Angewandte
Chemie
trans-Stilbene gave a 61% yield of product 18 as a 2:1
mixture of diastereomers. cis-Stilbene gave an identical
mixture but in a much lower yield of 21%. Cyclic substrates
such as 1-phenylcyclohexene (19, 27%), indene (20, 59%),
and dihydronaphthalene (21, 47%) could be used success-
fully. Preliminary studies show that the reaction is not limited
to styrenes: trans-dibenzoylethylene and allyl benzene pro-
vided the g-peroxyketones 22 and 23 in 47% and 12% yield,
respectively. The low yield of 23 suggests that a resonance-
stabilizing group on the olefin is beneficial. Vinyl pyridines
were completely unreactive, most likely because these basic
substrates deactivate the acid catalyst.
Further synthetic transformations of the g-peroxyketone
products were investigated with 1 as a model substrate
(Scheme 3a). In the presence of a catalytic amount of base
Scheme 4. Suggested mechanism of product formation.
the presence of a strong acid, tBuOOH adds to the ketone to
form 30 and subsequently alkenylperoxide 31. Such com-
pounds are reported to undergo facile homolytic bond
cleavage,[17c,d] which would generate the resonance-stabilized
ketone radical 32 and a tert-butoxyl radical 33. In the presence
of tBuOOH, a fast equilibrium exists between 33 and the tert-
butylperoxyl radical 34, favoring the latter.[17a] Addition of 32
to the styrene double bond would form 35, which reacts with
the peroxy radical 34 to give the final product 1. The selective
cross-coupling of 35 with 34 can be explained by the
“persistent radical” effect,[18] in that 34 is a persistent
radical[18b] and 35 a transient radical.
The observed selectivity for the linear versus branched
product (4a vs. 4b) could be explained by the rapid
decomposition of alkenylperoxides like 31, which gives the
less substituted ketone radical rather than the thermodynami-
cally preferred one owing to kinetic preferences. The
observations that Fe and Cu salts fail as catalysts and that
dimethyl malonate fails as a carbonyl component are both in
agreement with the proposed mechanism. Fe and Cu salts are
known catalysts for the decomposition of tBuOOH to
radicals, which can mediate reactions by hydrogen-atom
transfer (HAT) or by single-electron transfer (SET).[17a,19]
Also, the radical addition of activated methylene compounds
to olefins can be initiated by SET or HAT.[5,9] Accordingly, it
appears that neither SET nor HAT is operative in the present
system but rather the ketone radicals are formed directly. The
presence of Fe and Cu salts could shut down the formation of
30 and 31 by fast decomposition of tBuOOH while a malonate
could be too unreactive for the formation of ketene perketals
analogous to 31 under reaction conditions.
In conclusion, we have developed a multicomponent
alkylation–peroxidation reaction of styrene derivatives with
ketones by Brønsted acid catalysis, based on a previous
mechanistic study. The g-peroxyketone products thus
obtained can be further converted to synthetically useful
1,4-diketones, homoaldol products, and alkyl ketones. The
reaction is thus a valuable addition to the mostly metal-
mediated addition reactions employing activated methylene
compounds, which are less effective with simple ketones. The
mechanistic information gained by this study sheds further
light on the autoxidative coupling reaction with xanthene[1,2]
and could be useful for future developments.
Scheme 3. Further transformations of the g-peroxyketone products.
(Kornblum–DelaMare rearrangement[13]), 1 can be trans-
formed into the corresponding 1,4-diketone 24, an important
class of intermediates in the synthesis of heterocycles such as
pyrroles[14] and furans.[15] When 1 is hydrogenated using
palladium on carbon in acetonitrile, the corresponding
homoaldol product 25 is obtained. When the hydrogenation
is performed in methanol, complete reduction of the benzylic
position is observed, giving product 26. The reaction proceeds
via acetal 27, which can be isolated if the reaction is not
allowed to run to completion. The direct acid-catalyzed
transformation of 1,3-diketo-g-peroxides to furans has been
reported.[9i]
Finally, a one-pot synthesis of pyrrole 28, a nonsteroidal
antiinflammatory agent,[16] was performed (Scheme 3b). The
reaction of styrene and cyclohexanone as depicted in
Scheme 1 resulted in the formation of peroxide 9, which
was directly converted to the corresponding diketone by
adding one equivalent of DBU to the reaction mixture. After
removing acetonitrile, we followed the reported procedure[16]
and added aniline 29 together with acetic acid and heated the
mixture at reflux. Pyrrole 28 was obtained in 30% yield over
three steps using a single purification step and in 41% yield if
each product was purified before the next step.
Consistent with these results and previous investiga-
tions,[2,17] a plausible mechanism is shown in Scheme 4. In
Angew. Chem. Int. Ed. 2014, 53, 1 – 5
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3
These are not the final page numbers!