Manganaoxetane intermediate

Ugly facts, unfortunately, sometimes invalidate a beautiful mechanism. The Jacobsen epoxidation sometimes proceeds with loss of configurational purity of acyclic alkenes. This feature of the reaction can be explained by invoking radicals. Homolysis of the Mn C bond in the manganaoxetane intermediate would give a Mn(III) 1,4-diradical complex, and attack of the alkyl radical on O with displacement of Mn(II) would give the epoxide and regenerate the catalyst. 

The problem with this mechanism is that the square planar geometry of the salen ligand is unlikely to allow both a Mn—O and a Mn—C bond on the same face of the complex. An alternative mechanism that does not involve a Mn—C bond can be proposed. Electron transfer from the alkene to the Mn(IV) 0X0 complex might give a Mn(III) radical anion and an organic radical cation. Addition of the oxide to the cation would then give the same Mn(III) 1,4-diradical seen before. Whether or not a manganaoxetane intermediate actually occurs in the Mn-catalyzed epoxidation of alkenes is very controversial. 

---