Wittig reaction salt effects

This accounts for the considerable discrepancy between the alkene Z/E ratio found on work-up and the initial oxaphosphetan ais/trans ratio. By approaching the problem from the starting point of the diastereomeric phosphonium salts (19) and (20), deprotonation studies and crossover experiments showed that the retro-Wittig reaction was only detectable with the erythreo isomer (19) via the cis-oxaphosphetan (17). Furthermore, it was shown that under lithium-salt-free conditions, mixtures of (19) and (20) exhibited stereochemical drift because of a synergistic effect (of undefined mechanism) between the oxaphosphetans (17) and (18) during their decomposition to alkenes. 

Heterocycles.—The phosphonium salt (59) is an effective three-carbon synthon, as demonstrated by its reaction with enolates of /9-keto-esters (Scheme 20) to give cyclopentenyl sulphides via an intramolecular Wittig reaction.63 Ylides are also intermediates in the synthesis of dihydrofurans (60) from the cyclopropylphos-phonium salt (61) and sodium carboxylates (Scheme 21).64 Cumulated ylides are very useful for the synthesis of heterocyclic compounds, e.g. (62), from molecules which contain both an acidic Y—H bond and a carbonyl or nitroso-function, as shown in Scheme 22.65... 

Phosphorus ylides have been reviewed and an intermediate betaine lithium salt adduct (stabilized by complexation with lithium ions and by the chelating effect of pyridyl ligands) has been observed spectroscopically for the first time during the course of a Wittig reaction. ... 

The preparation of norbomadiene-fused thiophene (17) involved a double-Wittig reaction of 1,2-dione 15 with the bis-ylide derived from phosphonium salt 16 <99BCSJ1597>. The effect of the fused heteroaromatic ring of 17 (neighboring group participation) on electrophilic substitution of the norbomadiene ring was examined. 

The carbonyl- 14C KIEs in the title reaction system (equation 225), which gives a nearly 50 50 mixture of cis-trans isomers, depends very much on the ylide used430, and indicate that the reactions proceed via cycloaddition TS of considerable nucleophilic character, inferred also from the substituent effects studied. Positive p values indicate that the Wittig reaction is nucleophilic in nature. Assuming as before the four-centered TS, the authors430 conclude that the C—C bond formation is much advanced of the P—O bond formation in the TS and that the carbonyl-carbon KIE are expected to be larger for later TS salt-free reaction410 (more reactant-like for Li salt present in reaction). 

Whereas the effects of Li salts on Wittig reaction stereoselectivity are well known, effects of sodium salts have received less attention. As in the case of the Li salts, the presence of Na salts tends to increase the TJE ratio in alkene formation. However, this effect is observed only in those cases where the sodium salts are soluble and available for complexation with the reactants (Nal and NaB(C6H5)4). In cases where the metal ion is sequestered by an additive such as 15-crown-5 or DMSO or when the metal salt is not soluble in the reaction media (NaCl, NaBr), lower Z/E ratios in alkene formation are observed. 

Using functionalized 3,4-dialkyl pyridines, Shiao and co-workers have described an alternative synthesis of actinidine (183) based on mixed metal reactions with l-(alkoxycarbonyl)pyridinium salts. Thus, ethyl 3-iodopropi-onate was successively treated with zinc and cuprous cyanide, and after being cooled to —78°C the resulting solution was reacted with 5-methyl-methylnicotinoate (158) and the mixture warmed to room temperature to afford 159. Cyclization and decarboxylation was effected with sodium hydride followed by heating in aqueous solution to afford 160. A Wittig reaction on 160 gave the olefin 161, and catalytic hydrogenation (Pd-C) afforded ( )-actinidine (134) (Scheme 5) (183). 

Some of the most profound lithium ion effects are observed in diethyl ether, benzene, or toluene, and the literature on Wittig reactions under these conditions contains a number of results that appear to be somewhat contradictory. The reason for this lack of consistency may be related to issues of lithium halide or betaine adduct solubility. Thus, Bergelson and Shemyakin et al. (5c, d) were able to show that ylide solutions prepared in benzene behaved differently if they were filtered to remove precipitated salts prior to use (compare Table 11, entries 3 and 4 entries 47 and 48). Later, it was shown that filtered toluene solutions of Ph3P=CHCH3 obtained from the phos-phonium bromide using butyllithium contain <0.1% bromide according to elemental analysis (18b). Therefore, the dramatic change in the alkene ratio... 

---