferred dihedral angle in the
hydrogen bonding stabilizes the transition state of the epoxi- hydrogen-bonded transition state of the TFD epoxidation
dation. In protic media, the hydrogen bonding is exerted in-
constitutes approximately 130°, but contrary to DMD and for
termolecularly by the solvent, whereas in unpolar, non-hy- synthetic purposes important, the allylic alcohols and deriva-
drogen-bonding solvents intramolecular assistance through tives 1 and 3؊5 investigated here are chemoselectively epo-
the adjacent hydroxy functionality comes into the play and
the attack on the allylic alcohol moiety is favored. For chiral
xidized by TFD without formation of the corresponding eno-
nes.
Introduction
towards competitive allylic oxidation instead of the ex-
pected epoxidation.
After a decade of mechanistic studies on the reactivity of
dioxiranes,[1] recent work has been focused on the selec-
tivity of these useful oxidants.[2] Special attention was paid
to the π-facial selectivity of dimethyldioxirane (DMD) in
epoxidations[3][4][5], since such oxygen-transfer reactions
constitute a valuable tool in the diastereoselective syn-
thesis.[6] Recent results have demonstrated that DMD dis-
plays slightly higher diastereoselectivities as epoxidizing
agent than mCPBA when purely steric factors control the
π-facial preference.[3b][5] Moreover, in analogy to peracids,
also for DMD the transition state of the epoxidation is sta-
bilized by hydrogen bonding with OH[3][4] and NH2[7] func-
tionalities. This appreciable diastereoselectivity-directing
stabilization may either take place intermolecularly by a
protic solvent or intramolecularly by an allylic hydroxy or
amino group. Besides the stereoselectivity, the transition
state stabilization through hydrogen bonding is further sub-
In the present study we have examined the oxidation of
acyclic allylic alcohols 1 and the cyclic substrates 3؊5 by
methyl(trifluoromethyl)dioxirane (TFD). Two questions
were of interest: on one hand, how sensitive is the more
reactive TFD (compared to DMD) towards stabilization ef-
fects through hydrogen bonding by testing its regioselectiv-
ity and π-facial diastereoselectivity in the epoxidation; on
the other hand, how chemoselective is the more reactive
TFD (compared to DMD) as reflected in the extent of en-
one formation versus epoxidation. The results described
herein unequivocally display that the more reactive TFD is
definitely less regioselective, about as diastereoselective, but
significantly more chemoselective than DMD.
Results and Discussion
All TFD epoxidation of geraniol 1a and its derivatives
stantiated by kinetic competition experiments on the sol- 1bϪd were performed as with DMD[4a][4b] to only 30% con-
vent-dependent regioselectivity of the geraniol epoxida- version in order to prevent any bisepoxide formation. As
tion.[4a][4b] These selectivity trends reveal that hydrogen shown in Table 1, geraniol (1a) is preferentially epoxidized
bonding is weaker for DMD than for mCPBA.[8]
by TFD at the 6,7 position, but the regioselectivity de-
In competition with such hydroxy-directed dia- creases (entries 1Ϫ3) from that in the most polar MeOH/
stereoselective epoxidations, also allylic oxidation to the TFA (9:1) to the least polar CCl4/TFA (9:1) medium (TFA
corresponding enone has been observed for a variety of al- stands for 1,1,1-trifluoroacetone). In contrast, the geraniol
lylic alcohols in the DMD oxidations. Especially cyclic sub- methyl ether (1b) shows no detectable solvent effect (entries
strates and vinylsilanes show a pronounced tendency 4 to 6). For the trimethylsilyl ether (1c) and acetate (1d)
Eur. J. Org. Chem. 1998, 349Ϫ354
WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1998
1434Ϫ193X/98/0202Ϫ0349 $ 17.50ϩ.50/0
349

W. Adam, R. Paredes, A. K. Smerz, L. A. Veloza
Table 1. Solvent effects on the regioselectivity in the TFD epoxidation of geraniol (1a) and derivatives 1b؊d
FULL PAPER
entry
1
substrate
X
H
medium[a]
conversion[b]
[%]
6,7/2,3
distribution[c]
DMD[d]
TFD
1a
MeOH/TFA
(9:1)
21
73:27
88:12
2
3
1a
1a
H
H
TFA
22
25
68:32
57:43
74:26
51:49
CCl4/TFA
(9:1)
4
1b
Me
MeOH/TFA
(9:1)
31
68:32
88:12
5
6
1b
1b
Me
Me
TFA
32
32
73:27
71:29
88:12
87:13
CCl4/TFA
(9:1)
7
8
1c
1d
SiMe3
Ac
TFA
TFA
33
30
70:30
83:17
85:15
96:4
[a] TFA ϭ 1,1,1-trifluoroacetone. Ϫ [b] Mass balances >90%. Ϫ [c] Determined by 1H-NMR analysis directly on the crude product mixture
[d]
(error ±5% of the stated values). Ϫ Ref.[4b], values given for identical solvent mixtures with acetone instead of TFA.
