New routes to chiral phosphines

Abstract

The reaction of diphenyl(2-trimethylsilylethyl)phosphine with methyl trifluoromethanesulphonate gave a phosphonium salt from which the side-chain was cleaved with F ® , yielding diphenylmethylph osphine. A slower reaction with the triphenylsilyl-analogue gave the corresponding phosphine oxide. When the cleavage of a diphosphonium salt was attempted the product was a mixture of 1,3-bis-(diphenylphosphino)propane and diphenyl(2-trimethylsilylethyl)phosphine in 1:2 ratio showing the cleavage to be unselective. ß-Silylphosphonium salts derived from 2,2-dimethyl-H-methylene- 1 , 3-dioxolane reacted with F® to provide completely selective cleavage of the dioxolanyl side-chain, and a related bis-phosphonium salt gave only diphenyl(2-trimethylsilylethyl)phosphine under similar conditions. Greater selectivity towards fragmentation of the silylated side-chain was achieved with triethoxysilylethyl derivatives and 1, 3-bis-(diphenylphosphino)propane as well as 1,4-bis-(diphenylphosphino)butane were prepared in this way, in 86% and 9^% yield respectively. It was found that the dioxolanylmethyl side-chain was more readily cleaved from phosphonium salts than was 2-(triethoxylsily1)ethyl, however. In similar manner, reaction of 2-(trimethylsilyloxyethy1)- methyldiphenylphosphonium trifluoromethanesulphonate with Fö gave methyldiphenylphosphine oxide. The same reaction occurred when the corresponding hydroxyethylphosphine was treated with KH. This method was shown to be applicable to the synthesis of biphosphine oxides. Nevertheless, application to dioxolanylmethyl phosphonium salts again led to px^ef erential loss of that side-chain; this caused a change to utilisation of the trans-1,2-dimethylcyclobutyl backbone. The optically active o-anisylmethylphenylphosphine oxidewas prepared in both enantTomeric forms by a literature route. Reaction of the R-isomer with (iPr)2NLi and then acetone led to a new phosphine oxide which was reduced with E t 3N/Cl3$iH and reacted with methyl trifluoromethanesulphonate to giveR - (2-hydroxy-2-methylpropyl)(o-methoxyphenyl)methylphenylphosphonium trifluoromethanesplphonate. This reacted with KH to give the parent phosphineOKas R-enantiomer, optically pure within experimental error. “ A new route for the resolution of trans-cyclobutane-1,2-dicarboxylic acid has been developed. This involved the preparation of the bis-ephedrinium amide and its separationinto diastereomers by silica chromatography. The enantiomers were converted into SS and RR t r a n s - 1 ,2-cyclobutanedimethanolof 100% and 88% optical purTTy respectively. Whilst it proved possible to prepare the racemic bis-trifluoromethanesulphonate, it was an unstable compound and the optically active analogues could not be isolated in several attempts.