Molecule-Induced Homolysis versus "Concerted Oxenoid Oxygen Insertion" in the Oxidation of Organic Compounds by Dimethyldioxirane

Article abstract of DOI:10.1021/jo971226w

Evidence for a molecule-induced homolysis of dimethyldioxirane by several classes of organic compounds (alkanes, alkenes, ethers, alcohols, aldehydes, iododerivatives) is reported. Carbon-centered radicals, arising from alkanes, ethers, and aldehydes, are trapped by CBrCl₃ or protonated quinolines. The dramatic influence of oxygen in these reactions, as well as the formation of products of induced homolysis of the dioxirane by carbon-centered radicals (CH₄, CH₃OH, CH₃COOCH₃, ROCOCH₃, CH₃COOCH₂COCH₃), strongly supports a radical mechanism. With alkenes and iodo derivatives the induced homolysis would lead to diradical intermediates, whose very fast fragmentation would prevent detection, but circumstantial evidence supports a radical mechanism.

Full text of DOI:10.1021/jo971226w

254
J . Org. Chem. 1998, 63, 254-263
Molecu le-In d u ced Hom olysis ver su s “Con cer ted Oxen oid Oxygen
In ser tion ” in th e Oxid a tion of Or ga n ic Com p ou n d s by
Dim eth yld ioxir a n e
Anna Bravo, Francesca Fontana, Giovanni Fronza, Francesco Minisci,* and Lihua Zhao
Dipartimento di Chimica del Politecnico, via Mancinelli 7, I-20131 Milano, Italy
Received J uly 7, 1997X
Evidence for a molecule-induced homolysis of dimethyldioxirane by several classes of organic
compounds (alkanes, alkenes, ethers, alcohols, aldehydes, iododerivatives) is reported. Carbon-
centered radicals, arising from alkanes, ethers, and aldehydes, are trapped by CBrCl3 or protonated
quinolines. The dramatic influence of oxygen in these reactions, as well as the formation of products
of induced homolysis of the dioxirane by carbon-centered radicals (CH4, CH3OH, CH3COOCH3,
ROCOCH3, CH3COOCH2COCH3), strongly supports a radical mechanism. With alkenes and iodo
derivatives the induced homolysis would lead to diradical intermediates, whose very fast
fragmentation would prevent detection, but circumstantial evidence supports a radical mechanism.
In tr od u ction
LFER studies5,6 would indicate similar concerted elec-
trophilic transition states also for the oxidation of alkanes
both by DMD and by peracids (structures 2 and 3).
Dioxiranes represent a class of oxidants of great
current fashion for the oxidation of organic compounds.
While this interest is undoubtedly justified from the
theoretical standpoint, it is less so for synthetic and
practical purposes, despite the excessive emphasis given
to the synthetic applications of these oxidations1 and of
the potential accessibility of the simplest member of the
series, dimethyldioxirane (DMD), from acetone and po-
tassium monoperoxysulfate (tradename Curox or Caroate).
Actually, DMD is an expensive reagent because it is
obtained in about 5% yield from Caroate2 (78 g of Caroate
is consumed to prepare 1 g of DMD); moreover, it is
unstable, due to a fast radical chain decomposition which
can be initiated at room temperature by traces of oc-
casional impurities, and most oxidations carried out with
DMD can be achieved by simpler and cheaper oxidants.
This dualism (concerted oxygen insertion versus radi-
cal mechanism) is not limited to DMD, but it is quite
general (aromatic peracids,6 perfluorooxaziridines,7 Gif
reaction,8 cytochrome P450,9 and metalloporphyrin ca-
talysis and other metal salt complexes10 ). We have
recently reported evidence that also with peracids,11 Gif
reaction,12 and metalloporphyrin catalysis13 the oxidation
of alkanes can be explained by radical mechanisms.
In a recent report14 numerous authors have criticized
our mechanistic conclusions by stating that “the epoxi-
dation and oxygen insertion into alkane C-H bonds by
DMD do not involve detectable radical pathways”; more-
over, these authors reported that some of our experimen-
tal results could not be reproduced.
Several classes of organic compounds have been oxi-
dized by DMD or similar dioxiranes and some specific
behaviors are strictly related to the oxidation mechanism.
In preliminary papers we3 have reported evidence that
the oxidation of a variety of organic compounds (alkanes,
alcohols, ethers, aldehydes, alkenes, and alkyl iodides)
by DMD can be explained by radical mechanisms. These
reports were in clear contrast with the widely accepted
mechanism1 of “concerted oxenoid oxygen insertion”,
which postulates a butterfly type transition state (struc-
ture 1) for the oxidation of alkenes, similar to the one
originally suggested by Bartlett4 for the alkene epoxida-
tion by peracids.
In this paper we report the developments of our
preliminary data, experimental details, and further
(5) Murray, R. W.; Gu, H. J . Org. Chem.. 1995, 60, 5673.
(6) Schneider, H. J .; Mu¨ller, W. J . J . Org. Chem. 1985, 50, 4609.
(7) Arnone, A.; Cavicchioli, M.; Montanari, V.; Resnati, G. J . Org.
Chem. 1994, 59, 5511.
(8) Barton, D. H. R.; Doller, D. Acc. Chem. Res 1992, 25, 504.
(9) Cytochrome P450: Structure, Mechanism and Biochemistry;
Ortiz de Montellano, P. R., Ed., Plenum Press: New York, 1986.
(10) Metalloporphyrin catalyzed oxidations; Montanari, F., Casella,
L., Eds., Kluwer Academic Publishers: Dordrecht, 1994.
(11) Bravo, A.; Bjørsvik, H. R.; Fontana, F.; Minisci, F.; Serri, A. J .
Org. Chem. 1996, 61, 9409.
(12) Minisci, F.; Fontana, F.; Araneo, S.; Recupero, F.; Zhao, L.
Synlett 1996, 119.
(13) Minisci, F.; Fontana, F.; Araneo, S.; Recupero, F.; Banfi, S.;
Quici, S. J . Am. Chem. Soc. 1995, 117, 226.
X Abstract published in Advance ACS Abstracts, December 15, 1997.
(1) Murray, R. W. Chem. Rev. (Washington, D.C.) 1989, 89, 1187.
Adam, W.; Hadjiarapoglou, L. P. Top.Curr.Chem. 1993, 164, 45 and
references therein.
(2) Adam, W.; Bialas, J .; Hadjiarapoglou, L. P. Chem. Ber. 1991,
124, 2377.
(3) (a) Minisci, F.; Zhao, L.; Fontana, F.; Bravo, A. Tetrahedron Lett.
1995, 36, 1697. (b) Minisci, F.; Zhao, L.; Fontana, F.; Bravo, A.
Tetrahedron Lett. 1995, 36, 1895. (c) Bravo, A.; Fontana, F.; Fronza,
G.; Mele, A.; Minisci, F. J . Chem. Soc., Chem. Commun. 1995, 1573.
(d) Bravo, A.; Fontana, F.; Fronza, G.; Minisci, F.; Serri, A. Tetrahedron
Lett. 1995, 36, 6945.
(14) Adam, W.; Curci, R.; D′Accolti, L.; Dinoi, A.; M.; Fusco, C.;
Gasparrini, F.; Kluge, R.; Paredes, R.; Schulz, M.; Smerz, A. K.; Veloza,
L. A.; Weinko¨tz, S.; Winde, R. Chem. Eur. J . 1997, 105.
(4) Bartlett, P. D. Rec. Chim. Prog. 1950, 11, 47.