of geraniol, again the 6,7 regioisomer is favored (entries 7 tone/acetone as in pure acetone (cf. footnotes e and g in
and 8).
The π-facial diastereoselectivity was tested for the set of
methyl-substituted, acyclic allylic alcohols 1eϪi (Table 2).
The substrates 1e and (E)-1g without methyl substituents in
Table 3).
Regioselectivity
The regioselectivity data (Table 1) for geraniol (1a) and
the gem- or cis- positions react rather unselectively to yield its oxygen-functionalized derivatives 1b؊d express clearly
nearly 50:50 diastereoselectivities, irrespective of the solvent the hydroxy-directing effects in the TFD epoxidations.
used (Table 2, entries 1, 4, and 5). For the gem-disubstituted Analogous to DMD, in all solvent systems the terminal 6,7
substrate 1f with 1,2-allylic (A1,2) strain, the threo selectivity rather than the 2,3 double bond is preferentially epoxidized.
is enhanced in the CCl4/TFA (9:1) compared to TFA (en- Although both double bonds are trisubstituted, the nucleo-
tries 2 and 3). Still higher diastereoselectivities were ob- philicity of the 2,3 double bond is significantly lowered by
tained for the substrates (Z)-1g (entries 6 and 7) and 1h the inductively electron-accepting hydroxy group. As to the
(entries 8 and 9) with A1,3 strain, which again increased in solvent effects, methanol (entry 1) stabilizes both dipolar
the unpolar CCl4/TFA (9:1) mixture. For the derivative 1i transition states to the same extent, so that intramolecular
with both A1,2 and A1,3 strain, the diastereoselectivities are hydrogen bonding by the allylic hydroxy functionality does
the highest of the substrates examined herein (entries 10 not come into play and the regioisomeric product distri-
and 11).
bution is merely a function of the relative nucleophilicity of
Unexpected was the fact that the trans-disubstituted sub- the two double bonds. In contrast, in the least polar and
strate (E)-1g was cleanly epoxidized by TFD without allylic non-protic medium 9:1 CCl4 and trifluoroacetone (entry 3),
oxidation, whereas with DMD as much as 10% of the cor- external hydrogen bonding is absent and the dipolar tran-
responding enone was obtained.[4b] This encouraged us to sition state to afford the 2,3-epoxide is assisted through in-
probe more deliberately such allylic oxidations of TFD with tramolecular hydrogen bonding by the allylic hydroxy func-
cyclic allylic alcohols.[4a] The results in Table 3 show that tionality. The observed small but definite shift to more 2,3-
for TFD exclusively epoxidation takes place with increased epoxide provides evidence for the conformationally pre-
cis diastereoselectivities in the oxygen transfer for the par- ferred six-membered-ring, hydrogen-bonded transition state
ent 2-cyclohexenol (3) and for isophorol (4) in the less polar (Figure 1). Thus, the unfavorable π nucleophilicity of the
medium (entries 1Ϫ4). Even for the substrate 5 (enty 5), 2,3 double bond is counteracted by the favorable hydroxy-
which is the most susceptible one towards allylic oxidation group assistance. The stabilization is, however, less pro-
(DMD gave 57% enone[4a]), no allylic oxidation was ob- nounced for TFD than for DMD, as manifested by the
served. Noteworthy, the DMD oxidation of substrate 5 smaller solvent-induced shift in the regioselectivity for the
yields quite similar epoxide/enone ratios in 4:1 trifluoroace- former (compare last two columns in Table 1).
