Synthesis and carbon-carbon bond-forming reactions of tungsten, molybdenum, and rhenium enolates

Alkylation of sodium cyclopentadienyl(tricarbonyl)metal— lates of tungsten and molybdenum with o—chloro ketones and c—chloro esters gives stable tungsten and molybdenum enolates (23a—23d). Similar alkylation of sodium pentacarbonylrhenate provides rhenium enolates 24. Compounds 23 lOse carbon monoxide and rearrange to 3—oxaallyl complexes 25 upon irradia— tiøn. Compounds 25 react with benzaldehyde to give transition metal aldolates (26), which may be converted into the silylated aldol (27) and the corres ponding metal chloride (28) upon reaction with trimethylsilyl chloride. Rhenium enolate 27 undergoes thermal aldol reaction with benzaldehyde and Buffers exchange of one carbonyl ligand upon being heated with triphenylphosphine. The resulting cia monophosphine complex (30) and compound 27 itself both react with triphenylphosphine in refluxing acet*nitrile to give complex 31, in which the enolate moiety has been transferred from rhenium to the nitrile function. Several schemes for establishing catalytic cycles based on the foregoing reactions are suggested. Since the first report of the aldol addition reaction in 1838 (ref. 1), enolate ions have occupied a position of singular importance in organic chemistry. A large fraction of the most general synthetic methods involve these important intermediates. Nevertheless, until about 30 years ago, there was little direct study of enolate ions s.e, since most reactions in which they are involved were carried out under protic conditions where enolates are formed only as transient intermediates. Obvious exceptions are the enolates derived from 13—dicarbonyl compounds, which may be prepared in alcoholic, or even in aqueous solutions. The situation began to change in 1949 and 1950 when Frostick and Hauser introduced diisopropylaminomagnesium bromide as a catalyst for the Claisen condensation (ref. 2) and Hamell and Levine reported the first use of lithium diiso— propylamide (LDA) for the same purpose (ref. 3). These strong bases have the useful property of being soluble in aprotic solvents such as ether and THF, and allow the stoichiometric production of enolate salts. The first report of a stoichiometrically formed enolate salt came from Dunnavant and Hauser, who prepared the enolate of ethyl acetate with lithium amide in ammonia and demonstrated that it reacts with aldehydes and ketones to give 3—hydroxy esters in low yield (ref. 4). The utility of such preformed enolate salts was graphically demonstrated by M. W. Rathke in an important paper published in 1970 (ref. 5). Rathke showed that the sodium enolate is unstable even at —78 0C, but that the lithium enolate, prepared with lithium biB (trimethylsilyl)amide in THF, is stable indefinitely at —78 °C, and that solutions of ethyl lithio— acetate react smoothly with aldehydes and ketones to give the corresponding aldols in high yield. In the last decade, we have seen an intense investigation of the preparation, structures, and chemical reactivity of preformed metal enolates. The reader is referred to several recent review articles for a more complete coverage of the field (ref. 6—11). A large part of this interest has revolved about the increasing importance of stereocontrol in organic synthesis. In this regard, there has been a great deal of success in understanding and using stereoselec— tive aldol addition (ref. 8, 9 & 11) and alkylation (ref. 10) reactions. In the context of the aldol reaction, two fundamentally different kinds of stereoselectivity are possible and have been investigated. In reactions