Highly enantioselective ylide-mediated synthesis of terminal epoxidesw

The importance of chiral epoxides as synthetic building blocks in asymmetric synthesis is extremely difficult to overstate. While the highly enantioselective Sharpless, and JacobsenKatsuki, protocols for the epoxidation of internal alkenes are nowmature technologies of inestimable value, the conversion of terminal alkenes to enantioenriched 1-oxiranes has proven a considerably more difficult process to control. Arguably the current method of choice (in terms of product ee) for the catalytic synthesis of these molecules is the Co-salen complexcatalysed kinetic resolution of racemic epoxides. While progress towards the efficient asymmetric oxidation of terminal alkenes has been made – e.g. Fe/Mn-porphyrin complexes, chiral Ti and Pt-complexes and chiral dioxiranes (styrene substrates only), the difficulties in preparing terminal epoxides in 490% ee via alkene oxidation has fostered interest in alternative protocols. One methodology which holds promise is catalytic methylene transfer to aldehydes mediated by sulfonium ylides. Since it is often from the aldehyde that the alkene substrate for oxidation processes is prepared, a methylene transfer reaction would represent a more direct synthesis of the product, which could potentially be performed in an operationally simple manner, in a transition metal-ion free environment. The stabilisedand semi-stabilised sulfonium ylide-mediated asymmetric epoxidation of aldehydes catalysed by chiral sulfides has proven a highly useful process for the formation of 1,2-disubstituted epoxides with excellent product diastereoand enantioselectivity. In an unfortunate parallel to the alkene oxidation methodologies, the corresponding sulfonium ylide-mediated aldehyde oxidation to form terminal epoxides (nearly 40 years after the first disclosed asymmetric attempt) is characterised by moderate yields and low-moderate levels of product enantiomeric excess. For instance, both benchmark methodologies (A and B, developed by Goodman and Aggarwal respectively, Scheme 1) for the conversion of the archetypal substrate benzaldehyde (1) to styrene oxide (2) involve the employment of (super)stoichiometric loadings of a chiral sulfide and a Simmons–Smith type Zn-carbenoid, and furnish the product in o60% yield and ee. In an attempt to develop a more atom-economic process, we have shown that the ylide can be generated via an alkylation and subsequent deprotonation route (C, Scheme 1). Sulfide 6 could be utilised at 20 mol% loading if the alkylating agent and base were added portion-wise, however no progress was made towards improving upon the mediocre enantioselectivity which bedevils this (otherwise) potentially very useful reaction. With the goal of solving this problem, we reflected upon the likely causes of the low enantioselectivity. The first major difficulty as Aggarwal has pointed out associated with epoxidation using unstabilised ylides is irreversible betaine formation. Thus the enantioselectivity is derived from the face-selective addition of the ylide to the aldehyde alone. Our thinking behind the identification of the second problem, i.e. the root-cause of the poor facial selectivity which results from the use of the C2-symmetric catalyst 6, is outlined in Fig. 1. It is assumed that the aldehyde is likely to approach the ylide in such a way that: (a) it avoids the large catalyst substituent in the b-orientation (as drawn, Fig. 1A and B), (b) charge separation/ gauche interactions are minimised in the transition state, and c) that the large aryl aldehydic substituent is directed into the solvent. In this scenario one can see that attack at the aldehyde si-face (leading to the observed (R)-product enantiomer, Fig. 1B) appears favourable to attack at the re-face (Fig. 1A) due to an unavoidable steric clash between the carbonyl group (which increases as the tetrahedral betaine geometry is approached) and the large a-catalyst substituent. This may explain why enantioselectivity is unsatisfactory: one is attempting to act upon a relatively small steric discrepancy between the aldehydic Scheme 1 Benchmark methodologies for the asymmetric synthesis of styrene oxide (2) from benzaldehyde (1) by methylene transfer.