350
Eur. J. Org. Chem. 1998, 349Ϫ354

Dioxirane Oxidation of Allylic Alcohols by Methyl(trifluoromethyl)dioxirane
Table 2. Solvent effects on the diastereoselectivities in the dioxirane epoxidation of allylic alcohols 1
FULL PAPER
entry
substrate
R1
R2
R3
medium[a]
m.b.[b] [%]
conv. [%][c,d] threo/erythro
selectivity[d]
DMD[e]
TFD
1
2
1e
H
H
H
B
A
B
A
B
A
B
A
B
A
B
90
99
92
97
87
98
68
95
82
99
84
>95
>95
>95
>95
76
55:45
53:47
62:38
48:52
43:57
77:23
88:12
76:24
82:18
85:15
90:10
50:50
60:40
70:30
53:47[f]
56:44[f]
67:33
85:15
76:24
82:18
87:13
91:9
1f
Me
Me
H
H
H
3
1f
H
H
4
(E)-1g
(E)-1g
(Z)-1g
(Z)-1g
1h
Me
Me
H
H
5
H
H
6
H
Me
Me
Me
Me
Me
Me
95
7
H
H
93
8
H
Me
Me
H
>95
>95
>95
60
9
1h
H
10
11
1i
Me
Me
1i
H
[a]
[b]
[c]
[d]
1
A: 1,1,1-trifluoroacetone (TFA), B: CCl4/TFA (9:1). Ϫ Mass balance. Ϫ Yield >95% in all cases. Ϫ Determined by H-NMR
analysis of characteristic signals (error ±5% of the stated values). Ϫ Ref.[4b], values given for identical solvent mixtures with acetone
instead of TFA. Ϫ Epoxide/enone ratio ca. 90:10.
[e]
[f]
Table 3. Diastereoselectivities in the dioxirane epoxidations of cyclic allylic alcohols 3؊5
entry
R1
R2
R3
medium[a]
m.b.[b]
[%]
conv.[c,d]
[%]
cis/trans
selectivity[d]
DMD[e]
TFD
1
2
3
4
5
3
3
4
4
5
H
H
H
A
B
A
B
A
85
96
70
78
84
70
80
57:43
70:30
88:12
94: 6
43:57 [52:48]
66:34 [61:39]
83:17 [77:23]
96: 4 [92: 8]
H
H
H
Me
Me
H
Me
Me
H
Me
Me
tBu
100
92
99
73:27[f]
60:40[f] [44:56][g]
[a]
[b]
[c]
A: 1,1,1-trifluoroacetone (TFA), B: CCl4/TFA (9:1); for DMD, acetone instead of TFA applies (cf. ref.[4a]). Ϫ
Mass balance. Ϫ
Yield >95% in all cases. Ϫ Determined by 1H-NMR analysis of characteristic signals (error ±5% of the stated values). Ϫ [e] For DMD
[d]
also appreciable amounts of enone were obtained, cf. ref.[4b], the epoxide/enone ratio (chemoselectivity) is given in brackets. Ϫ [f] Relative
[g]
configuration with respect to the hydroxy group. Ϫ
acetone and for an 4:1 mixture of TFA/acetone.
The epoxide/enone ratio is the same (within the experimental error) for pure
Figure 1. Stabilization of the spiro transition state through intra-
regioselective epoxidation of the methyl ether 1b. In this
case, no allylic hydroxy proton is available to assist intra-
molecularly in the epoxidation of the less nucleophilic 2,3
double bond. Hence, the ratio of 6,7- to 2,3-epoxides is
within the experimental error the same in all three solvents
employed (Table 1, entries 4Ϫ6) and the regioselectivity is
similar to that for geraniol (1a) in methanol as cosolvent
(entry 1). Moreover, the corresponding trimethylsilyl ether
1c with similar electronic properties as the methyl ether 1b
(enty 7) yields a regioselectivity identical to that for the
methyl ether 1b (entry 5). The acetate 1d (entry 8), however,
molecular hydrogen bonding
The assistance through intramolecular hydrogen bonding shows a more pronounced 6,7 regioselectivity, which is due
by the allylic hydroxy group is further substantiated by the to the stronger deactivation of the 2,3 double bond by the
Eur. J. Org. Chem. 1998, 349Ϫ354
351

W. Adam, R. Paredes, A. K. Smerz, L. A. Veloza
FULL PAPER
higher inductive electron-withdrawal of the acetoxy func- 11) than for derivative 1h with only A1,3 (entry 9) or 1f with
tionality.
only A1,2 strain (entry 3).