S0022-3263(97)01226-7 CCC: $15.00 © 1998 American Chemical Society
Published on Web 01/03/1998

Mechanism for Oxidation by Dimethyldioxirane
J . Org. Chem., Vol. 63, No. 2, 1998 255
Ta ble 1. Rea ction of Ad a m a n ta n e w ith DMD in th e
Ta ble 2. Oxid a tion of Cycloh exa n e w ith DMD in th e
a
a
P r esen ce of CBr Cl3
P r esen ce of CBr Cl3
CBrCl3
(mmol)
conversion
(%)b
1
(%)
2
(%)
3
(%)
4
(%)
5
(%)
6
(%)
CBrCl3 (mmol)
conversion (%)b
7 (%)
8 (%)
9 (%)
0.1
0.2
0.4
0.6
0.8
1.0
2.0
4.0
30.3
32.1
46.8
52.2
58.3
63.2
65.4
70.7
35.1
47.7
58.4
65.6
71.5
79.9
89.1
94.7
24.4
23.2
20.3
18.8
16.3
13.1
7.7
40.6
29.1
21.3
15.6
12.2
7.0
0.1
0.2
0.4
0.6
0.8
1.0
2.0
57
63
68
72
75
76
80
12.2
23.1
37.4
42.1
46.4
51.1
56.8
2.3
4.3
6.8
7.6
8.5
0.7
1.4
2.2
2.4
2.7
3.0
3.3
77.2
62.6
53.1
46.2
38.3
34.0
24.1
1.1
0.8
0.8
0.6
0.5
0.4
0.3
1.8
1.2
0.9
0.5
0.4
0.3
0.1
9.1
13.7
3.2
traces
5.2
a
a
1 ) 1-bromoadamantane, 2 ) 2-bromoadamantane, 3 )
1-chloroadamantane, 4 ) 1-adamantanol, 5 ) 2-adamantanol, 6
7 ) cyclohexyl bromide, 8 ) cyclohexanol, 9 ) cyclohexanone.
b
Conversions based on DMD (cyclohexane 4 mmol, DMD 1 mmol)
b
) 2-adamantanone. Conversion of adamantane based on DMD
(4 mmol of adamantane are reacted with 1 mmol of DMD).
cyclohexyl and adamantyl radicals are involved in the
halogenation: cyclohexyl and 2-adamantyl radicals only
abstract bromine atoms (eq 2), whereas 1-adamantyl
radical abstracts both bromine and chlorine atoms (eq
3).16,17
results, which, in our opinion, strongly support the
radical character of several oxidations by DMD.
Resu lts a n d Discu ssion
Oxid a tion of Alk a n es. The oxidation of unactivated
C-H bonds is a subject of great interest both from the
theoretical and the applicative standpoint. A variety of
oxidants and catalytic processes have been utilized.6-10
As concerns the oxidation by DMD, the overall reaction
eq 1 is very exothermic because the energies of the newly
formed bonds are considerably higher than those of the
broken ones.
•CCl3 or •CBrCl2 can induce several radical chain
processes3a,18
The energy of the O-O bond in DMD is very low, due
to ring strain: it has been evaluated15 at about 10 kcal
mol-1. Despite the large exothermicity, the reaction is
highly regio-, chemo-, and stereoselective.1 These and
other mechanistic features (kinetics, isotope effect) have
induced most authors to exclude the intermediate forma-
tion of radicals. The high sensitivity to polar effects in
the oxidation of alkanes by DMD and by aromatic
peracids (F ) -2.76 in the oxidation of substituted
cumenes by DMD5 and F* ) -2.2 in the oxidation of
substituted alkanes by aromatic peracids16 ), in addition
to a high degree of configurational retention, would
support a concerted electrophilic mechanism for both
oxidants (structures 2 and 3). We have recently reported
evidence pointing out that the oxidation of alkanes by
aromatic peracids is a radical process and that the high
regio-, chemo-, and stereoselectivity must be ascribed to
enthalpic, polar, and cage effects.11
The fact that in the absence of DMD no reaction takes
place under the same conditions clearly indicates that
DMD induces the radical halogenation of alkanes. An
explanation, recently suggested19 for our results, involves
an electron-transfer process between DMD and CBrCl3
(eq 9).
Our preliminary results,3 which are now more com-
pletely developed and reported in detail, supporting a
“radical oxygen rebound” mechanism3a in the oxidation
of alkanes by DMD were based on the following reactions.
(a) When the oxidation of alkanes by DMD was carried
out at room temperature in the presence of variable
amounts of CBrCl3, halogenation competes with oxidation
of the alkane,3a the halogenation increasing with the
concentration of CBrCl3 (Tables 1 and 2). The fact that
with cyclohexane and adamantane the halogenated com-
pounds are cyclohexyl bromide and 1-bromo, 1-chloro-,
and 2-bromoadamantane indicate beyond any doubt that
This explanation is really surprising: the electron-
transfer from a strongly electron-deficient substrate with
high oxidation potential, such as CBrCl3, to DMD appears
extremely unlikely.
(b) When the oxidation of cyclohexane or adamantane
by DMD was carried out in the presence of quinaldine
or lepidine and CF3COOH, quinaldine N-oxide and lepi-
dine N-oxide were the main reaction products, alkane
oxidation was significant, and small amounts of quino-
(17) Tabushi, T.; Aoyama, Y.; Kojo, S.; Hamuro, J .; Yoshida, Z. J .
Am. Chem. Soc. 1972, 94, 1177.
(15) Harding, L. B., Goddard, W. A., III J . Am. Chem. Soc. 1978,
100, 7180.
(16) Minisci, F.; Fontana, F.; Zhao, L.; Banfi, S.; Quici, S. Tetrahe-
dron Lett. 1994, 35, 8033.
(18) Vanni, R.; Garden, S. J .; Banks, J . T.; Ingold, K. U. Tetrahedron
Lett. 1995, 36, 7999.
(19) Curci, R.; Dinoi, A.; Fusco, C.; Lillo, M. A. Tetrahedron Lett.
1996, 37, 249.

256 J . Org. Chem., Vol. 63, No. 2, 1998
Bravo et al.
Ta ble 3. Oxid a tion of Alk a n es u n d er Oxygen (P r oced u r e
A) or u n d er Ar gon (P r oced u r e B) Atm osp h er e
(1) Relevant amounts of ROCOMe were formed in
addition to ROH, which is formed in eq 1, whereas CH4,
MeOH, MeCOOMe, and MeCOOCH2COMe were all
formed from DMD.
convn
ROH (%) +
ketone (%)
alkane (RH) (%)a
ROCOMe (%) proc
(2) The hydroxylation of the alkane is much more
selective than its acetoxylation; thus, with adamantane
the ratio 1-AdOH/2-AdOH is >50, whereas the ratio
1-AdOCOMe/2-AdOCOMe is about 4.
(3) The conversions of alkanes are significantly lower
in the absence of oxygen.
cyclohexane
cyclohexane
34.4 cyclohexanol 2.7
cyclohexanone 72.3
18.7 cyclohexanol 3.1
cyclohexanone 53.6
41.1 ROH 91.7
traces
A
B
35.3
cumene
cumene
ethylbenzene 57
traces
86.3
traces
A
B
A
12.6 ROH 9.6
ROH 2.4
Ph-CO-Me 74.8
ethylbenzene 18.2 ROH 2.2
PhCOMe 53.6
The formation of CH4, MeOH, MeCOOMe, and Me-
COOCH2COMe in the absence of oxygen can be reason-
ably explained only by radical chains, which involve Me•,
16.4
B
A
B
adamantane 63
1-AdOH 96.3
2-AdOH 1.7
1-AdOH 52.4
2-AdOH 0.8
traces
•
R•, and CH2COMe radicals according to eqs 12-16.
adamantane 38
1-MeCOOAd 37.0
2-MeCOOAd 8.2
a
Conversion of the alkane.
lines, substituted in the 2- and 4-positions by methyl,
cyclohexyl, and adamantyl groups, were also obtained.3b
The only reasonable explanation for the formation of
these compounds involves the reaction of cyclohexyl,
adamantyl, or methyl radicals with protonated quino-
lines20 (eqs 10 and 11).
