Radical versus "oxenoid" oxygen insertion mechanism in the oxidation of alkanes and alcohols by aromatic peracids. New synthetic developments

Article abstract of DOI:10.1021/jo961366q

Evidences concerning a novel free-radical mechanism for the oxidation of alkanes by aromatic peracids are reported. The mechanism involves hydrogen abstraction from the OH group of peracids by an aroyloxyl radical; the acylperoxyl radical thus generated is responsible for the selective oxidation. The reaction is affected by the presence of oxygen and it is inhibited by TEMPO and by solvents forming hydrogen bonds with peracids. A more simple and effective synthetic procedure has been developed, on the basis of the autoxidation of aliphatic and aromatic aldehydes in the presence of alkanes. It is also shown that the previously reported inertness of alcohols toward peracids must be ascribed to solvent effects, due to the formation of hydrogen bonds; in suitable solvents alcohol oxidation smoothly occurs.

Full text of DOI:10.1021/jo961366q

J . Org. Chem. 1996, 61, 9409-9416
9409
Ra d ica l ver su s “Oxen oid ” Oxygen In ser tion Mech a n ism in th e
Oxid a tion of Alk a n es a n d Alcoh ols by Ar om a tic P er a cid s. New
Syn th etic Develop m en ts
Anna Bravo, Hans-Rene´ Bjorsvik, Francesca Fontana, Francesco Minisci,* and Anna Serri
Dipartimento di Chimica del Politecnico, via Mancinelli 7, I-20131 Milano, Italy
Received J uly 19, 1996X
Evidences concerning a novel free-radical mechanism for the oxidation of alkanes by aromatic
peracids are reported. The mechanism involves hydrogen abstraction from the OH group of peracids
by an aroyloxyl radical; the acylperoxyl radical thus generated is responsible for the selective
oxidation. The reaction is affected by the presence of oxygen and it is inhibited by TEMPO and by
solvents forming hydrogen bonds with peracids. A more simple and effective synthetic procedure
has been developed, on the basis of the autoxidation of aliphatic and aromatic aldehydes in the
presence of alkanes. It is also shown that the previously reported inertness of alcohols toward
peracids must be ascribed to solvent effects, due to the formation of hydrogen bonds; in suitable
solvents alcohol oxidation smoothly occurs.
An “oxenoid” oxygen insertion mechanism has been
suggested to explain unusual selectivities in the oxidation
of unactivated C-H bonds by a variety of oxidants and
catalytic systems (dioxiranes,1 perfluorooxaziridines,2 Gif
systems,3 cytochrome P450,4 metalloporphyrins, and
other metal salt complexes5), sometimes characterized by
enzyme-like specificity.
The suggested radical mechanism13 involves a chain
process initiated by the homolysis of the peroxidic bond
(eq 1), with propagation steps characterized by hydrogen
abstraction from unactivated C-H bonds and induced
decomposition of the peracid by alkyl radical (eqs 2-4).
+ •OH
Ph
C
O
C
O
O
OH
k2
Ph
C
O
O•
(1)
We and other research groups have recently reported
evidences suggesting that oxidations by Gif systems,6-8
dioxiranes9,10 and metalloporphyrin catalysis11,12 can be
explained by radical mechanisms. This dualism (oxenoid
oxygen insertion versus radical mechanism) is quite
general. Aromatic peracids represent another class of
oxidants, characterized by high selectivity in the oxida-
tion of nonactivated C-H bonds, for which both mecha-
nisms have been proposed.
Ph• + CO2
Ph
O•
(2)
k2 = ~106s–1
Ph
C
O
O• (Ph•) + H–R
PhCOOH (Ph–H) + R•
(3)
(4)
R• + PhCOOOH
ROH + PhCOO•
This mechanism, however, is not consistent with the
selectivity of the process. Aryl radicals can be ruled out
as hydrogen-abstracting species as the oxidation has been
successfully carried out in benzene as solvent: it is
known14 that the addition rate of phenyl radical to
benzene (∼106 M-1 s-1) is at least of the same order of
magnitude as the rate of hydrogen abstraction from
unactivated C-H bonds (105-106 M-1 s-1). Moreover, the
regio- and chemoselectivity of hydrogen abstraction by
phenyl radical is rather low,14 while this oxidation is
highly selective.
In spite of the clearly incongruous rate constants
reported in a recent book on free radicals15 (1.4 × 106
M-1 s-1 at 22 °C and 2.5 × 103 M-1 s-1 at 25 °C for
hydrogen abstraction by benzoyloxyl radicals respectively
from cyclohexane and THF), very reliable absolute rate
constants have been reported by the research groups of
Ingold16 and Tokumaru,17 who have clearly shown that
hydrogen abstraction from alkanes by aroyloxyl radicals
can compete with decarboxylation, which is subject to
dramatic solvent effects and also to substituent effects.16
X Abstract published in Advance ACS Abstracts, December 1, 1996.
(1) Adam, W.; Hadjiarapoglou, L. P. Top. Curr. Chem. 1993, 49, 2227
and references therein.
(2) Arnone, A.; Cavicchioli, M.; Montanari, V.; Resnati, G. J . Org.
Chem. 1994, 59, 5511.
(3) Barton, D. H. R.; Doller, D. Acc. Chem. Res. 1992, 25, 504.
Barton, D. H. R.; Hill, D. R. Tetrahedron Lett. 1994, 35, 1431.
(4) Cytochrome P450: Structure, Mechanism and Biochemistry;
Ortiz de Montellano, P. R., Ed.; Plenum Press: New York, 1986.
(5) Metallo-porphyrin catalyzed oxidation; Montanari, F.; Casella,
L., Eds.; Kluwer Academic Publishers: Dordrecht, 1994. Kaufman, M.
D.; Grieco, P. A.; Bougie, D. W. J . Am. Chem. Soc. 1993, 115, 11648;
Nomura, K.; Vemura, S. J . Chem. Soc., Chem. Commun. 1994, 129.
(6) Minisci, F.; Fontana, F. Tetrahedron Lett. 1994, 35, 1427. Minisci,
F.; Fontana, F.; Araneo, S.; Recupero, F. J . Chem. Soc., Chem.
