Epoxide of a ketene acetal. The first 2,2-dialkoxyoxirane to be isolated

Dimethoxycarbene, generated at 110 degreesC in benzene by thermolysis of 2, 2-dimethoxy-5,5-dimethyl-Delta (3)-1,3,4- oxadiazoline, reacted with cycloh exanone to afford 2,2-dimethoxyspiro[2.5]-1-oxaoctane. It is the first oxir ane of a ketene acetal that could be isolated and characterized by H-1 NMR- , C-13 NMR-, and IR spectroscopy. The identical oxirane might be expected f rom conrotatory cyclization of the appropriate carbonyl ylide. That ylide w as generated under identical conditions by thermolysis of an oxadiazoline p recursor (3,4-diaza-2,2-dimethoxy-1-oxaspiro[4.5]dec-3-ene) (14). The ylide could either cyclize or fragment to dimethoxycarbene and cyclohexanone. Ad dition of 4-tert-butylcyclohexanone, to trap dimethoxycarbene in competitio n with the cyclohexanone generated from 14 and, to leave the ylide closure pathway as the only route to the oxirane, showed that the carbonyl ylide do es cyclize. However, fragmentation of the carbonyl ylide is relatively fast compared to its cyclization and most of it fragments to dimethoxycarbene a nd cyclohexanone. Oxirane formation from the carbene and ketone is probably either a concerted cycloaddition or it occurs in two steps, by nucleophili c attack at the carbonyl carbon to form the C-C bond first, prior to ring c losure. If the carbene is bonded first to O of the carbonyl group, as it is in the carbonyl ylide, subsequent formation of the C-C bond is too slow, r elative to fragmentation of the ylide, to afford the oxirane ring efficient ly.