We can evaluate a rate constant of 2.7 × 103 M-1 s-1
for eq 15 from the known absolute rate constant of
hydrogen abstraction from EtCOEt21 (7.4 × 104 M-1 s-1
)
by Me• and from the relative rates of hydrogen abstrac-
tion MeCOMe:EtCOEt (1:27)22 always by Me•. At low
conversion of DMD it was therefore possible to ap-
proximately evaluate the relative rates of eqs 12 and 15
from the amounts of MeOH + MeCOOMe generated by
the induced homolysis of DMD by Me• (eqs 12 and 13)
and from the amount of MeCOOCH2COMe generated by
hydrogen abstraction from acetone by Me• (eqs 15 and
16). Reaction 12 appears to be >500 times faster than
reaction 15; thus, the induced homolysis of DMD by Me•
is very fast, >106 M-1 s-1
. If we consider that the
â-scission of alkoxyl radicals23,24 (for cumyloxyl radical,
rate constants ranging from 2.6 × 105 to 1.9 × 106 s-1 at
30 °C has been reported24 for the â-scission in several
solvents) and the hydrogen abstraction by alkoxyl radi-
cals from C-H bonds are also fast processes (for cyclo-
hexane a rate constant of 1.0 × 106 M-1 s-1 has been
reported),24 the kinetic length of the radical chain of eqs
12-16 must be rather high. This means that a large
quantity of DMD is consumed by small amounts of chain-
initiating radicals.
(c) The presence of oxygen has a dramatic influence
on the oxidation of alkanes by DMD,3c as the results of
Table 3 indicate. The reproducibility of these results is
strictly related to a careful elimination of even the
smallest traces of oxygen. The absence of oxygen has
several consequences never before described in the
numerous reports1 on alkane oxidation by DMD:
(21) Gilbert, B. C.; Norman, R. O. C.; Placucci, G.; Sealy, R. C. J .
Chem. Soc. Perkin Trans.2, 1975, 885.
(20) Minisci, F. Top. Curr. Chem. 1976, 62, 1; Minisci, F. Substituent
Effects in Free-Radical Chemistry; Viehe, H., Ed.; D. Reidel: Dordrecht,
1986; p 391. Minisci, F.; Fontana, F.; Vismara, E. Heterocycles 1989,
28, 489. Minisci, F.; Fontana, F.; Vismara, E. J . Heterocycl. Chem.
1990, 27, 79.
(22) Singh, M.; Murray, R. W. J . Org. Chem. 1992, 57, 4263.
(23) Howard, J . A.; Scaiano, J . C. Landolt-Bo¨rnstein; 1984; Vol. 13,
p 19.
(24) Avila, D. V.; Brown, C. E.; Ingold, K. U.; Lusztyk, J . J . Am.
Chem. Soc. 1993, 115, 466.

Mechanism for Oxidation by Dimethyldioxirane
J . Org. Chem., Vol. 63, No. 2, 1998 257
Sch em e 1
All these results (a-c) show beyond any doubt that
radicals are involved in the oxidation of alkanes by DMD.
The crucial mechanistic problem concerns the ques-
tion: is the oxidation of alkanes by DMD strictly related
to the radical chains described by eqs 2-16 or are these
radical reactions competitive chain processes indepen-
dently initiated by DMD? In other words, either the
initial reaction of the alkane with DMD leads to radicals,
which on one hand determine the alkane oxidation and
on the other hand initiate the radical chains of eqs 2-16,
or there are two completely different independent mech-
anisms, a “concerted oxenoid oxygen insertion”, which
does not involve radicals, and the radical chains of eqs
2-16, induced by DMD and due to accidental impurities
or to reactions such as eq 9, as has been recently
suggested.19
The old concept of molecule-induced homolysis was
thoroughly developed by one of the fathers of radical
chemistry, Walling,26 in his works concerning the reac-
tions of peroxides and alkyl hypochlorites with alkenes,
alkynes, and electron-rich substrates, such as amines.
As far as we know, the molecule-induced homolysis of
peroxides by alkanes was never previously observed or
suggested. We believe that this induced homolysis is
reasonable on the basis of thermochemical evaluations:
two bonds, R-H in the alkane and O-O in DMD, are
broken and a strong O-H bond is formed; for tertiary
C-H bonds the process is almost thermoneutral. With
other peroxides, which have energies >30 kcal mol-1 for
the O-O bonds, the molecule-induced homolysis is too
endothermic to take place; the induction, on the opposite,
is possible with alkenes, due to the much lower energy
of the π bond. The transition state for the induced
homolysis by alkanes is shown by structure 4 (eq 20).
Our interpretation involves an “induced homolysis” of
DMD by the alkane3c,d with formation of a radical pair,
whose coupling in the solvent cage leads to the oxidation
products; radicals escaping from the cage can initiate the
radical chains of eqs 2-16 (Scheme 1).
We have emphasized3c the fact that only a few radicals
escape from the solvent cage, because the radical pair is
necessarily generated in a singlet state and small amounts
of air oxygen inhibit the radical chains; the kinetic
lengths of these chains in the absence of oxygen are,
however, high enough to allow a considerable amount of
DMD to react.
The inhibition by oxygen is due to the formation of
peroxyl radicals3c (eq 17), which, under the reaction
conditions, are unable to effectively sustain the chain,
because the hydrogen abstraction from the alkane (eq 18)
(10-3-10-2 M-1 s-1) is too slow at room temperature.3c
The fate of peroxyl radicals is then a bimolecular self-
reaction, as in classical autoxidation processes25 (eq 19).
This transition state does not involve a “concerted
oxygen insertion”, but it has a clear-cut radical character
with some charge separation (as for the classical polar
effect in radical reactions). This may well explain the
high regio- and chemoselectivity: the marked discrimi-
nation among primary, secondary, and tertiary C-H
bonds depends on the fact that hydrogen abstraction is
largely endothermic in the first two cases; on the other
hand, the high sensitivity to the electron availability of
the C-H bond is related to the polar effect. The high
regioselectivity in the hydroxylation of adamantane (1-
Ad-OH/2-Ad-OH > 50) and the much lower selectivity
in acetoxylation (1-Ad-OCOMe/2-Ad-OCOMe ∼ 4) clearly
indicate that hydrogen abstraction from Ad-H takes place
by two different mechanisms: the slow induced homolysis
for the alcohol (Scheme 1) and mainly the fast eq 13 for
the ester.
The high degree of configurational retention is due to
the fast coupling of the singlet radical pair in the cage.
Adam and Curci14,19 have repeatedly stated that we
interpret the alkane oxidation by DMD by a radical chain
mechanism, quoting ref 3c; this is completely incorrect.
In ref 3c we clearly report that in the presence of air
oxygen only the coupling of the radical pair in the cage
takes place. In the absence of oxygen, radical chain
processes occur, but they concern different reactions (eqs
2-16) and form different reaction products.
After the publication of our preliminary results, an
attempt has been reported14,19 to reconcile the “concerted”
and radical mechanisms (Scheme 2).