Commun. 1994, 823. Minisci, F.; Fontana, F.; Araneo, S.; Recupero,
F. Tetrahedron Lett. 1994, 35, 3759. Minisci, F.; Fontana, F.; Bravo,
A.; Yan, Y. M. Gazz. Chim. Ital., 1996, 126, 85. D. H. R. Barton does
not consider anymore “Gif chemistry” the oxidations by GoAggIV and
GoAggV systems: Synlett, in press.
(7) Snelgrove, D. W.; MacFaul, P. A.; Ingold, K. U.; Wayner, D. D.
M. Tetrahedron Lett. 1996, 37, 823.
(8) Newcomb, M.; Simakov, P. A.; Park, S. Tetrahedron Lett. 1996,
37, 819.
(9) Minisci, F.; Zhao, L.; Fontana, F.; Bravo, A. Tetrahedron Lett.
1995, 36, 1697, 1895. Bravo, A.; Fontana, F.; Fronza, G.; Mele, A.;
Minisci, F. J . Chem. Soc., Chem. Commun. 1995, 1573. Bravo, A.;
Fontana, F.; Fronza, G.; Minisci, F.; Serri, A. Tetrahedron Lett. 1995,
36, 6945.
(10) Vanni, R.; Garden, S. J .; Banks, J . T.; Ingold, K. U. Tetrahedron
Lett. 1995, 36, 7999.
(11) Minisci, F.; Fontana, F.; Araneo, S.; Recupero, F.; Banfi, S.;
Quici, S. J . Am. Chem. Soc. 1995, 117, 226.
(14) Kryger, R. G.; Lorand, J . P.; Stevens, N. R.; Herron, N. R. J .
Am. Chem. Soc. 1977, 99, 7589. Scaiano, J . C.; Stewart, L. C. J . Am.
Chem. Soc. 1983, 105, 3609.
(15) Reference 13a, p 290.
(12) Meunier, B. Chem. Rev. 1992, 92, 1411.
(16) Chateauneuf, J .; Lusztyk, J .; Ingold, K. U. J . Am. Chem.
Soc.1987, 109, 897; 1988, 110, 2877, 2886.
(17) Misawa, H.; Sawabe, K.; Takahara, S.; Sakuragi, H.; Tokumaru,
K. Chem. Lett. 1988, 355.
(13) (a) Fossey, J .; Lefort, D.; Sorba, J . Free Radicals in Organic
Chemistry; Wiley: New York, 1995; p 217. (b) Tokumaru, K.; Sima-
mura, O. Bull. Chem. Soc. J pn. 1962, 35, 1678.
S0022-3263(96)01366-7 CCC: $12.00 © 1996 American Chemical Society

9410 J . Org. Chem., Vol. 61, No. 26, 1996
Bravo et al.
Sch em e 1
phenyl (6) is an important byproduct; also in this case,
the presence of oxygen inhibits the formation of 6.
In CHCl3 solution the results are similar to those
obtained in CH2Cl2 and DCE (compounds 1-5 are
formed), with the difference that a considerable amount
of 1-chloroadamantane is also formed.
By using a more effective chlorine-transfer solvent,
such as CCl4, chlorination of adamantane with low
selectivity (mono- and dichloro derivatives) is by far
prevailing, while m-dichlorobenzene and hexachloro-
ethane were identified as byproducts.
A still more effective halogen-transfer reagent, such
as CBrCl3, is effective also at low concentrations, and the
ratio between oxidation and halogenation products is
strictly related to CBrCl3 concentration. Bromoadaman-
tanes are prevailing, but 1-chloroadamantane is a sig-
nificant byproduct. The oxidation is substantially inhib-
ited by the use of t-BuOH as solvent or in the presence
of TEMPO in DCE solution. The presence of cyclohex-
anone also inhibits adamantane oxidation. The results
are reported in Table 1.
The effect of the solvent on the rate of adamantane
oxidation by m-CPBA is shown in Table 2: the reaction
is slower in benzene than in DCE, and it is practically
inhibited in t-BuOH solution. The results of the oxida-
tion of adamantane by aldehydes (butanal, benzaldehyde,
and m-chlorobenzaldehyde) and oxygen are reported in
Table 3. The reaction was carried out at 70 °C in the
absence of initiator and at 40 °C in the presence of (t-
BuOOCO)2 as initiator. In the presence of an excess of
cyclohexanone, the reaction is not inhibited; 1-hydroxy-
adamantane and 1-adamantyl esters similar to 2 are the
main reaction products.
Oxid a tion of Cycloh exa n e. Cyclohexanol, cyclohex-
anone, and caprolactone are the reaction products of
cyclohexane oxidation in DCE or CH2Cl2 by m-CPBA.
Chlorobenzene (5) is a significant byproduct of the
reaction; t-BuOH again inhibits the reaction. In CCl4
solution chlorination of cyclohexane exclusively takes
place and m-dichlorobenzene and hexachloroethane are
significant reaction products. Bromocyclohexane is the
reaction product in DCE solution in the presence of
CBrCl3, and m-chlorobromobenzene is a byproduct. The
results are reported in Table 4. In a competitive experi-
ment, adamantane was >50 times more reactive than
cyclohexane.
However, the observed selectivity for hydrogen abstrac-
tion (5.3 × 105 and 7.3 × 105 M-1 s-1 at 24 °C in CCl4
respectively for cyclohexane and THF),16 is very low and
it is not consistent with the high regio- and chemoselec-
tivity observed in the oxidation by aromatic peracids.
Moreover, eq 4 does not satisfactorily explain the ob-
served stereoselectivity. The high degree of configura-
tional retention in the oxidation of epimeric cycloalkanes
has in facts led other authors18 to exclude a radical
mechanism and to suggest an oxenoid oxygen insertion
mechanism based on a cyclic transition state with some
charge separation, as first proposed by Bartlett19 (Scheme
1).
A quite similar transition state has been suggested in
alkane oxidation by dioxiranes; LFER studies with both
oxidants (dioxirane and peracids) would indicate that
these hydroxylations occur in a concerted electrophilic
process (Scheme 1).20 A very recent evolution of the
“concerted oxenoid mechanism of insertion” by dioxiranes
has been reported21 according to Scheme 2.
Though a concerted insertion mechanism by definition
does not involve intermediates, the radical pair depicted
in Scheme 2 is considered not distinct. This lack of
distinction would prevent the coupling of the “singlet”
radical pair in the cage and would lead to a hydroxyl
transfer and to a not better defined “radical leakage”.
