.
Angewandte
Communications
À
Scheme 6. Suggested C C bond cleavage through intramolecular
oxygen atom rearrangement.
MeBr could not be found by GC/MS. To examine this
hypothesis, the reaction was further investigated in 18O2
atmosphere. 18O-labeled products [18O]-2aa or [18O]-2ab
were detected, thereby indicating that one of the oxygen
atoms of the ester originated from molecular dioxygen
(Scheme 7). At the same time, the suggested pathway
shown in Scheme 6 is excluded.
Scheme 5. Isolation, transformation, and X-ray crystal structure of the
key intermediate 4a. Thermal ellipsoids are set at 50% probability.[8]
4a was a key intermediate in this transformation (Scheme 5,
I). Two control experiments without either K2CO3 or O2 were
performed. However, no product 2a was detected (Scheme 5,
II and III), and 4a was recovered totally. These results
strongly indicated that both K2CO3 and O2 are necessary for
the conversion of dimer intermediate 4a to the final product
2a.
Since the dimer 4a was a key intermediate, we presumed
that 1a should be converted to 4a without BnBr. To our
surprise, no desired product 4a was detected without BnBr,
and substrate 1a was decomposed completely (Table 2,
entry 1). Obviously, BnBr is essential to form dimer 4a.
Scheme 7. Isotopic labeling experiment with 18O.
Subsequently, a plausible pathway through [4+2] and
[2+2] cycloaddition of O2 was taken into consideration (see
the Supporting Information, Part V). Since singlet oxygen
could perform the [4+2] and [2+2] cycloaddition,[4] reaction
of 1a with singlet oxygen (generated from triplet oxygen and
light in the presence of TPP (tetraphenylporphyrin) as
a sensitizer) was carried out (see the Supporting Information,
Part V).[4d] The reaction performed with singlet oxygen
showed no improvement in both yield and reaction time
compared to that performed in triplet oxygen, thus inspiring
us to consider another pathway.
On the basis of the above results, a superoxide radical
mechanism for this transformation is illustrated in
Scheme 8.[5] In this transformation, substrate 1a could form
the dimeric intermediate 4a in the presence of halide and
base in an aerobic atmosphere via radicals B and C.[6] Also
a peroxyl radical of THF as a radical initiator has been
thought over (see the Supporting Information, Part VI), but it
was excluded, because the formation of 4a in this route must
be involved with BnBr according to entry 1 in Table 2. The
conjugated diene intermediate D was easily formed under
basic conditions and O2. Then auto-oxidation through reac-
tion of radical D with O2 could generate superoxide radical
E.[5] Further intramolecular cycloaddition to the ketone
would form the corresponding oxygenic radical F. Intermedi-
ate F would then capture a benzyl radical to give intermediate
G, and subsequent fragmentation of G would produce the
desired ester 2a and byproduct 5, which resulted in benzyl
formate 6.[7] To our delight, when the reaction was monitored
by GC/MS, the byproduct benzyl formate was unambiguously
detected, thus solidly indicating 6 was the byproduct of
oxidative cleavage of a-hydroxy ketone (Figure 1). Though
Table 2: Effects of halides in the transformation of 1a to 4a.[a]
Entry
Change from the “standard conditions”[a]
Results[b]
1
2
3
4
4
without BnBr
decomposed
decomposed
decomposed
2a: 65%
KI (4.0 equiv) instead of BnBr
KBr (4.0 equiv) instead of BnBr
BnCl (4.0 equiv) instead of BnBr
with TEMPO (2.0 equiv)
2a: 47%
[a] For the reaction conditions: Table 1, entry 18. TEMPO=2,2,6,6-
tetramethylpiperidine-N-oxyl. [b] Yields of isolated products.
Subsequently, KI and KBr also proved to be ineffective to this
transformation (Table 2, entries 2 and 3). However, when
BnCl was used, 4a could be observed by GC/MS, and 2a
could be isolated in 65% yield (Table 2, entry 4). Hence, we
conclude that not only K2CO3 and O2, but also organic halides
are necessary for this transformation to dimer intermediate
4a. By using the radical scavenger TEMPO (2.0 equiv), the
yield of 2a was sharply decreased from 80% to 47%, thus
indicating that this transformation may involve a radical
process.
The origination of the two oxygen atoms in ester 2a is
a crucial question for understanding the mechanism of C C
À
À
bond cleavage. First of all, a C C bond cleavage that occurs
through intramolecular oxygen atom rearrangement is sug-
gested in Scheme 6. Both atoms Oa and Ob are from substrate
1a, and MeBr is considered to be the byproduct. However,
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2012, 51, 12570 –12574