The explanation given to Scheme 2 is, in our opinion,
contradictory and ambiguous; the reaction is considered
(25) Walling, C. Active Oxygen in Chemistry; Foote, C. S., Valentine,
J . S., Greenberg, A., Liebman, J . F., Eds.; Blackie Academic Profes-
sional: New York, 1995; p 24.
(26) Walling, C.; Chang, Y. W. J . Am. Chem. Soc. 1954, 76, 4878;
Walling, C.; Heaton, L.; Tanner, D. D. J . Am. Chem. Soc. 1965, 87,
1715.

258 J . Org. Chem., Vol. 63, No. 2, 1998
Bravo et al.
Sch em e 2
“an essentially concerted oxenoid mechanism of insertion
(I); in the rate-determining step, this should present no
distinct radical nor carbenium ion character, but after
initiate the radical chains of eqs 12-16, which are broken
by the presence of a small amount of oxygen.
Oxid a tion of Eth er s a n d Alcoh ols. Dialkyl ethers
are very easily oxidized by DMD in the R-position.27 It
has been reported that “the high selectivity for the
observed R-functionalization speaks for a nonradical C-H
insertion. In fact, also some â-attack would be expected
in a classical pathway involving RO• radicals; for in-
stance, approximately 13% attack at â-CH2 was found
for the reaction of THF with HO• ”.27 This statement is
really surprising, because it is well-known that oxygen-
centered radicals selectively attack the R-position of alkyl
ethers, with the exception of HO• radical, which is an
extremely reactive and unselective species, mainly for
enthalpic reasons. The facile R-autoxidation of ethers is
notorious because of the production of hazardous perox-
ides.28
Again, when we carried out the oxidation of diethyl
ether or of THF by DMD in the presence of protonated
quinaldine, N-oxidation was the main reaction but small
amounts of methyl, ethyl and R-tetrahydrofuranyl radi-
cals were trapped by the heterocyclic ring. Methyl
radical is generated from DMD by reactions similar to
eqs 12,14, 16; R-tetrahydrofuranyl radical is originated
by hydrogen abstraction from THF and ethyl radical
comes from diethyl ether29 according to eq 21.
the transition state some radical character develops (II)”
19
The contradiction concerns the fact that a “concerted
mechanism of insertion” does not involve, by definition,
any intermediate; moreover, a transition state with no
charge separation cannot explain the high influence of
polar effects. If II is a radical pair, we do not understand
why it should have only some radical character (a radical
is a well-defined species, characterized by an unpaired
electron) and why the coupling of the radical pair should
fail to occur, while a hydroxyl transfer takes place. If II
is not a radical pair, it becomes to us a mysterious
species, as one transition state cannot, by definition, give
different reactions. Structure II would then give either
hydroxyl transfer or a radical leakage, competitively. Yet,
it is also stated that “the oxygen insertion into alkanes
by DMD does not involve detectable radical pathways”;14
in this case we do not understand the meaning of “radical
leakage”. Besides, the collapse of the concerted transition
state I into products is much more favorable on enthalpic
grounds than would be its evolution to a radical pair. It
appears to us that this attempt to reconcile concerted and
radical mechanisms does not work, as the two mecha-
nisms are alternative. To verify if the radical chains of
eqs 12-16 are really initiated by radicals, escaped from
the cage according to Scheme 1, and not by accidental
initiation, we have carefully investigated the decomposi-
tion of DMD. We conducted several experiments, both
under argon and under oxygen, in the presence as well
as in the absence of adamantane, at room temperature
(20 °C), by always using the same solution of DMD, to
avoid the possibility that radical-chain-initiating impuri-
ties could affect the results.
In the absence of adamantane, DMD is not signifi-
cantly decomposed after 6 h, either under argon or under
oxygen. In the presence of adamantane under argon 76%
of DMD reacts in 6 h and the reaction products are
1-adamantanol (27%), traces of 2-adamantanol (<1%),
1-acetoxyadamantane (18.8%), and 2-acetoxyadamantane
(4.2%); the balance of the reacted DMD is completed by
CH4, MeOH, MeCOOMe, and MeCOOCH2COMe. In the
presence of adamantane under oxygen, 71% of DMD
reacts in 6 h and the reaction product is mostly 1-ada-
mantanol (>90%), with small amounts of 2-adamantanol
and 1- and 2-acetoxyadamantane.
Since 1- and 2-acetoxyadamantane, CH4, MeOH,
MeCOOMe, and MeCOOCH2COMe (on the whole 65%
of the reacted DMD) are certainly formed under argon
by a radical mechanism (eqs 12-16), while in the absence
of adamantane under the same conditions no reaction
occurs, the only reasonable explanation of these results
(Occam’s razor) involves the induced homolysis of DMD
by adamantane. The few radicals escaping the cage
Also in this case, when the reaction was performed
under argon in the presence of diisopropyl ether or
cyclohexanol, the main reaction products (CH4, MeOH,
MeCOOMe, and MeCOOCH2COMe) came from the radi-
cal decomposition of DMD; if diethyl ketone was added
to the reaction mixture, the R-acetoxy derivative (EtCO-
CH(Me)OCOMe) was a significant reaction product (Et-
COEt is 27 times more reactive than acetone toward
methyl radical).22 Since in the absence of diisopropyl
ether or of cyclohexanol the same solution, under the
same conditions, does not give decomposition of DMD,
(27) Curci, R.; D′Accolti, L.; Fiorentino, M.; Fusco, C.; Adam, W.;
Gonzalez-Nunez, M. E.; Mello, R. Tetrahedron Lett. 1992, 33, 4225.
(28) ref 25 pag.29.
(29) Minisci, F.; Coppa, F.; Fontana, F.; Zhao, L. Gazz. Chim. Ital.
1993, 123, 613.

Mechanism for Oxidation by Dimethyldioxirane
J . Org. Chem., Vol. 63, No. 2, 1998 259
the reasonable conclusion is that diisopropyl ether and
cyclohexanol induce the homolysis of DMD. Because of
enthalpic and polar effects, ethers and alcohols induce
the homolysis of DMD more effectively than do the
corresponding alkanes; e.g. cyclohexanol is about 103
times more reactive than cyclohexane.3a
The radical decomposition of dioxiranes, induced by
alkyl ethers, has also been recently reported by other
authors,30 although they did not suggest a mechanism
to explain the formation of radicals; besides, their
interpretations are not consistent with the reactivity of
the involved species: for instance, the formation of the
acetoxyl derivatives has been explained by cross-coupling
between acetoxyl and alkyl radicals (eq 22).
reaction products when the reaction was carried out in
the presence of oxygen; lepidine N-oxide was formed in
significant amount (7-10%), and minor amounts (1-4%)
of 2,4-dimethyl-, 2-acetyl-4-methyl-, 2-pivaloyl-4-methyl-,
and 2-tert-butyl-4-methylquinoline were also obtained.
Thus methyl, pivaloyl and tert-butyl radicals are certainly
involved in the oxidation. tert-Butyl radical clearly arises
from the unimolecular decarbonylation of pivaloyl radical
(eq 25).