We now report evidences concerning a novel free-
radical mechanism, different from the one described by
eqs 1-4, for the oxidation of alkanes and alcohols by
peracids, and new synthetic developments arising as
direct consequence of this interpretation.
Resu lts
We have investigated the oxidation of adamantane,
cyclohexane, cyclohexanol, cyclopentanol, 1-heptanol,
2-heptanol, and benzyl alcohol at different temperatures
in several solvents by m-chloroperbenzoic acid (m-CPBA)
or by aldehydes and molecular oxygen.
Oxid a tion of Ad a m a n ta n e. Adamantane has been
oxidized by m-CPBA in several solvents at temperatures
ranging between 18 and 65 °C in an atmosphere of air,
oxygen, or argon. In CH2Cl2, ClCH2CH2Cl (DCE), and
benzene solutions two main reaction products are formed
from adamantane, with high regioselectivity: 1-hydroxy-
adamantane (1) and 1-adamantyl-m-chlorobenzoate (2);
minor amounts of 2-hydroxyadamantane (3) and traces
of adamantanone (4) are also formed. In methylene
chloride or DCE solution under an argon atmosphere,
chlorobenzene (5) is a significant reaction product at 18
°C and its amount increases with temperature. At 18
°C under oxygen atmosphere 5 is practically absent. No
chlorobenzene is formed under an argon atmosphere, but
in the absence of adamantane. When benzene is utilized
as solvent no chlorobenzene is formed, but m-chlorobi-
Oxid a tion of Alcoh ols. Cyclohexanol and cyclopen-
tanol are oxidized by m-CPBA respectively to caprolac-
tone and valerolactone in all the solvents investigated
(CH2Cl2, DCE, CHCl3, CCl4, benzene); t-BuOH, when
used as solvent, inhibits the oxidation of both alcohols.
Chlorobenzene is a byproduct in all the reactions but the
ones carried out in benzene or CCl4 solution, in which
m-dichlorobenzene is the byproduct. The oxidation of
1-heptanol gives heptanoic acid as the main reaction
product and heptanal and hexyl formate as minor prod-
ucts. Similarly, benzyl alcohol mainly gives benzoic acid,
with minor amounts of benzaldehyde; 2-heptanol is
oxidized to hexyl acetate, 2-heptanone, and methyl hex-
anoate. The results are reported in Table 5. The effect
of the solvent on the oxidation rate of cyclohexanol by
m-CPBA is reported in Table 6. The reaction is slower
in benzene than in CH2Cl2, and it is substantially
inhibited in t-BuOH.
(18) Schneider, H. J .; Mu¨ller, W. J . Org. Chem. 1985, 50, 4609. Tori,
M.; Matsuda, R.; Asakawa, Y. Tetrahedron Lett. 1985, 26, 227.
(19) Bartlett, P. D. Rec. Chim. Progr. 1950, 11, 47.
(20) Murray, R. W.; Gu, H. J . Org. Chem. 1995, 60, 5673.
(21) Curci, R., Dinoi, A.; Fusco, C.; Lillo, M. A. Tetrahedron Lett.
1996, 37, 249.
The oxidation of cyclohexanol has also been carried out
under autoxidation conditions of butanal: cyclohexanone
and caprolactone are the main reaction products (Table

Radical versus “Oxenoid” Oxygen Insertion Mechanism
Sch em e 2
J . Org. Chem., Vol. 61, No. 26, 1996 9411
Ta ble 1. Oxid a tion of Ad a m a n ta n e by m -CP BAa
reaction
time, h
mono- and
2, %c 3, %c dihaloadamantanesc 5, %d
entry
solvent
conversion, %b T, (°C)
1, %c
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
DCEe
30.2
29.4
44.1
32.1
31.3
15.3
49.9
18
18
65
20
20
40
60
65
18
18
65
65
65
65
40
40
65
100
100
24
100
100
24
24
24
96
96
24
24
24
24
24
24
24
70.1
77.2
68.0
72.2
74.2
66.5
39.4
∼1.0
75.2
74.5
84.6
25.1
7.3
19.3
11.4
10.5
16.3
15.4
10.4
3.5
1.6
1.3
2.0
1.1
1.3
1.9
1.2
18.1
DCEf
DCEg
CH2Cl2
CH2Cl2
CH2Cl2
24.9
16.6
e
f
g
31.3
9.5
g
CHCl3
18.1 (1-chloro)
88.2
g
CCl4
∼99
benzeneh
benzenei
benzenel
9.4
11.4
21.0
73.3
74.9
95.4
3.1
18.3
17.6
10.8
4.8
0.8
0.9
2.4
0.8
0.3
DCEg (CBrCl3, 0.25 mmol)
DCEg (CBrCl3, 0.5 mmol)
DCEg (CBrCl3, 1 mmol)
53.3
77.2
85.1
1.2
g
CH2Cl2 (TEMPO 0.5 mmol)
traces
traces
t-BuOHg
4.2
DCEm (cyclohexanone 3 mmol)
a
b
DCE ) 1,2-dichloroethane. 1 mmol of adamantane, 1 mmol of m-CPBA in 5 mL solvent. Conversion of adamantane. c Yields based
on converted adamantane. Yields based on m-CPBA. e Argon atmosphere. f Oxygen atmosphere. Air atmosphere. Argon atmosphere,
5% of 6 based on m-CPBA is formed. Oxygen atmosphere, no 6 is formed. Air atmosphere, 18% of 6, based on m-CPBA, is formed.
Caprolactone is the only reaction product.
d
g
h
i
l
m
Ta ble 2. Solven t Effect on th e Ra te of Oxid a tion of
hydroperoxides considerably decrease the rate of reaction
5.11,12 Thus, it seems reasonable to assume a high rate
constant and a marked solvent effect for reaction 6 as
well, on the basis of a similar difference (∼17 kcal/mol)23
between the bond energies of ArCOO-H and ArCOOO-H
and similar absolute rate constants in hydrogen abstrac-
tion from unactivated C-H bonds by t-BuO• radical11,22
and ArCOO• 16 radicals (105-106 M-1 s-1).