Noteworthy is the fact that the preference for the 6,7
Figure 2. Diastereomeric transition states in the epoxidation of al-
double-bond epoxidation is generally less pronounced for
TFD than for DMD (compare the last two columns in
Table 1), i.e. TFD as the more reactive oxidant responds
less sensitively to the relative nucleophilicity of the two
double bonds. Only in the CCl4 solvent mixture (Table 1,
entry 3) shows TFD a slightly higher propensity towards
6,7 epoxidation than DMD. In this non-polar medium, the
intramolecular stabilization by the allylic hydroxy func-
tionality is coming fully into play and the inversion of the
general trend reflects the more pronounced hydrogen bond-
ing for DMD as compared to TFD.
lylic aclcohol 1i
Diastereoselectivity
As a consequence of stabilization effects through intra-
molecular hydrogen bonding, the diastereoselectivity
should also be directed in the TFD epoxidation of chiral
allylic alcohols by the hydroxy group, as already demon-
strated for DMD[4]. Indeed, for TFD a similar solvent de-
pendence of the threo/erythro diastereoselectivity is ob-
served in the epoxidation of the acyclic substrates 1e؊i as
for DMD (compare the last two columns in Table 2). The
substrates 1e and (E)-1g without any allylic strain[9] (entries
1, 4, and 5), do not display any diastereoselectivity nor sol-
vent effects in the dioxirane epoxidations. Therefore, the
lack of allylic strain provides no steric differentiation of the
two π faces irrespective whether a hydrogen-bonded stabi-
lization operates in the transition state. For the substrate 1f
Furthermore, also the diastereoselectivity data for the
epoxidation of the cyclic allylic alcohols 3Ϫ5 (Table 3) by
TFD is in good accord with the results for DMD.[4a][4b]
Therefore, also these stereochemical results substantiate the
conclusions derived from the acyclic systems (Figure 2) in
regard to the suggested transition-state geometry.
Chemoselectivity
As stated initially, our interest in the cyclic substrates
with only A1,2 strain (entries 2 and 3) and for the derivatives 3Ϫ5 was to test the chemoselectivity of the more reactive
(Z)-1g and 1h with only A1,3 strain (entries 6Ϫ9), the dia- TFD in terms of enone formation by CH insertion versus
stereoselectivities are always significantly higher in the un- epoxidation. In this context, the finding was unexpected
polar 9:1 TFA/CCl4 solvent mixture (entries 3, 7, and 9) that for TFD all three substrates were exclusively epoxi-
than in the more polar TFA (entries 2, 6, and 8). As sug- dized without even traces of enone (Table 3). Analogous to
gested in Figure 1, in the non-polar solvent mixture the the enone formation in the oxidation of allylic alcohols by
transition state geometry is best fixed by intramolecular hy- VO(acac)2/tBuOOH as oxidant,[11] a convincing mechan-
drogen bonding with the allylic hydroxy group and, hence, istic rationale is yet to be given for the unusual DMD reac-
the better the π-facial control. Notably, as for DMD, also tivity.
for TFD (compare last two columns in Table 2), A1,3 strain
In view of the higher propensity of TFA versus acetone
in the cis-substituted substrates (Z)-1g (entry 7) and 1h (en- to form hydrates with water, we suspected that the allylic
try 9) is more effective in dictating the threo selectivity than alcohol was converted to the corresponding hemiacetal
the A1,2 strain in the gem-substituted derivative 1f (entry 3). structure 9 in the TFA medium and thereby the substrate
This is unequivocally demonstrated by the stereochemical was deactivated towards allylic CH oxidation to the enone.
probe,[10] namely the chiral allylic alcohol 1i (entry 11) with This is akin to the lack of CH bond oxidation in acetates
both A1,2 and A1,3 strain, which displays the highest threo and also acetyl derivatives of allylic alcohols.[3b][4a][4b][12]
diastereoselectivity of all the acyclic substrates (Table 2) However, that this explanation does not apply is convinc-
examined here.