A careful investigation has been carried out for the oxi-
dation of phenylacetaldehyde by DMD, to recognize the
origin of these radicals; again, the same solution of DMD
in acetone has been utilized for all the experiments. In
the presence of oxygen, phenylacetic acid was the only
reaction product from the aldehyde (81% conversion of
DMD, 75% yield of phenylacetic acid in 6 h at 20 °C);
under the same conditions, in the absence of DMD, no
substantial oxidation of phenylacetaldehyde occurs. Un-
der argon atmosphere, at 20 °C for 6 h, DMD was con-
verted for 78%, but the conversion of phenylacetaldehyde
was only 6% (4% gave benzyl acetate and 2% phenylacetic
acid); most of the DMD was converted into CH4, MeOH,
MeCOOMe, and MeCOOCH2COMe. At 60 °C under
argon the conversion of DMD was complete, but phenyl-
acetaldehyde conversion only increased up to 16%, with
formation of benzyl acetate (14%) and phenylacetic acid
(2%). In the absence of phenylacetaldehyde at 20 °C
under argon, no substantial decomposition of the same
solution of DMD occurred. The only reasonable mecha-
nism for the formation of benzyl acetate involves the fast
decarbonylation32 of the acyl radical outside the cage (eq
26)
This cross-coupling, however, can be excluded by the
fact that the Ingold-Fischer “persistent radical effect”31
is not fulfilled (both radicals are transient) and, above
all, by the high rate constant (>109 s-1) of the unimo-
lecular decarboxylation of the acetoxyl radical (eq 23)
The only reasonable mechanism for the formation of
the acetoxyl derivative is the one depicted in eq 14.
The conclusion is that ethers and alcohols are oxidized
by DMD through a mechanism which is identical with
the one operating for alkanes, the only differences being
the contribution of polar effects and the lower bond
strength and the correspondingly lower activation energy
for the reaction of the R-C-H bond in ethers and alcohols.
Oxid a tion of Ald eh yd es. The oxidation of acetalde-
hyde, pivalaldehyde, and phenylacetaldehyde by DMD
in the presence of air oxygen mainly leads to the corre-
sponding carboxylic acids with high selectivity. Polar and
enthalpic factors suggested that the molecule-induced
homolysis could be still easier in these cases (eq 24)
The benzyl radical induces the chain decomposition of
DMD, giving benzyl acetate (eq 27)
These results indicate beyond any doubt that phenyl-
acetaldehyde induces the radical decomposition of DMD
under argon because only a small amount of phenylacetic
acid is obtained, while all the reaction products come
from the radical reaction of DMD.
The presence of a small amount of air oxygen inhibits
the radical chain processes, favoring the formation of
phenylacetic acid, as in the oxidation of alkanes, ethers,
and alcohols. Thus, the oxidation mechanism of all these
classes of organic compounds by DMD is basically identi-
cal, due to the molecule-induced homolysis of DMD by
C-H bonds of relatively low energy and high electron
availability. The relevant fact is that the numerous
authors who investigated the oxidations by DMD did not
realize that the success of these reactions was strictly
The coupling of the radical pair in the cage gives the
carboxylic acid. The induced homolysis is more exother-
mic by 5-6 kcal mol-1 compared to tertiary C-H bonds
in alkanes, due to the lower energy of the C-H bond in
the aldehyde group. Protonated quinaldine and lepidine
have been once again utilized in the oxidation of acetal-
dehyde and pivalaldehyde to intercept radicals outside
the solvent cage. Acetic and pivalic acids were the main
(30) Ferrer, M.; Sanchez-Baeza, F.; Casas, J .; Messeguer, A. Tetra-
hedron Lett. 1994, 35, 2981.
(31) MacFaul, P. A.; Arends, I. W. C. E.; Ingold, K. U.; Wayner, D.
D. M. J . Chem. Soc. Perkin 2 1997, 135. Fischer, H. J . Am. Chem.
Soc. 1986, 108, 3925. Bravo, A.; Bjørsvik, H. R.; Fontana, F.; Liguori,
L.; Minisci, F. J . Org. Chem. 1997, 62, 3849.
(32) Lunazzi, L.; Ingold, K. U.; Scaiano, J . C. J . Phys. Chem. 1983,
87, 528.

260 J . Org. Chem., Vol. 63, No. 2, 1998
Bravo et al.
Sch em e 3
3 would give cyclization in competition with fragmenta-
tion, because steric factors make the fragmentation less
favorable in the case of fullerene diradical compared with
simple alkenes. The fact that small amounts of 1,3-
dioxolane were obtained in the epoxidation of cis-3-
hexene also supports this interpretation because the
cyclization is sterically more favorable for the cis than
for the trans isomer.
An interesting point concerns the behavior of 1,1-
diphenyl-2-vinylcyclopropane toward DMD;14 the fact
that the epoxide is the only reported reaction product
would indicate that either an induced homolysis does not
occur or the fragmentation of the diradical is faster than
the cyclopropylcarbinyl rearrangement. Certainly the
fragmentation of the diradical is more exothermic than
the rearrangement, and this could explain both the lack
of rearrangement and the high stereoselectivity. How-
ever, a question arises from the very high rate of
fragmentation of the diradical: does it make any sense
to distinguish between “induced homolysis” and “con-
certed oxygen insertion” with alkenes, considering the
very short lifetime of the possible diradical in Scheme
3? We believe that the distinction still makes sense, even
if the detection of the radical intermediate is quite a
problem and the evidence can only be circumstantial.
Actually, we consider the formation of 1,3-dioxolane,
observed in some cases33,34 in competition with the
formation of the epoxide, as a significant, although
indirect, proof of a radical mechanism. If the epoxidation
mechanism is the same by DMD and by m-chloroperben-
zoic acid (m-CPBA), namely an electrophilic concerted
mechanism, we should expect, at least qualitatively, the
same behavior with both reagents. To verify this aspect
we have investigated the relative reactivities for the
oxidation of cyclohexene and quinoline: cyclohexene
epoxide and quinoline N-oxide were the only reaction
products in separate experiments with both oxidants.
However, in the competitive oxidation of equimolar
amounts of cyclohexene and quinoline, only cyclohexene
epoxide was formed by DMD, while with m-CPBA cyclo-
hexene epoxide and quinoline N-oxide were formed in a
13:87 ratio, despite the reduced reactivity of quinoline,
due to the broad extent of protonation by the peracid.
Quinoline is oxidized by an electrophilic process and it
appears unlikely that such a dramatic inversion of
reactivity should occur if the epoxidation of cyclohexene
by DMD would also take place by the same mechanism.
This result, though, can be explained by assuming that
m-CPBA acts as an electrophilic reagent toward both
substrates, while DMD oxidizes cyclohexene through a
fast induced homolysis and quinoline by a much slower
electrophilic process.
related to the presence of a small amount of air oxygen
in the reaction medium.
Oxid a tion of Alk en es. The molecule-induced ho-
molysis of peroxides and other electron-deficient reagents
with weak bonds (e.g. O-Cl, O-Br) by alkenes has been
thoroughly investigated.26 The very weak peroxidic bond
in DMD (about 10 kcal mol-1) suggested that the induced
homolysis should be much easier compared with common
peroxides, i.e., more exothermic by at least 20 kcal mol-1
.
Thus we considered the possibility that also the epoxi-
dation of alkenes by DMD could be explained by the
molecule-induced homolysis according to Scheme 3.
On the other hand, if an induced homolysis of DMD
occurs with alkanes, ethers, alcohols, and aldehydes, the
process should be extremely more favored with alkenes,
due to the weak π bond and the high electron availability.
This hypothesis has been recently challenged:14 “Mecha-
nistically more relevant is the fact that, were a radical
DMD epoxidation to apply, (Scheme 3), cycloadducts
should be formed, since it is well established that a
diradical intermediate of this type would preferentially
cyclize rather than undergo fragmentation.33 The cy-
clization would have essentially no activation energy,
whereas as much as 10-15 kcal mol-1 would be required
for the fragmentation, because a relatively strong C-O
bond is broken and a strained product (epoxide) formed”.14
Now this thermochemical evaluation is completely er-
roneous and the report of ref 33 is quite incorrect. The
authors clearly mistake the C-O bond energy of a
molecule for that of the same bond in the â-position to a
radical; this is an elementary concept in radical chem-
istry. Actually, the C-O bond, broken in Scheme 3, is
an extremely weak bond (<20 kcal mol-1) and the C-O
bond formed (epoxide) is >60 kcal mol-1
. Thus the
fragmentation of the diradical in Scheme 3, with forma-
tion of the epoxide and of acetone, is exothermic by at
least 40 kcal mol-1; this means, for a radical â-scission,
no activation enthalpy.