Ad a m a n ta n e by m -CP BA (20° C)a
conversion, %b
reaction
time, h
DCE
benzene
t-BuOH
4
12
24
48
72
96
5.1
9.4
13.3
21.6
27.9
33.9
2.7
3.9
5.3
8.1
10.6
12.9
traces
0.4
0.6
Ar
C
O
O• + Ar
C
O
OOH
Ar
C
O
OH + Ar
C
O
OO•
(6)
a
1 mmol of adamantane, 1 mmol of m-CPBA in 5 mL of solvent,
b
air atmosphere. Conversion of adamantane.
3). Competitive experiments indicate that the R-hydroxy
C-H bond in cyclohexanol is about three times more
reactive than the C3-H bond in adamantane, while
benzyl alcohol is slightly less reactive than 1-heptanol
(relative rates 1:1.2)
Reaction 6 could therefore successfully compete, in
suitable solvents, with reactions 2 and 3, whose rate
constants have been evaluated16 to be in the range 105-
106 M-1 s-1. This means it could play a relevant role in
the oxidation of unactivated C-H bonds by peracids.
Adamantane is a useful substrate for the investigation
of the relationship between the structure of a free radical
and its behavior as hydrogen-abstracting species. The
selectivity of abstraction between tertiary and secondary
C-H bonds is scarcely influenced by the enthalpic effect24
(bromine atom, highly sensitive to C-H bond energies,
shows low selectivity24), but highly influenced by the
polar effect24 (high selectivity by R2NH•+ radicals, Minisci
chlorination,25 particularly sensitive to polar effects26).
This behavior is not unexpected, on the grounds of the
C3-H bond energy in adamantane, evaluated27 to be 3.7
Discu ssion
The novel interpretation of the oxidation of unactivated
C-H bonds by peracids was originated by previous
studies11,12 concerning the reaction of alkoxyl radicals
with hydroperoxides (eq 5). Reaction 5 is very fast, about
k5
t-BuO• + HOOBu-t
t-BuOH + t-BuOO•
(5)
k5 = 2.5 × 108 M–1s–1 at 22 °C
2 orders of magnitude faster than hydrogen abstraction
from unactivated C-H bonds11,22 by t-BuO• radical,
mainly due to enthalpic effects (bond energies 88 and 103
kcal/mol respectively for t-BuOO-H and t-BuO-H).
Moreover, solvents which form hydrogen bonds with
(23) Reference 13a, p 298.
(24) Minisci, F.; Fontana, F.; Zhao, L.; Banfi, S.; Quici, S. Tetrahe-
dron Lett. 1994, 35, 8033.
(25) Deno, N. C. In Methods in Free Radical Chemistry; Huyser, E.
S., Ed.; M. Dekker: New York, 1975, p 143.
(26) Bernardi, R.; Galli, R.; Minisci, F. J . Chem. Soc. B 1968, 324.
Minisci, F. In Substituent Effects in Free-Radical Chemistry; Viehe,
H., Ed.; Reidel: Dordrecht, 1986; p 391.
(27) Kruppa, G. H.; Beauchamp, J . L. J . Am. Chem. Soc. 1986, 108,
2162.
(22) Minisci, F.; Fontana, F.; Araneo, S.; Recupero, F.; Zhao, L.
Synlett 1996, 119. Avila, D. V.; Ingold, K. U.; Lusztyk, J .; Green, W.
H.; Procopio, D. R. J . Am. Chem. Soc. 1995, 117, 2929.

9412 J . Org. Chem., Vol. 61, No. 26, 1996
Ta ble 3. Oxid a tion of Ad a m a n ta n e a n d Cycloh exa n ol by Ald eh yd es a n d O2
Bravo et al.
a
substrate
aldehyde
butanal
benzaldehyde
benzaldehyde
m-Cl-benzaldehyde
m-Cl-benzaldehyde
conversion, %b
T, °C
1, %c
3, %c
1-adamantyl esters, %c
adamantane
adamantane
adamantaned
adamantaned
adamantanee
47.1
31.2
36.9
34.2
20.1
70
70
40
40
70
78.2
1.2
13.1
12.2
6.1
8.2
7.3
84.1
82.4
75.3
72.4
1.4
1.3
2.1
2.8
cyclohexanone, %c
caprolactone, %c
cyclohexanol
butanal
53.4
70
61.4
16.6
a
2 mmol aldehyde in 5 mL DCE are dropped during 3 h to a solution of 1 mmol substrate in 5 mL DCE and oxygen is bubbled for 24
b
d
h. Conversion of the substrate. c Yields based on converted substrate. As in (a) by using 0.1 mmol (t-BuOOCO)2. e As in (a) in the
presence of 3 mmol cyclohexanone.
Ta ble 4. Oxid a tion of Cycloh exa n e by m -CP BAa
conversion,
cyclohexanol,
cyclohexanone,
caprolactone,
mono- and
b
c
c
c
solvent
DCE
%
%
%
%
dihalocyclohexanes, %c
5, %d
32.3
21.2
3.1
49.4
45.6
21.2
24.4
40.2
traces
traces
20.3
21.2
15.7
39.9
33.6
23.4
23.4
25.4
0.9
CH2Cl2
t-BuOH
e
CCl4
88.1
83.8
DCEf (CBrCl3)
a
b
5 mmol of cyclohexane, 1 mmol of m-CPBA in 5 mL of solvent at 65 °C for 24 h, air atmosphere. Conversion of cyclohexane based
on m-CPBA. c Yields based on converted cyclohexane. Yields based on m-CPBA. e m-Dichlorobenzene (15.5%) and CCl3CCl3 (3.2%) are
d
formed. f m-Chlorobromobenzene (6.4%) is formed.