ingly demonstrated by the control experiment (cf. footnotes
For convenience, the two possible diastereomeric tran- e and g in Table 3) with substrate 5, in which the epoxide/
sition states of the oxygen transfer for the stereochemical enone ratio was for DMD the same (within the experimen-
probe 1i are exhibited in Figure 2. Since with both TFD tal error) in the 4:1 TFA/acetone mixture as in pure acetone.
and DMD the same threo selectivity is observed (entry 11), Hence, even though TFD is the more reactive oxidant than
the same CϭCϪCϪO dihedral angle (α ca. 130°)[4a][4b] ap- DMD, it is also significantly more chemoselective in terms
plies. Consequently, it is not surprising that A1,3 dominates of epoxidation versus allylic oxidation.
over A1,2 strain. In fact, a slight synergistic effect operates
This apparent violation of the reactivity-selectivity prin-
since the preference for the threo attack is significantly ciple (RSP) is remarkable since the regioselectivity of the
larger for substrate 1i with both A1,3 and A1,2 strain (entry geraniol epoxidation (Table 1) followed the expected reac-
352
Eur. J. Org. Chem. 1998, 349Ϫ354

Dioxirane Oxidation of Allylic Alcohols by Methyl(trifluoromethyl)dioxirane
FULL PAPER
were used as received, solvents were purified and dried by reported
standard methods. The hitherto unknown epoxide 6,7-2c was
fully characterized.
General Procedure for the Epoxidations of Alkenes by Methyl(tri-
fluoromethyl)dioxirane: The alkene 1, 3, 4, or 5 was dissolved in an
organic solvent (cf. Table 1Ϫ3) and 0.3Ϫ1.1 equiv. of methyl-
(trifluoromethyl)dioxirane (0.2Ϫ0.45  solution in acetone) was ra-
pidly added at 0Ϫ10°C. The solution was stirred at this tempera-
ture until the peroxide test (KI/HOAc) was negative. The solvent
was removed (20°C, 20Ϫ100 Torr) to afford a mixture of the corre-
sponding epoxides in high purity and the product mixture analyzed
by 1H NMR spectroscopy. The quantitative results are summarized
in Tables 1Ϫ3.
tivity trend, i.e. the more reactive TFD is also the less re-
gioselective. Presumably, for competitive epoxidations with
similar transition states, as is unquestionably the case in the
regioselectivity for geraniol (1a), the RSP is obeyed and
many cases have been reported.[13] Nevertheless, for pro-
cesses with different transition states, as must be undoubt-
edly the case for the competitive allylic oxidation versus
epoxidation, the RSP fails.
2,2-Dimethyl-3-[3-methyl-5-trimethylsilyloxy-(E)-pent-3-enyl]-
oxirane (6,7-2c): A 0.08- solution of dimethyldioxirane (5.45 ml,
0.44 mmol) in 1,1,1-trifluoroacetone was added to 300 mg (1.33
mmol) alkene. The solution was stirred at ca. 20°C for 15 min and
the solvent was removed (20°C, 20 Torr) to yield, at a conversion
of 35%, 312 mg of a colorless oil. The two regioisomers were sepa-
rated by low-temperature (Ϫ10°C) column chromatography on sil-
ica gel with petroleum ether/Et2O (4:1) as eluent to yield 100 mg
(32%) of the regioisomerically pure epoxide 6,7-2c. Ϫ 1H NMR
(250 MHz, CDCl3): δ ϭ 0.12 (s, 9 H), 1.25 (s, 3 H), 1.30 (s, 3 H),
1.66 (s, 3 H), 1.59Ϫ1.71 (m, 2 H), 2.03Ϫ2.27 (m, 2 H), 2.71 (dd,
J1 ϭ 6.4, J2 ϭ 6.0 Hz, 1 H), 4.15 (d, J ϭ 6.4 Hz, 2 H), 5.36 (mdd,
J1 ϭ 6.7, J2 ϭ 6.4 Hz, 1 H). Ϫ 13C NMR (63 MHz, CDCl3): δ ϭ
0.0 (3ϫ q) 16.3 (q), 18.7 (q), 24.8 (q), 27.1 (t), 36.1 (t), 58.4 (s),
59.3 (t), 64.0 (d), 124.3 (d), 136.7 (s). Ϫ IR (CCl4): ν˜ ϭ 2940, 1440,
1370, 1240, 1110, 1060, 870, 840 cmϪ1. Ϫ calcd. C13H26OSi (242.4):
C 64.41, H 10.81; found: C 63.91, H 11.08.
