Allylic alcohols: ideal radical allylating agents?

Abstract

Radical allylations represent effective routes to various alkenes, but to date they have relied chiefly on organostannane derivatives and still suffer from significant limitations with respect to the substitution pattern of the starting allylating agent. Indeed, while substituents at the β-position relative to the radical leaving group are well-tolerated, introduction of α-substituents induces a major complication due to the rapid and usually irreversible isomerization of the starting allylating agents. Although a number of research groups have made substantial efforts to develop heavy-metal-free radical allylations, methods compatible with α-substitution of the allylating agent are still scarce. Furthermore, quite a few systems are limited by the relative inaccessibility of the substrates. This Account summarizes our sustained efforts regarding the development of allylic alcohols into "ideal" radical allylating agents and presents published as well as some unpublished results. The systems we have developed combine the use of readily available xanthates and allylic alcohol derivatives under metal-free conditions to furnish not only alkenes but also aldehydes and saturated and unsaturated ketones through the virtually unprecedented homolytic cleavage of the normally strong C-O or C-C bond. The former route hinges on first converting the allylic alcohol into a 2-fluoro-6-pyridoxy derivative by reacting the corresponding alcoholate with 2,6-difluoropyridine, while the latter relies on attaching a cumyl group to the carbon bearing the free allylic alcohol. Either substrate is then exposed to the action of a suitable xanthate in the presence of a stoichiometric amount of a peroxide, usually lauroyl peroxide (DLP) in refluxing ethyl acetate or di-tert-butyl peroxide (TBHP) in refluxing chlorobenzene for the more difficult cases. Even though C-O or C-C bond homolysis leads to a stabilized 2-fluoro-6-pyridinyloxyl radical or a cumyl radical, respectively, the β-scission in both cases is relatively slow and at the lower limit of useful elementary radical steps. The kinetic barrier of the fragmentation can nevertheless be overcome because of the long relative lifetime of radicals generated by the degenerate transfer of the xanthate group, and this is a key element for success. This novel technology offers numerous advantages. The starting activated allylic alcohol derivatives are readily accessible in two steps from aldehydes or ketones. They can also be obtained by base-induced opening of epoxides. Numerous functional groups are tolerated under the mild reaction conditions for the radical addition-elimination, as nicely illustrated by over 150 examples of radical allylations, not all of which can be included in the present Account. In addition, substitution at both the α- and β-positions of the allylating agent is possible, a rare feature in this area.