Ta ble 5. Oxid a tion of Alcoh ols by m -CP BAa
T,
°C
reaction
time, h
conversion,
b
alcohol
solvent
%
reaction products, %c
5, %d
cyclohexanol
cyclohexanol
cyclohexanol
cyclohexanol
cyclohexanol
cyclohexanole
cyclohexanol
cyclohexanol
cyclopentanol
cyclopentanol
cyclopentanol
1-heptanol
DCE
DCE
60
20
40
18
60
60
60
18
60
60
60
60
60
60
40
60
60
60
20
70
27
120
24
5
20
120
24
24
24
21
21
21
24
24
24
24
75.1
85.2
52.4
66.3
93.1
51.6
35.5
3.3
74.2
27.8
traces
70.1
77.7
41.2
82.9
85.6
58.8
41.2
caprolactone (87.1)
caprolactone (89.2)
caprolactone (84.1)
caprolactone (86.2)
caprolactone (78.8)
caprolactone (81.4)
caprolactone (82.3)
caprolactone (48.2), cyclohexanone (31.6)
valerolactone (79.2)
valerolactone (86.3)
traces
heptanoic acid (68.2), hexyl formate (7.1), heptanal (3.3)
2-heptanone (43.1), hexyl acetate (39.7), methyl hexanoate (2.5)
2-heptanone (14.1), hexyl acetate (64.9), methyl hexanoate (2.8)
2-heptanone (38.3), hexyl acetate (41.2), methyl hexanoate (2.6)
2-heptanone (32.7), hexyl acetate (46.2), methyl hexanoate (2.8)
2-heptanone (21.6), hexyl acetate (62.7), methyl hexanoate (3.1)
benzaldehyde (27.4), benzoic acid (52.3)
9.9
4.2
10.9
3.8
CH2Cl2
CH2Cl2
CHCl3
CCl4
benzene
t-BuOH
DCE
benzene
t-BuOH
DCE
5.5
8.9
6.9
2-heptanol
2-heptanol
2-heptanol
DCE
CH2Cl2
CH2Cl2
CHCl3
CCl4
10.4
8.3
2-heptanol
2-heptanole
benzyl alcohol
DCE
12.2
a
b
1 mmol of alcohol, 2 mmol of m-CPBA in 5 mL solvent, air atmosphere. Alcohol conversion. c Yields based on alcohol conversion.
Yields based on m-CPBA. e Ratio 1:1 between alcohol and m-CPBA, m-dichlorobenzene (8%) as byproduct.
d
SCE),29 which explains the high sensitivity to polar
effects for hydrogen abstraction by electrophilic radicals.
Ta ble 6. Solven t Effect on th e Ra te of Oxid a tion of
Cycloh exa n ol by m -CP BA (18 °C)a
conversion, %b
The results of Table 1 clearly indicate that free radicals
are certainly involved in the oxidation of adamantane by
m-CPBA. Esters such as 2 have never been previously
reported in alkane oxidation by peracids. This had been
considered18 a significant evidence for an “oxenoid”
oxygen insertion mechanism, against electrophilic or free-
radical mechanisms. The only reasonable mechanism for
the formation of chlorobenzene (5) in CH2Cl2, CHCl3, or
DCE solution and of m-chlorobiphenyl (6) in benzene
solution involves decarboxylation of m-chlorobenzoyloxyl
radical (eq 2) and subsequent hydrogen abstraction from
the solvent (S-H, eqs 7 and 8) or homolytic substitution
of benzene (eqs 9 and 10) by m-chlorophenyl radical. The
different results obtained under oxygen and under argon
atmospheres represent a further strong evidence that free
radicals are involved in the reaction.
reaction
time, h
CH2Cl2
benzene
t-BuOH
12
24
48
72
96
12.2
21.3
34.1
46.4
56.8
66.6
4.2
7.4
12.5
15.8
18.6
21.3
traces
0.9
1.6
120
2.3
a
1 mmol of cyclohexanol, 1 mmol of m-CPBA in 5 mL of solvent,
b
air atmosphere. Conversion of cyclohexanol.
kcal/mol greater than C3-H bond energy in isobutane,
and therefore quite close to C2-H bond energy in alkanes.
Moreover, 1-adamantyl cation is considerably more stable
than its (secondary) 2-isomer,28 and the oxidation poten-
tial of adamantane (Eox ) 2.72 V vs SCE) is considerably
lower than that of 2,3-dimethylbutane (Eox ) 3.45 V vs
In CHCl3 solution the results are similar to those
obtained in CH2Cl2 and DCE, with the difference that
considerable amounts of 1-chloroadamantane are also
(28) Olah, G. A.; Liang, G.; Mateescu, G. D. J . Org. Chem. 1974,
39, 3750.
(29) Albini, A. Personal communication.

Radical versus “Oxenoid” Oxygen Insertion Mechanism
J . Org. Chem., Vol. 61, No. 26, 1996 9413
•
m-ClC6H4COO•
m-ClC6H4 + CO2
m-ClC6H4–H + S•
tion mechanism, suggest a general free-radical chain
mechanism. Above all, the fact that the formation of the
free-radical products 5 and 6 occurs at room temperature
in the presence of adamantane, but not in its absence
under the same conditions, strongly suggests that ada-
mantane induces the homolysis of m-CPBA.
(7)
(8)
•
m-ClC6H4 + H–S
H
•
(9)
m-ClC6H4
+
•
m-ClC6H4
Since the free-radical mechanism previously13 reported
is not consistent with the chemo-, regio-, and above all
stereoselectivity observed, we suggest a novel general
free-radical chain mechanism based on eq 6, which would
explain the formation of all the reaction products. The
acylperoxyl radical generated in eq 6 selectively abstracts
hydrogen from adamantane, leading to adamantyl radi-
cal, which rapidly reacts with the peracid in the cage with
formation of 1, 2, and 3 (eq 16).
H
•
+ m-ClC6H4COOOH
m-ClC6H4
m-ClC6H4
+ m-ClC6H4COO• + H2O
(10)
formed. The result is surprising, as CHCl3 has been one
of the most utilized solvents in alkane oxidation and until
now alkane chlorination has never been reported.18 The
only reasonable explanation we can see for the formation
of 1-chloroadamantane is chlorine abstraction from CHCl3
by 1-adamantyl radical (eq 11). On the other hand, it is
Ar
C
O
OO• + H–Ad
(a)
(b)
O• + AdOH
Ar
C
O
OOH •Ad
Ar
C
O
cage
1–Ad• + CHCl3
1–AdCl + •CHCl2
(16)
(11)
OOAd + •OH
Ar
C
O
well-known30 that tertiary alkyl radicals abstract hydro-
gen and chlorine atoms from CHCl3 by quite similar
rates.