Conclusion
In conclusion, our present study on the regioselectivity
and diastereoselectivity of TFD oxidations for a variety of
substrates manifests that also for this highly reactive dioxir-
ane, intramolecular stabilization through hydrogen bonding
operates in the epoxidation of allylic alcohols. The ge-
ometry of the postulated six-membered-ring cyclic tran-
sition state is similar to the one postulated for DMD, as
derived from the diastereoselectivities displayed by the chi-
ral allylic alcohols. Furthermore, the more nucleophilic
double bond is preferentially epoxidized, provided no
hydroxy-directing effect counteracts this nucleophilicity
trend. These facts also imply that as for DMD,[14] the epox-
idation by TFD proceeds by an oxenoid attack of the dioxi-
rane on the double bond to form a partially polarized tran-
sition state. Of synthetic importance is the fact that TFD
possesses a much higher chemoselectivity than DMD, since
none of the allylic alcohols 1 and 3Ϫ5 were oxidized to the
corresponding enone, which would have been hardly antici-
pated.
[1] [1a]
W. Adam, R. Curci, J. O. Edwards, Acc. Chem. Res. 1989,
[1b]
22, 205Ϫ211. Ϫ
1187Ϫ1201. Ϫ
R. W. Murray, Chem. Rev. 1989, 89,
R. Curci, in Advances in Oxygenated Pro-
[1c]
cesses; A. L. Baumstark, Ed.; JAI: Greenwich, CT, 1990; Vol
[1d]
2, Chapter I. Ϫ
W. Adam, L. P. Hadjiarapoglou, R. Curci,
R. Mello, in Organic Peroxides; W. Ando, Ed.; Wiley, New
York, 1992; Chapter 4, pp 195Ϫ219. Ϫ [1e] W. Adam, L. Hadjia-
rapoglou, Top. Curr. Chem. 1993, 164, 45Ϫ62.
[2] [2a]
Financial support by the Deutsche Forschungsgemeinschaft
(Schwerpunktprogramm “Peroxidchemie: Mechanistische und
präparative Aspekte des Sauerstofftransfers”) and the Fonds der
Chemischen Industrie is gratefully appreciated. L. A. V. thanks
COLCIENCIAS for a doctoral fellowship (1995Ϫ1996) to conduct
some of her research work in Würzburg.
R. Curci, A. Dinoi, M. F. Rubino, Pure Appl. Chem. 1995,
[2b]
67, 811Ϫ822. Ϫ
W. Adam, A. K. Smerz, Bull. Soc. Chim.
Belg. 1996, 105, 581Ϫ599.
[3]
[4]
[5]
[3a] R. W. Murray, M. Singh, B. L. Wiliams, H. Moncrieff, Tetra-
[3b]
hedron Lett. 1995, 36, 2437Ϫ2440. Ϫ
R. W. Murray, M.
Singh, B. L. Williams, H. M. Moncrieff, J. Org. Chem. 1996,
61, 1830Ϫ1841.
[4a] W. Adam, A. K. Smerz, Tetrahedron 1995, 51, 13039Ϫ13044.
[4b]
Ϫ
W. Adam, A. K. Smerz, J. Org. Chem. 1996, 61,
[4c]
3506Ϫ3510. Ϫ
W. Adam, R. Paredes, A. K. Smerz, L. A.
Experimental Section
Veloza, Liebigs Ann. 1997, 547Ϫ551.
E. Vedejs, W. H. Dent III, J. T. Kendall, P.A. Oliver, J. Am.
Chem. Soc. 1996, 118, 3556Ϫ3567.
1
General Aspects: H- and 13C-NMR spectra were recorded with
[6]
[7]
a Bruker AC 200 and AC 250 spectrometer by using CDCl3 as
internal standard. Potassium iodide starch paper (Merck) was used
P. Besse, H. Veschambre, Tetrahedron 1994, 50, 8885Ϫ8927.
`
´
G. Asensio, R. Mello, C. Boix Bernadini, M. E. Gonzalez Nu-
n˜ez, G. Castellano,
for the peroxide tests.
