Sigmatropic rearrangements of the enolate of alpha-(allyloxy)-propiophenone and allyl aryl ethers: A kinetic study.

Abstract

In the first part of this thesis, the competition between the (3,3) Claisen rearrangement and the (2,3) Wittig rearrangement in the rearrangement of the enolates of $\alpha$-(allyloxy)-propiophenone was examined. It was found that the addition of solvating agents such as 18-crown-6 ether, which presumably increases the separation of the enolate from the cation, promoted formation of the (2,3) Wittig rearrangement product, however, the ratio was not large enough to merit synthetic utility. Also investigated was the effect of various substituents (methoxy, hydroxy, silyloxy, potassium salt of the hydroxylated allyl aryl ether and the "naked" oxygen anion, obtained by desilylation of the silyl ether with tris(dimethylamino)sulfur (trimethylsilyl)difluoride (TAS-F)) on the rate and mechanism of the (3,3) sigmatropic rearrangements of allyl aryl ethers. It was found that these electron-donating substituents have the greatest effect on the rate of rearrangement when at the ortho position of the aryl ring and virtually no effect at the para position. The rate of rearrangement increases according to the electron-donating ability of the substituent in the order: H $\sim$ OCH$\sb3$ $<$ OSiR$\sb3$ $<$ OH $<$ O$\sp-$K$\sp+$ $<$ O$\sp-$TAS$\sp+$ for the ortho- and meta-substituted allyloxybenzenes. The effect of a substituent at the para-position was found to be very minimal with only the naked oxyanion causing any rate increase over that of allyloxybenzene. Product characterizations are reported along with product rate constants. Heats of formation and geometries of the transition states (TS) of the rearrangements of allyloxybenzene and 2-allyloxyphenol, as predicted by a semi-empirical molecular orbital study, are also reported. The two methods used to locate the TS, MNDO and AM1, provided different heats of formation and transition state geometries: those predicted by MNDO are TS with very little C-O bond breakage ($\sim$15%) and a great degree of C-C bond formation (50-60%), corresponding to the previously postulated cyclohexadiyl-like TS; and those predicted by AM1 are much looser TS where the C-O bond breaking is much more advanced ($\sim$80%) whereas little C-C bond formation ($\sim$20%) has occurred. This aromatic-like TS predicted by the AM1 method is in agreement with the reported TS predicted from kinetic isotope studies. (Abstract shortened with permission of author.).