By using a more effective chlorine-transfer solvent,
such as CCl4, free-radical chlorination of adamantane
(eqs 12 and 13) is by far prevailing and the byproducts,
m-dichlorobenzene and hexachloroethane, clearly arise
from eqs 14 and 15. CBrCl3 is effective as halogen
It is known that (alkyl) peroxyl radicals do not abstract
hydrogen atoms from alkanes at a significant rate (∼10-2
M-1 s-1 ) at room temperature, but benzoyl peroxyl
radicals are about 800-900 times as reactive in hydrogen
atom abstraction from hydrocarbons as HOO• (which has
a “normal” (alkyl) peroxyl reactivity).32 Thus reaction 16
should be fast enough to carry the chain.33 Equation 16a
prevails over eq 16b, and eqs 6 and 16a are the propaga-
tion steps of the free-radical chain for the formation of
the main reaction product 1. The competitive incursion
of eq 7 leads to the classical reaction products arising
from aryl radical (5 and 6), which do not interrupt the
radical chain. In fact, in benzene solution (formation of
6) the acyloxyl radical is regenerated by eq 10, while in
hydrogen donor solvents (CH2Cl2, DCE) leading to 5,
radical S• in eq 8 is probably oxidized by the peracid
according to eq 17 (S ) ClCH•CH2Cl in DCE solution),
even if this oxidation is certainly affected by the nucleo-
philic character of the alkyl radical.
Ad• + CCl4
AdCl + •CCl3
(12)
Ad–H + •CCl3
Ad• + H–CCl3
Cl
(13)
•
+
•CCl3
+ CCl4
Cl
(14)
(15)
Cl
2
•CCl3
Cl3C–CCl3
transfer reagent also at low concentration and, even if
bromoadamantanes are prevailing, 1-chloroadamantane
is a mechanistically significant byproduct, as will be
discussed later on.
Thus, there is no doubt concerning the involvement of
free radicals in the oxidation of adamantane by m-CPBA,
but the overall behavior could be, in principle, explained
by two simultaneous competitive mechanisms: free radi-
cal and oxenoid oxygen insertion. However, the signifi-
cant formation of free-radical products (5 and 6) even at
room temperature (18 °C), where the thermal homolysis
of the peroxidic bond occurs only in negligible traces,
according to the O-O bond energy (34-38 kcal/mol)
evaluated31 for peracids, makes the simultaneous pres-
ence of two competitive mechanisms unlikely. Moreover,
the inhibition by TEMPO, the halogenation of adaman-
tane in the presence of CHCl3, CCl4, and CBrCl3 and the
formation of 1-adamantyl m-chlorobenzoate, whose for-
mation cannot be explained by an oxenoid oxygen inser-
•
ClCH2CHCl + ArCOOOH
ArCOO• + ClCH2CHClOH
ClCH2CHO + HCl (17)
The high regioselectivity between tertiary and second-
ary C-H bonds in adamantane oxidation is mainly
explained by the electrophilic character of the acylperoxyl
radical, which is higher than that of other oxygen-
centered radicals, due to the electron-withdrawing effect
of the acyl group on the peroxyl radical, to the higher
stability of 1-adamantyl as compared to its (secondary)
2-isomer cation28 and to the low oxidation potential of
adamantane.29
The high degree of configurational retention in the
oxidation of epimeric cycloalkanes18 should be due to a
fast cage reaction of the alkyl radical with the peracid
(eq 16, an oxygen rebound mechanism).
(30) Frith, P. G.; McLauchlan, K. A. J . Chem. Soc. Faraday Trans.
2 1976, 72, 87. Duetsch, H. R.; Fischer, H. Int. J . Chem. Kinet. 1982,
14, 659.
(31) Liktenshtein, G. I. Zh. Fiz. Chim. 1962, 36, 1503. Curci, R.;
Edwards, J . D. In Catalytic Oxidations with Hydrogen Peroxide as
Oxidant; Strukul, G., Ed.; Kluwer Academic Publishers: Dordrecht,
1992; p 48.
(32) Zaikov, G. E., Howard, J . A.; Ingold, K. U. Can. J . Chem. 1969,
47, 3017.
(33) We thank K. U. Ingold for this suggestion.

9414 J . Org. Chem., Vol. 61, No. 26, 1996
Bravo et al.
The behavior observed in the presence of oxygen at
room temperature (5 and 6 are not formed) is explained
by the fact that oxygen intercepts all carbon-centered
radicals outside the solvent cage. The fact that the
amount of chlorobenzene is higher in refluxing CH2Cl2
(40 °C) than in DCE at 65 °C (Table 1) under an air
atmosphere could appear surprising, since one would
expect an increase in decarboxylation of m-chloroben-
zoyloxyl radical at higher temperature, which actually
occurs in DCE solution (Table 1). The larger amount of
chlorobenzene in refluxing CH2Cl2 must therefore be
ascribed to the different concentration of oxygen in the
two solvents.
The first approach is based on the use of an aliphatic
aldehyde, butanal, because aliphatic peracids are unsuit-
able for the oxidation of alkanes. This is due, according
to our interpretation, to the fact that decarboxylation of
aliphatic acyloxyl radicals is much faster34 (∼1010 s-1
)
than that of aromatic acyloxyl radicals16 (∼106 s-1), which
determines a fast free-radical chain decomposition of the
peracid35 (eqs 21 and 22) and avoids the intermolecular
reaction corresponding to eq 6.
k
k ~ 1010s–1
ROH + RCOO•
(21)
(22)
RCOO•
R• + CO2
R• + RCOOOH
The fact that under argon, but in the absence of
adamantane at room temperature, 5 and 6 are not formed
suggests the possibility that adamantane might induce,
to a small extent, the homolysis of the peracid (eq 18),
thus initiating the free-radical chain of eqs 6 and 16a,
or in any case that it might sustain the chain process
much more effectively than the solvents (eqs 16 and 17).
Actually, the autoxidation of butanal in the presence
of adamantane or cyclohexanol provides reaction products
similar to those observed with aromatic aldehydes or
m-CPBA (Table 3). This result is strong evidence that
the species responsible for the oxidation is the acylperoxy
radical, formed according to eq 19 and not the peracid,
formed in situ.
The other approach was derived from the observation
that cyclohexanone is much more reactive (Baeyer-
Villiger) than adamantane toward m-CPBA. In the
presence of adamantane and cyclohexanone, only the
latter is oxidized (eq 23), which means that cyclohex-
anone completely inhibits adamantane oxidation by
m-CPBA.
Ad• + H2O + ArCOO•
(18)
Ad H + HOOOCAr
Equation18 can represent an alternative explanation
for the formation of 2 through coupling of the radical pair
(cage effect); in this case reaction 18 should take place
to a higher extent than it would as simple initiation
reaction of a chain process.
