892
Bull. Chem. Soc. Jpn. Vol. 82, No. 7 (2009)
Baeyer-Villiger Oxidation
AIBN/O2
Table 1. Baeyer-Villiger Oxidation of Cyclopentanone (1a)
with H2O2 Generated In Situ from Benzhydrol (2a) and O2
by NHPI/AIBN Systema)
O
O
NO
NOH
Conv./%
Yield/% (select/%)b)
Ratio/mmol
(1a/2a)
Entry
O
O
1a
2a
3ac)
4
PINO
NHPI
OH
Ph
OH
O2
1
2
3
4/6
6/6
8/6
4/6
4/6
4/6
4/6
4/6
4/6
4/6
4/6
4/(6)l)
94
78
66
25
28
66
85
95
96
99
98
62
99
88
99
99
93
99
99
99
99
99
99
®
71(76)
49(63)
42(64)
trace
90
82
78
50
71
90
99
99
99
99
99
®
Ph
Ph
Ph
2a
A
4d)
5e)
6f)
7g)
8h)
9i)
10j)
11k)
12
O
HO
Ph
HO
OO
Ph
OOH
Ph
9(32)
H2O2
+
Ph
Ph
Ph
42(64)
64(75)
62(65)
83(86)[76]
81(82)
71(72)
20(32)
4
B
C
O
HO OOH
O O
1
H2O2
+
n
n
O O
H H
E
n
n
1
D
a) 1a (4-8 mmol) was allowed to react with 2a (6 mmol) in the
presence of NHPI (0.6 mmol) and AIBN (0.3 mmol) under O2
(1 atm) in CH3CN (3 mL) at 75 °C for 22 h, followed by
evaporation (30 mmHg) at room temperature, and then HFIP
(6 mL) was added and stirred under air at 60 °C for 24 h.
b) Yield and selectivity were determined by GC analysis.
c) Yield based on 1a. Number in the bracket shows isolated
yield. d) Reaction was performed without AIBN. e) Reaction
was performed without NHPI. f) NHPI (0.3 mmol) and AIBN
(0.15 mmol) were used. g) Reaction time was 12 h in 1st step.
h) Reaction was performed under Ar in 2nd step. i) p-
TsOH¢H2O (0.02 mmol) was added in 2nd step. j) HFIP (3 mL)
was used. p-TsOH¢H2O (0.02 mmol) was added in 2nd step
under Ar. k) EtOAc was used instead of CH3CN. l) Number in
parenthesis shows the amount (mmol) of isopropyl alcohol
(2b).
O
HFIP
O
(n = 0,1,2)
n
3
Scheme 2. A possible reaction path for the BVoxidation of 1
with 2a and O2 by NHPI/AIBN system.
of 2a as a H2O2 source under these conditions to give 3b
in slightly lower yield (44%). This shows that 2b is a less
effective H2O2 source than 2a (Entry 3). However, excess 2b
was used to give 3b in good yield (74%) (Entry 4). Cyclo-
heptanone (1c) was a slightly reluctant substrate for the BV
reaction giving rise to 3c in 55% yield (Entry 7). However,
cyclooctanone (1d) underwent the BV oxidation with difficulty
under the present reaction conditions, forming ¡-hydroxy
cyclooctanone in preference to lactone 3d (Entry 8). A similar
observation was reported for the BVoxidation of 1d by Nozaki
et al.6 Cyclododecanone (1e) and cyclopentadecanone (1f) were
also oxidized to the corresponding lactones, 3e and 3f, in
69% and 81% yields, respectively (Entries 9 and 10). The
compound 3f, which is known as muscone, is an important
perfume. Adamantanone (1g) was converted to 4-oxatricyclo-
[4.3.1.13,8]undecan-5-one (3g) in good yield (93%) (Entry 11).
4-tert-Butylcyclohexanone (1h) was completely converted into
lactone (3h) (99%) (Entry 12).
lactone 3a in satisfactory yield. Needless to say, the reaction in
the absence of either AIBN or NHPI brought about 3a in trace
or poor yield, respectively (Entries 4 and 5). When the amount
of AIBN and NHPI was halved under these conditions, the
yield of 3a decreased to 42% (Entry 6). The reaction time in
step 1 was shortened to 12 h to form 3a in slightly lower yield
(64%) (Entry 7). The BVoxidation of cyclohexanone with 50%
H2O2 is reported to be accelerated by adding p-toluenesulfonic
acid ( p-TsOH) in HFIP.4b Thus, the reaction in step 2 was run
by adding a very small amount of p-TsOH to result in 3a in
good yield (83%) (Entry 9). By adding p-TsOH, the amount of
HFIP could be reduced by half (3 mL) to give 3a in satisfactory
yield (81%) (Entry 10). The reaction in EtOAc afforded almost
the same result as that in CH3CN (Entry 11). Unfortunately, the
reaction using isopropyl alcohol (2b) in place of 2a as a H2O2
source afforded 3a in low yield (20%) and selectivity (32%)
(Entry 12).
On the basis of these results, various cycloalkanones were
allowed to react under several reaction conditions (Table 2).
Cyclohexanone (1b) was reacted under the same conditions
as Entry 1 in Table 1 to give ¾-caprolactone (3b) in 68% yield
(Entry 1). In contrast to the reaction of 1a where the yield of 3a
was improved by adding a small amount of p-TsOH in step 2,
no such positive effect was observed in the case of the reaction
of 1b (Entry 2). Isopropyl alcohol (2b) was employed in place
A conceivable reaction pathway in the reaction of 1 is well
represented as follows (Scheme 2).
Phthalimide-N-oxyl radical (PINO) abstracts a hydrogen
atom from 2a to form 1-hydroxy-1,1-diphenylmethyl radical
(A). Under O2, the resulting radical A is readily trapped by O2
to give a hydroperoxide (C) which quickly liberates H2O2 and
benzophenone 4. For 1a,7a 1b,7b and 1c,7a it is well-known that
cycloalkanones 1 react with H2O2 to form 1,1¤-dihydroxydi-
cycloalkyl peroxides E probably through 1-hydroxy-1-cyclo-
alkylhydroperoxides (D).7 Therefore, it is believed that the
formed H2O2 reacts readily with ketones 1 to form D which
then undergoes the rearrangement to lactones 3 via E assisted
by HFIP. The most important feature of the present method-
ology is that highly reactive anhydrous H2O2 is generated in
situ from 2a and O2 by the action of the NHPI. Thus,