Moreover, the cyclization of the diradical of Scheme 3
leading to the 1,3-dioxolane is, obviously, largely exo-
thermic, but entropic factors would favor the fragmenta-
tion. Reference 33 only reports that the oxidation of
fullerene-C60 by DMD leads to both the epoxide and the
1,3-dioxolane in a 1:1.5 ratio; the comments of the
authors in ref 33 are “the determination of the mecha-
nism of the formation of the epoxide and the dioxolane
requires further studies”. If we consider that small
amounts of dioxolane were also formed during the
epoxidation of simple alkenes, such as cis-3-hexene,34 we
believe that the results with fullerene33 represent, on the
opposite, a significant evidence of the induced homolysis
also for epoxidation: the same diradical adduct of Scheme
Further circumstantial evidence for a radical mecha-
nism was provided by the results obtained with R-meth-
ylstyrene under nitrogen: the epoxide was obtained in
86% yield. The careful GC-MS analysis revealed the
presence of minor amounts of 2-phenylpropenol and of
2-phenylpropenal. During the GC-MS analysis, part of
the epoxide rearranges to R-phenylpropionaldehyde: this
thermal rearrangement is well-known for styrene ep-
oxide. We explain these results by a fast but reversible
induced homolysis of DMD by the alkene, with which
allylic hydrogen abstraction can compete to a small
extent (Scheme 4).
(33) Elemes, Y.; Silverman, S. K.; Sheu, C.; Kao, M.; Foote, C. S.;
Alvarez, M. M.; Whetten, R. L. Angew. Chem., Int. Ed. Engl. 1992,
31, 351.
(34) Murray, R. W.; Ramachandran, V. Photochem. Photobiol. 1979,
30, 187.
It was recently reported14 that the oxidation of R-meth-
ylstyrene by DMD exclusively gave the epoxide and that

Mechanism for Oxidation by Dimethyldioxirane
J . Org. Chem., Vol. 63, No. 2, 1998 261
Sch em e 4
attempts to detect minor amounts of 2-phenylpropenol
and 2-phenylpropenal failed. Our preliminary results
were reported without experimental details, which are
now described in the Experimental section. On the other
hand, our result is not unprecedented: the peroxyacid
epoxidation of 4,4-dimethyl-D5 steroids is straightfor-
ward,35 while on the contrary, by using DMD the double
bond is not attacked and only allylic oxidation occurs.
Oxidation of Iodo Der ivatives. Aryl iodides smoothly
react with DMD at 0-20 °C in acetone to give mainly
ArIO or ArIO2, depending on the amount of DMD; in the
presence of AcOH the corresponding iodosoacetate, ArI-
(OAc)2, is formed.3d
Cyclohexyl iodide mainly gives trans-2-iodocyclohex-
anol, while in the presence of AcOH trans-2-iodocyclo-
hexyl acetate is formed;3d this is due to the instability of
the alkyliodoso derivatives: they eliminate IOH, whose
trans addition to cyclohexene gives the reaction products
(eq 28)
circumstantial, as for the reaction with alkenes. Com-
petitive reaction with iodobenzene and quinoline were
carried out in acetone solution by DMD and by m-CPBA;
both reagents oxidize the substrates to iodosobenzene and
to quinoline N-oxide, respectively. However, in competi-
tive experiments the reactivity is totally opposite, in that
DMD exclusively oxidizes iodobenzene, while m-CPBA
only attacks quinoline. The result is well-explained if
we assume, as in the case of alkenes, that iodobenzene
induces the homolysis of DMD according to eq 28.
Con clu sion s
The molecule-induced homolysis of DMD by alkanes,
ethers, alcohols, and aldehydes, through hydrogen ab-
straction from relatively weak C-H bonds of high
electron availability, is considered responsible for the
oxidation of these classes of compounds. The oxidation
occurs by cross-coupling of the radical pair in the solvent
cage, while the few radicals escaping from the cage can
initiate radical chains involving DMD. The effect of the
presence of oxygen, CBrCl3, or protonated quinoline
supports this interpretation. It is suggested that an
induced homolysis could occur also for the epoxidation
of alkenes and for the oxidation of iodo derivatives, but
in these cases the evidence is only circumstantial.
Exp er im en ta l Section
Gen er a l Meth od s. All the solvents and reagents were
obtained from commercial sources and were used without
further purification. The acetone solutions of DMD were
prepared according to the literature procedure,2 and the
concentration of DMD was evaluated by iodometric titration.
All the reaction products were commercially available or were
previously prepared by different procedures in our laboratory,
as was the case for the quinoline derivatives;20 they were
utilized for the qualitative identification (GC-MS) and the
quantitative analysis (GC and NMR).
The electrophilic oxidation of iodo derivatives by
peroxides is well-documented.36 We have, however,
considered the possibility that a radical mechanism could
be operating with DMD (eq 28), on the grounds that iodo
derivatives appear to react very rapidly with oxygen- and
carbon-centered radicals37 (106-109 M-1 s-1 at rt).
Gen er a l P r oced u r e for th e Oxid a tion of Ad a m a n ta n e
a n d Cycloh exa n e by DMD in th e P r esen ce of CBr Cl3. The
hydrocarbon (4 mmol) was added at 20 °C to a solution of 1
mmol of DMD in 10.4 mL of acetone, prepared according to
the known procedure,2 and in which variable amounts of
CBrCl3 were dissolved, as reported in Tables 1 and 2. After 2
h, the solution was analyzed by GC and GC-MS. With
adamantane six products were formed: 1-bromoadamantane
(1), 2-bromoadamantane (2), 1-chloroadamantane (3), 1-ada-
mantanol (4), 2-adamantanol (5), and 2-adamantanone (6). The
products were identified by comparison with authentic com-
mercial samples. The results are reported in Table 1. With
cyclohexane the reaction products were cyclohexyl bromide (7),
cyclohexanol (8), and cyclohexanone (9); the results are
reported in Table 2.
The fragmentation of the diradical adduct is largely
exothermic and it is almost certainly extremely fast; thus,
also in this case, mechanistic evidence could only be
(35) Marples, B. A.; Muxworthy, J . P.; Baggaley, K. H. Tetrahedron
Lett. 1991, 32, 533.
(36) Davidson, R. I.; Kropp, P. J . J . Org. Chem. 1982, 47, 1904 and
references therein.
(37) Minisci, F. In Sulfur-Centered Reactive Intermediates in Chem-
istry and Biology: Chatgilialoglu, C., Asmus, K. D., Eds.; NATO ASI
Series A; Plenum Press: New York, 1990; Vol. 197, p 303 and
references therein; Baciocchi, E.; d'Acunzo, F.; Galli, C.; Ioele, M. J .
Chem. Soc. Chem. Commun. 1995, 429.
In the absence of DMD under the same conditions no
reaction took place.

262 J . Org. Chem., Vol. 63, No. 2, 1998
Bravo et al.
Oxid a tion of Ad a m a n ta n e a n d Cycloh exa n e by DMD
in th e P r esen ce of P r oton a ted Qu in a ld in e or Lep id in e.