The inhibition in t-BuOH solution is related, in our
opinion, to a solvent effect, which inhibits reaction 6 by
forming hydrogen bonds between peracid and solvent; the
same effect has been observed11,22 for reaction 5 in t-BuOH
solution, and the analogy between hydroperoxides and
peracids appears to be particularly strict in this context.
The phenomenon is quite general; it has been observed
with adamantane (Table 2), cyclohexane (Table 4), and
cyclohexanol (Table 6).
The oxidation in benzene solution contrasts with the
previous report18 stating that aromatic nuclei prevent
reaction at tertiary C-H bonds in alkylbenzenes, due to
“the action of the benzene ring as hydrogen bond accep-
tor, thus deactivating the peracid”. Actually, benzene
decreases, to some extent, the rate of oxidation of
adamantane and cyclohexanol by m-CPBA (Tables 2 and
6), but it does not prevent the reaction. The inertness of
alkylbenzenes toward m-CPBA must be ascribed, in our
opinion, to polar effects (electron-withdrawing effect of
the phenyl group, i.e., cumene is less reactive than
cyclohexane toward t-BuO• radical for polar reasons11 in
spite of the lower energy of the benzylic C-H bond).
Acylperoxyl radicals and peracids are classical inter-
mediates in the autoxidation of aldehydes (eqs 19 and
20), in which both acylperoxyl radicals and peracids are
involved. We have obtained further evidence of the
O
O
+ ArCOOH (23)
+ ArCOOOH
O
However, autoxidation of m-chlorobenzaldehyde in the
presence of adamantane and cyclohexanone leads to
adamantane oxidation, with results similar to those
obtained with m-CPBA in the absence of cyclohexanone.
This result provides a further strong evidence that acyl-
peroxy radical is responsible for adamantane oxidation.
Some differences characterize the oxidation of cyclo-
hexane as compared to adamantane: cyclohexyl m-
chlorobenzoate is formed only in traces and bromocyclo-
hexane is the only product in the presence of CBrCl3,
while adamantane gives significant amounts of 1-chlo-
roadamantane, in addition to 1-bromoadamantane.
We explain these differences on the grounds of the
more marked nucleophilic character of 1-adamantyl
radical compared to cyclohexyl radical, as a consequence
of the particular stability of 1-adamantyl cation.28 This
polar character is reflected, for halogen abstraction and
for oxidation by peracids (eqs 24 and 25), in transition
states with more charge separation.
δ+
δ–
•
•
1–Ad• + ClCHCl2
1–Ad----Cl----CHCl2
1–AdCl + CHCl2
R
C
O + O2
R
C
O
OO•
(19)
(24)
H
O
OH
O
•
OH
OO•
R
R
C
OOH + R
C
O
δ+
+
R
C
O
(20)
δ–
1–Ad----
O
1–AdOCOAr + •OH
(25)
1–Ad• +
Ar
Ar
O
O
O
mechanism described by eqs 6 and 16 by investigating
the autoxidation of aldehydes in the presence of ada-
mantane and cyclohexanol. We have taken advantage
of two different approaches for obtaining evidence that
acylperoxyl radicals were responsible for the oxidation.
(34) Turetskaya, E. A.; Skakorskii, E. D.; Rikov, S. V.; Glazkov,
Yu.V.; Ol’dekop, Yu.A. Dokl. Akad. Nauk BSSR 1980, 24, 57.
(35) Lefort, D.; Paquat, C.; Sorba, J . Bull. Soc. Chim. Fr. 1959, 1385.

Radical versus “Oxenoid” Oxygen Insertion Mechanism
J . Org. Chem., Vol. 61, No. 26, 1996 9415
On the other hand, the peculiar behavior of 1-adaman-
tyl radical in chlorine and bromine abstraction from
CBrCl3, quite different from that of simple alkyl radicals,
which only abstract bromine, was first reported by
Ru¨chardt36 and confirmed by Tabushi.37 We believe that
a transition state similar to the one in eq 24 is responsible
for this behavior. A ratio of about 1:20 between 1-chloro-
and 1-bromoadamantane is a valuable diagnostic crite-
rion37 for the formation of 1-adamantyl radical in the
presence of CBrCl3.
OH
OH
+ •CCl3
+ CCl4
+ HCCl3
•
(27)
(28)
OH
O
+ •CCl3 + HCl
•
alcohol oxidation the involvement of acylperoxyl radicals
according to eq 16. This is particularly favored by the
stability of the R-hydroxy carbonium ion. Contrary to
the oxidation of cyclohexanol by m-CPBA, in which
caprolactone is the main reaction product (Table 5), in
this case cyclohexanone is prevailing, due to the low
concentration of the peracid formed from the aldehyde.
The high regioselectivity between C3-H and C2-H in
adamantane and the high chemoselectivity in competitive
oxidation of cyclohexane and adamantane (adamantane
is >50 times more reactive than cyclohexane) is also
explained by charge separation in the transition state (eq
26).
Con clu sion
The most significant evidences concerning a novel
mechanism for the selective oxidation of alkanes and
alcohols by peracids according to eqs 6 and 16 involve
the following aspects: (i) the formation of products from
aryl radical; (ii) the effect of oxygen on product distribu-
tion; (iii) induced homolysis of peracids by alkanes and
alcohols; (iv) halogen transfer from halogenated solvents
or reagents toward the carbon-centered radicals; (v)
oxidation of alkanes and alcohols during the autoxidation
of aldehydes; (vi) inhibition effect of solvents, which form
hydrogen bonds with peracids (this latter evidence ex-
plains the previously reported inertness of alcohols
toward peracids); (vii) inhibition by TEMPO.