Ϫ
Methyl(trifluoromethyl)dioxirane
J. Org. Chem. 1995, 60, 3692Ϫ3699.
(TFD)[15] and dimethyldioxirane (DMD)[16] were prepared accord-
ing to the described literature procedure, whereby essentially ke-
tone-free solutions were obtained by the extraction procedure. The
starting materials 1b[17], 1c[18], and 1d[19] were prepared according
to literature procedures, 1f؊i[4a] and 3؊5[4b] were made as earlier
described. The product ratios were determined by comparison of
[8]
W. Adam, H.-G. Brünker, A. S. Kumar, E.-M. Peters, K. Peters,
U. Schneider, H. G. von Schnering, J. Am. Chem. Soc. 1996,
118, 1899Ϫ1905.
[9]
R. W. Hoffmann, Chem. Rev. 1989, 89, 1841Ϫ1860.
W. Adam, B. Nestler, Tetrahedron Lett. 1993, 34, 611Ϫ614.
P. Kocovsky, J. Chem. Soc. Perkin Trans. 1 1994, 1759Ϫ1763.
[10]
[11]
[12] [12a]
´
´
G. Asensio, R. Mello, M. E. Gonzalez-Nun˜ez, G. Castel-
the spectral data with those given in the literature for 2,3-2a[4a]
,
,
lano, J. Corral, Angew. Chem. Int. Ed. 1996, 35, 217Ϫ218. Ϫ
[12b]
W. Adam, F. Prechtl, M. J. Richter, A. K. Smerz, Tetra-
6,7-2a[4a]
, , , , ,
2,3-2b[22] 6,7-2b[17] 2,3-2c[21] 2,3-2d[22] 6,7-2d[22]
hedron Lett. 1993, 8427Ϫ8430.
2e؊i[4a], 3[4b], 4[4b], and 5[4b] or by independent preparation accord-
ing to literature procedures. The commercial compounds 1a and 1e
[13] [13a]
E. Buncel, H. Wilson, J. Chem. Ed. 1986, 64, 476Ϫ482. Ϫ
B. Giese, Angew. Chem. Int. Ed. 1977, 16, 125Ϫ136.
[13b]
Eur. J. Org. Chem. 1998, 349Ϫ354
353

W. Adam, R. Paredes, A. K. Smerz, L. A. Veloza
FULL PAPER
[14]
[18]
[19]
[20]
[21]
[22]
W. Adam, R. Curci, L. DЈAccolti, A. Dinoi, C. Fusco, F. Gas-
E. Negishi, S. Chatterjee, H. Matsushita, Tetrahedron Lett.
1981, 22, 3737Ϫ3740.
parrini, R. Kluge, R. Paredes, M. Schulz, A. K. Smerz, L. A.
Veloza, S. Weinkötz, R. Winde, Chem. Europ. J. 1997, 3,
105Ϫ109.
H. Shibuya, N. Murakami, F. Shimada, J. Kitagawa, Chem.
Pharm. Bull. 1991, 39, 1302Ϫ1304
[15]
R. Mello, M. Fiorentino, O. Sciacovelli, R. Curci, J. Org. Chem.
M. Neli, B. Rindone, S. Tollari, L. Valletta, Gazz. Chim. Ital.
1987, 117, 207Ϫ212.
1988, 53, 3891Ϫ3893.
[16] [16a]
R. W. Murray, R. J. Jeyaraman, J. Org. Chem. 1985, 50,
A. Yosuda, S. Tamaka, H. Yamamoto, H. Nozuki, Bull. Chim.
Soc. Jpn 1979, 52, 1752Ϫ1756.
[16b]
2847Ϫ2853. Ϫ
Chem. Ber. 1991, 124, 2377.
M. A. Schwartz, T. J. Dunn, J. Am. Chem. Soc. 1972, 94,
4205Ϫ4212.
W. Adam, L. Hadjiarapoglou, J. Bialas,
G. Cicala, R. Curci, M. Fiorentino, O. Laricchiuta, J. Org.
Chem. 1982, 47, 2670Ϫ2673.
[17]
[97240]
354
Eur. J. Org. Chem. 1998, 349Ϫ354