The hydrocarbon (4 mmol), the heteroaromatic base (1 mmol),
and CF3COOH (1 mmol) were added to a solution of 1 mmol
of DMD in 11.2 mL of acetone. The reaction was carried out
at 0 °C for 8 h or at 50 °C for 1 h. The solution was analyzed
by GC and GC-MS. With cyclohexane in the presence of
quinaldine five products were formed, in the amounts reported
in brackets respectively for 0 and 50 °C: cyclohexanol (4.3 and
6.4%), cyclohexanone (1.8 and 3.8%), quinaldine N-oxide (28.5
and 22.8%), 2,4-dimethylquinoline (3.6 and 2.4%), and 2-meth-
yl-4-cyclohexylquinoline (0.6 and 2.2%). With adamantane in
the presence of lepidine, lepidine N-oxide was the main
reaction product at 50 °C (38.2%), 1-adamantanol was a
significant reaction product (17.8%), and a small amount
(1.8%) of 2-(1-adamantyl)-4-methylquinoline was also formed.
The reaction products were identified by comparison with
authentic commercial samples, with the exception of 4-cyclo-
hexyl-2-methylquinoline and 2-(1-adamantyl)-4-methylquino-
line, previously prepared by known procedures.20
Oxid a tion of Alk a n es by DMD in th e P r esen ce a n d in
th e Absen ce of Oxygen . Gen er a l P r oced u r e A. A solution
of 1 mmol of DMD in 10.8 mL of acetone was purged with
oxygen for 10 min and then kept under oxygen for the duration
of the experiment. Alkane (1 mmol) was added and the
solution was kept at 20 °C for 8 h. The solution was directly
analyzed by GC and GC-MS by using authentic samples for
the identification and the quantitative analysis.
Gen er a l P r oced u r e B. The reaction was carried out as
in method A, but the acetone solution was purged with argon
and kept under argon for the duration of the experiments.
The results are reported in Table 3.
of acetone. After 8 h the solution was analyzed by GC and
GC-MS: quinaldine N-oxide was formed in 12% yield, 2,4-
dimethylquinoline in 1.2% yield, and 2-methyl-4-ethylquinoline
in 3.2% yield. Authentic samples were used for the identifica-
tion and the quantitative analysis. Iodometric titration re-
vealed that 79% of DMD had reacted, leading mainly to
ethanol, acetaldehyde, and acetic acid (a quantitative evalu-
ation was not carried out) as previously reported.30
Similarly, THF gave 23% of quinaldine N-oxide, 3.2% of 2,4-
dimethylquinoline, and 2.8% of 4-(a-tetrahydrofuranyl)-2-
methylquinoline.
In d u ced Hom olysis of DMD by Diisop r op yl Eth er
u n d er Ar gon . A solution of 1 mmol of DMD and 4 mmol of
diisopropyl ether in 11 mL of acetone was purged with argon
for 10 min and then kept under argon for 6 h at 22 °C.
Iodometric titration revealed that 83% of DMD had reacted;
NMR analysis of the acetone solution, carried out as above-
described, has shown the presence of CH4, CH3COOCH3 (38.4%
based on DMD), CH3OH (12.1%), and acetoxyacetone (11.8%).
The same experiment, carried out in the presence of diethyl
ketone (2 mL), gave 9.6% of R-acetoxydiethyl ketone and 2.1%
of acetoxyacetone.
In the absence of diisopropyl ether under the same condi-
tions, no substantial decomposition of DMD occurred.
Oxid a tion of Cycloh exa n ol by DMD u n d er Ar gon a n d
u n d er Oxygen . Meth od A. The reaction was carried out
under argon as for diisopropyl ether. GC and NMR analyses
revealed the formation of cyclohexanone (45% based on DMD),
CH3CO-OCH3 (27.2%), CH3OH (13.5%), acetoxyacetone (12.1%),
and CH4.
Meth od B. As in method A under oxygen, cyclohexanone
was obtained in 87% yield.
Com p lete An a lysis of th e Oxid a tion of Ad a m a n ta n e
by DMD u n d er Ar gon . The reaction was carried out as in
method B with adamantane at 23 °C. After 6 h iodometric
titration revealed that 76% of the DMD had reacted. GC
analysis of the solution revealed the formation of 1-adaman-
tanol (27%), traces of 2-adamantanol (<1%), 1-acetoxyada-
mantane (18.8%), 2-acetoxyadamantane (4.2%), and acetoxy-
acetone (7.2%).
Deuterated acetone (0.2 mL) was added to 0.8 mL of the
reaction solution and the resulting solution was analyzed by
NMR. The deuterated solvent was used to provide a deuter-
ium signal for the instrument lock. For cost reasons the
reaction was carried out in nondeuterated acetone, so the NMR
spectrum is dominated by the solvent signal. Such a very
intense peak precludes the detection of weak signals and thus
it was suppressed using the presaturation technique: a long
soft RF pulse (55-60 dB attenuation) was applied at the
frequency of acetone (2.05 ppm) during the recycle delay
followed by the acquisition of the spectrum. To avoid any
saturation of the signals the spectra were acquired with a long
recycle delay (10-15 s) between pulses. Due to the presatu-
ration RF field, the signals near the region of 2 ppm are
distorted and partially suppressed; thus, in this region they
are difficult to assign and cannot be integrated.
We observe clean singlets due to the presence of methane
(0.15 ppm, CH4), methanol (3.29 ppm, CH3), methyl acetate
(3.58 ppm, OCH3) and acetoxyacetone (4.70 ppm, -OCH2-).
All these products have been unequivocally identified by
addition to the solution of traces of the authentic samples. The
signals have been carefully integrated to determine the
relative quantities of the products in solution. The integration
was done after a very careful baseline correction of the
spectrum and phase adjustment of the signals to minimize
errors. The evaluation of methane, in that it is a gas, can be
only qualitative. In this way we have determined, for methyl
acetate, methanol, and acetoxyacetone, a ratio 3.2:0.8:1 and,
considering that the yield of acetoxyacetone, determined by
GC, was 7.2%, the yields of methanol and methyl acetate were
5.8% and 23.1%, respectively.
Oxid a tion of Aceta ld eh yd e a n d P iva la ld eh yd e by
DMD in th e P r esen ce of P r oton a ted Qu in a ld in e or
Lep id in e. Method A. A 10 mmol portion of acetaldehyde, 1
mmol of quinaldine, and 1 mmol of CF3COOH were dissolved
at 0 °C in a solution of 1 mmol of DMD in 10.8 mL of acetone.
After 8 h the solution was analyzed by GC and GC-MS: acetic
acid was obtained in 77% yield, quinaldine N-oxide in 7% yield,
2,4-dimethylquinoline in 4.1% yield, and 2-methyl-4-acetylquin-
oline in 4.2% yield.
Meth od B. As in method A but using pivalaldehyde and
lepidine instead of acetaldehyde and quinaldine, lepidine
N-oxide (60%) and pivalic acid (26%) were the main reaction
products, but small amounts of 4-methyl-2-tert-butylquinoline
(0.4%) and 2-pivaloyl-4-methylquinoline (1.1%) were also
determined.
Oxid a tion of P h en yla ceta ld eh yd e by DMD u n d er
Oxygen a n d u n d er Ar gon . Meth od A. A 1 mmol sample
of phenylacetaldehyde was dissolved at 20 °C in a solution of
1 mmol of DMD in 11.4 mL of acetone under an oxygen
atmosphere. After 6 h, 76% conversion of phenylacetaldehyde
was observed, with formation of phenylacetic acid in 98%
selectivity.