δ+
OO----H----Ad
δ–
O
R
O
O
+ Ad•
(26)
+ HAd
R
R
OO•
OOH
This conclusion, of a high sensitivity to polar effects
by acylperoxy radicals, was in striking contrast with
previous reports38 stating that primary and secondary
alcohols are unreactive with peracids (perbenzoic acid
decomposes in ethanol solution in two months at room
temperature39). In fact, hydrogen abstraction from C-H
bonds in the R-position of alcohols is particularly sensitive
to the electrophilic character of the abstracting radical,
due to the particular stability of the R-hydroxy carbonium
ion. Also the results obtained with cyclohexane (Table
4), with cyclohexanone and caprolactone prevailing over
cyclohexanol even at low conversion (∼10%), are in
contrast with the same reports38,39 because they indicate
that cyclohexanol is much more reactive than cyclohex-
ane toward m-CPBA. Actually, primary and secondary
alcohols smoothly react in CH2Cl2, CHCl3, CCl4, or DCE
solution, giving the corresponding carbonyl compound
and the further oxidation products (lactones, esters,
carboxylic acids) (Table 5). Thus the previously38,39
reported inertness of primary and secondary alcohols
towards peracids must be ascribed to the fact that alcohols
were utilized as solvents, which inhibit reaction 6 by
forming hydrogen bonds with peracids. This is clearly
shown by the results in Table 6.
Competitive experiments, in fact, indicate that the
R-hydroxy C-H bond in cyclohexanol is about 3 times
more reactive than a C3-H bond in adamantane, while
benzyl alcohol is slightly less reactive than 1-heptanol
(relative rates 1:1.2) in spite of the weaker benzylic C-H
bond, once again supporting the influence of the polar
effect on oxidation rates.
The oxidation of alcohols is more effective in CHCl3 or
CCl4 solution; in these cases radical chains involving
•CCl3 radicals (eqs 27 and 28) are superimposed on the
oxidation chain (eqs 6 and 16) and contribute to the
overall oxidation of alcohols.
The high regio- and chemoselectivities are due to the
marked electrophilic character of acylperoxyl radicals,
and the previously reported stereoselectivity is ascribed
to a fast cage reaction of alkyl radicals with peracids.
Exp er im en ta l Section
Gen er a l Meth od s. Mass spectra were performed on a
GLC-MS Finnigan TSQ70 instrument, using a Varian 3700
gas chromatograph, equipped with SBP-1 fused silica column
(30 m × 0.2 mm i.d., 0.2 µm film thickness) and helium as
carrier gas.
GLC analyses were performed on a Dani 6500 capillary gas
chromatograph, equipped with a 25 m × 0.25 mm i.d. SBP-5
fused silica column (1 µm film thickness) at a hydrogen flow
rate of 8 cm3 min-1, PTV injector, flame ionization detector.
All of the solvents and the reagents were obtained from
commercial sources and were used without further purifica-
tion.
tert-Butyl peroxalate, (t-BuOOCO)2, was prepared according
to the literature procedure.40 All the reaction products were
commercially available and were utilized for the qualitative
identification (GLC-MS) and the quantitative analyses (GLC).
Gen er a l P r oced u r e for th e Oxid a tion of Alk a n es a n d
Alcoh ols. The procedure is extremely simple: m-CPBA and
the alkane or alcohol were dissolved in the solvent and stirred
at the temperature and for the time reported in Tables 1-6.
Most of the experiments were carried out under an air
atmosphere, while for few experiments (Table 1) the solution
was purged with argon or oxygen for 10 min before adding
m-CPBA and kept under argon or oxygen for the duration of
the experiment. The solvents and the concentration of the
reagents are reported in Tables 1-6. At the end of the
reaction, an internal standard was added and the solution was
directly analyzed by GLC-MS and GLC.
The oxidation of cyclohexanol under autoxidation
conditions of butanal gives cyclohexanone and caprolac-
tone as the main reaction products, supporting also for
Com p etitive Oxid a tion of Ad a m a n ta n e a n d Cycloh ex-
a n e. Five millimole of cyclohexane, 1 mmol of adamantane,
(36) Ru¨chardt, C.; Herwig, K.; Eichler, S. Tetrahedron Lett. 1969,
421.
(37) Tabushi, I.; Aoyama, Y.; Kojo, S.; Hamuro, J .; Yoshida, Z. J .
Am. Chem. Soc. 1972, 94, 1177.
(38) Bouillon, G.; Lick, C.; Schank, K. The Chemistry of Peroxides;
Patai, S., Ed.; Wiley: New York, 1983; p 295.
(39) Tokumaru, K.; Simamura, O.; Fukuyama, M. Bull. Chem. Soc.
J pn. 1962, 35, 1674.
(40) Bartlett, P. D.; Benzing, E. P.; Pincock, R. E. J . Am. Chem. Soc.
1960, 82, 1762.

9416 J . Org. Chem., Vol. 61, No. 26, 1996
Bravo et al.
and 1 mmol of m-CPBA were dissolved in 10 mL of DCE. The
solution was stirred for 48 h at 20 °C under air atmosphere
and analyzed by GLC-MS and GLC. The conversion of
adamantane was 20.9%, while the conversion of cyclohexane
was <1%; 1 and 2 were the reaction products, and only traces
(<0.2%) of 3, cyclohexanol, and cyclohexanone were formed.
The relative rate between adamantane and cyclohexane oxida-
tion can be thus evaluated as >50.
ratio. The C3-H of cyclohexanol, this way, is 3 times more
reactive than the C3-H of adamantane.
Oxid a tion of Ad a m a n ta n e a n d Cycloh exa n ol by Ald e-
h yd es a n d O2. P r oced u r e A. One millimole of cyclohexanol
or adamantane was dissolved in 5 mL of DCE. The solution
was purged with O2 for 10 min, and a solution of 2 mmol of
the aldehyde in 5 mL of DCE was dropped under stirring
during 3 h at 70 °C. O2 was introduced in the solution at a
rate of 2 mL/min for the duration of the experiment (24 h at
70 °C). At the end of the reaction, an internal standard was
added, and the solution was analyzed by GLC-MS and GLC.
P r oced u r e B. As in procedure A, with the differences that
0.1 mmol of (t-BuOOCO)2 was added to the aldehyde solution
in DCE and that the reaction temperature was 40 °C.
The results are reported in Table 3.
Com p etitive Oxid a tion of Ad a m a n ta n e a n d Cycloh ex-
a n ol. One millimole of cyclohexanol, 1 mmol of adamantane,
and 1 mmol of m-CPBA were dissolved in 10 mL of DCE. The
solution was stirred for 48 h at 18 °C under an air atmosphere
and analyzed by GLC-MS and GLC. The conversions of
cyclohexanol and adamantane were respectively 8.5% and
11.3%. Caprolactone, 1, and 2 were obtained in a 1.0:1.1:0.2
J O961366Q