Meth od B. The reaction was carried out under argon as
in method A. Only 6% of phenylacetaldehyde was converted,
with 67% selectivity in benzyl acetate and 33% in phenylacetic
acid; CH3COOCH3 was the main reaction product (42%), with
minor amounts of CH3OH (9.1%), acetoxyacetone (8.4%), and
CH4. At 60 °C the conversion of phenylacetaldehyde increases
to 16% with 87% selectivity in benzyl acetate and 13% in
phenylacetic acid.
Com p etitive Oxid a tion of Cycloh exen e a n d Qu in olin e
by DMD a n d m -CP BA. By DMD. A 4 mmol portion of
cyclohexene and 4 mmol of quinoline were dissolved in a
solution of DMD (1 mmol) in 10.5 mL of acetone at 18 °C. After
4 h, GC analysis only revealed the formation of cyclohexene
epoxide, without traces of quinoline N-oxide.
By m -CP BA. A 4 mmol portion of cyclohexene and 4 mmol
of quinoline were dissolved in a solution of DMD (1 mmol) in
10.5 mL of acetone at 18 °C. After 4 h, GC analysis revealed
the formation of 13% of cyclohexene epoxide and 87% of
quinoline N-oxide.
Oxid a tion of Dieth yl Eth er a n d THF by DMD in th e
P r esen ce of P r oton a ted Qu in a ld in e. Diethyl ether (10
mmol), quinaldine (1 mmol), and CF3COOH (1 mmol) were
dissolved at 0 °C in a solution of 1 mmol of DMD in 11.2 mL
Oxid a tion of r- a n d â-Meth ylstyr en e by DMD. r-Meth -
ylstyr en e. A 2 mmol sample of R-methylstyrene was dis-

Mechanism for Oxidation by Dimethyldioxirane
J . Org. Chem., Vol. 63, No. 2, 1998 263
Sch em e 5
Sch em e 6
solved at 20 °C in a solution of DMD (1 mmol) in 11 mL of
acetone under nitrogen. After 6 h the GC analysis revealed
the presence of 86% of the epoxide and minor amounts of
byproducts. Careful GC-MS analysis revealed the presence
of small amounts of 2-phenylpropenol (6%) and 2-phenylpro-
penal (5%); the compounds were identified by comparison with
MS of authentic samples. During the GC-MS analysis, 51%
of the epoxide rearranges to R-phenylpropionaldehyde. The
analyses were performed on a GLC-MS Finnigan TSQ70
instrument, using a Varian 3700 gas chromatograph, equipped
with a SBP-1 fused silica column (30 m × 0.2 mm i.d., 0.2 mm
film thickness) and helium as carrier gas. An analytical
experiment with pure R-methylstyrene epoxide has verified
its partial rearrangement to R-phenylpropionaldehyde.
â-Meth ylstyr en e. Under the same conditions, â-methyl-
styrene substantially gave only the epoxide, with a selectivity
>98%.
Com p etitive Oxid a tion of Iod oben zen e a n d Qu in olin e
by DMD a n d m -CP BA. (A) By DMD. A 4 mmol portion of
iodobenzene and 4 mmol of quinoline were dissolved in a
solution of DMD (1 mmol) in 10.5 mL of acetone at 22 °C. After
4 h, GC, GC-MS, and NMR analyses only revealed the
presence of iodosobenzene, without traces of quinoline N-oxide.
(B) By m -CP BA. The reaction was carried out as in
method A with m-CPBA instead of DMD. The analyses
revealed only the presence of quinoline N-oxide, without traces
of iodosobenzene.
ing the bimolecular self-reaction, to give cyclohexene (eq 30)
Reaction 30, however, cannot take place to a significant
extent outside the solvent cage, for obvious kinetic reasons.
All the possible X• radicals are transient and the steady-state
concentration of cyclohexyl radical can only be very low, due
to the high rate constants for the reaction of alkyl radicals
with DMD39 (for Me•, eq 12, we have evaluated a rate constant
>106 M-1 s-1) and with acetone, used as solvent (for Me•, eq
15, we have evaluated a rate constant >103 M-1 s-1). The only
reasonable explanation for cyclohexene formation is, in our
opinion, the competition of the radical pair in the cage between
coupling and disproportionation (Scheme 6).
This kind of competition in bimolecular reactions between
two radicals is well-known in radical chemistry.
The kinetic isotope effect reported by ref 38 is in full
agreement with this explanation: the higher isotopic effect
was observed with cyclohexanone (in this case hydrogen
abstraction takes place by the slow induced homolysis, Scheme
6), while the lower isotopic effect occurred with cyclohexyl
acetate, in which hydrogen abstraction from cyclohexane
mainly takes place by the fast eq 13; cyclohexene epoxide
formation shows an intermediate value of the isotopic effect,
since the first hydrogen abstraction is slow and the second is
extremely fast (Scheme 6). This is also in full agreement with
the regioselectivity of adamantane oxidation, discussed in this
paper: the hydroxylation is highly selective, while the ace-
toxylation is much less so. In the gas phase the oxidation of
cyclohexane by DMD is more complex:38 only very low conver-
sion (2%) was achieved, cyclohexene epoxide and cyclohexyl
acetate were not detected and the isotope effect for cyclohex-
anone formation is much lower. Moreover, 1,1,3-trimethylcy-
clohexane, derived from methyl radical, was formed in sig-
nificant amounts. Clearly the mechanism in the gas phase is
quite different.
Note Ad d ed in P r oof. An interesting paper38 concerning
the mechanism of the oxidation of alkanes by DMD was
published after the submission of this paper.
The oxidation of cyclohexane and perdeuteriocyclohexane
was carried out by DMD in the absence of oxygen in order to
measure the kinetic isotope effect.38 In addition to cyclohex-
anone and cyclohexyl acetate, already reported for the first
time in our preliminary paper,3c a considerable amount (ca.
10% of the reaction products) of cyclohexene epoxide was
obtained. We have verified that a byproduct in the oxidation
of cyclohexane, which we had not previously3c identified in our
experiments, is actually cyclohexene epoxide. To explain their
results, Asensio and co-workers38 suggest two competitive
mechanisms: (i) an electrophilic oxygen insertion for the for-
mation of cyclohexanone (Structure 2 in our paper) and (ii) a
radical mechanism, in which cyclohexyl radical is formed and
then evolves, following two competitive paths, leading to cyclo-
hexyl acetate and cyclohexene epoxide respectively (Scheme
5).
The mechanism for the formation of cyclohexyl acetate is
identical with the one previously reported by us3c (eq 14), but
the mechanism for cyclohexene formation has no kinetic
basis: certainly, cyclohexyl radical cannot spontaneously lose
a hydrogen atom to yield cyclohexene (we do not believe that
this was what the authors38 meant in reporting Scheme 5),
but it could react with another intermediate radical, X•, includ-
Su p p or tin g In for m a tion Ava ila ble: GC-MS analysis,
MS of 2-phenylpropenol and 2-phenylpropenal from the reac-
tion of R-methylstyrene with DMD, and NMR spectra before
and after the reaction of adamantane with DMD, acquired by
using the presaturation technique (5 pages). This material is
contained in libraries on microfiche, immediately follows this
article in the microfilm version of the journal, and can be
ordered from the ACS; see any current masthead page for
ordering information.
J O971226W
(39) A reviewer suggested that the rate of eq 14 could be close to
diffusion-controlled limit; we agree and thank the reviewer for this
suggestion.
(38) Asensio, G.; Mello, R.; Gonzalez-Nun˜ez, M. E.; Boix, C.; Royo,
J . Tetrahedron Lett. 1997, 38, 2373.