Mechanism of the Shi Epoxidation
The epoxidizing species is believed to be a dioxirane, which is a powerful epoxidation reagent. These are not indefinitely stable, but can be generated in situ by oxidation of a ketone with potassium peroxymonosulfate (Oxone). The sulfate - as a good leaving group - facilitates the ring closure to the dioxiranes. As the ketone is regenerated, only catalytic amounts of it are needed. In addition, chiral ketones can be used for a catalyzed, enantioselective epoxidation, since the ketone substituents are close to the reacting center.
Reactions are conducted in buffered, often biphasic mixtures with phase transfers catalysts. Addition of K2CO3 to the reaction mixture increases the rate of formation of the dioxirane but also lowers the stability of Oxone. However, a higher pH also disfavors the Bayer-Villiger Oxidation as a side reaction, so the catalysts remain more active. Therefore the autodecomposition of Oxone at high pH can be overridden if the ketone is sufficiently reactive. The enhancements in reaction rate can also be explained by a higher nucleophility of Oxone under more basic conditions. In any case, a careful use of buffered media is often needed.
The reactivity of the ketones can be increased by electron-withdrawing groups in the α-position. From early attempts at building active catalysts, it was learned that trifluoromethyl ketones improved the activity, but other electron-withdrawing groups can also be used. These factors also lower the rate of the Bayer-Villiger Oxidation. As ketones with a hydrogen in the α-position are prone to racemization, chiral elements have often been placed in other positions. Some early catalysts are shown here:
In 1996, a fructose-derived ketone was developed as a highly effective epoxidation catalyst. This ketone can be synthesized in two steps from the very cheap chiral starting material D-fructose by ketalization and oxidation. As L-fructose can be synthesized from L-sorbose, the enantiomer of this catalyst is also conveniently available.
In this catalyst, the stereocenters are close to the reacting center, so the stereochemical communication between substrate and catalyst is efficient. The presence of fused rings or quaternary centers α to the carbonyl group minimizes epimerization of the stereogenic centers. Electron-withdrawing substituents activate the carbonyl.
For this catalyst, the conversion of trans-β-methylstyrene at a pH at higher than 10 increased by 10 fold from that at a lower pH (7-8) and the enantioselectivity remained high (90-92%). This dramatic pH effect allows the use of a catalytic amount of the ketone. A pH of 10.5 can be conveniently achieved by adding K2CO3 as the reaction proceeds. This first catalyst allowed highly enantioselective conversions of trans-disubstituted and trisubstituted olefins, although the enantiomeric excesses remained low for cis-olefins and terminal olefins.
A spiro transition state seems to be favored due to a stabilizing oxygen lone pair interaction with the π* orbital of the alkene, which cannot be achieved in the planar transition state.
The main competing mode is the planar transition state shown; this is somewhat more favored with trisubstituted olefins if R' is bulky (a), whereas bulkier R substituents disfavor the planar transition state (b).
Later developments enabled the conversion of cis-substituted alkenes and terminal olefins by varying the substitution pattern of the catalyst. For example a Boc-protected lactam allows the conversion of cis-olefins.
Here, an interaction between groups with a π-system and the spiro oxazolidinone can be assumed, so conjugated styrenes and enynes give products in high enantiomeric excess:
A recent publication also shows selective conversions of terminal olefins. Here, the planar transition state is favored due to steric reasons. With an N-tolyl lactam ketone, the attractive interaction between aryl substituents of the olefin and the catalyst could be improved even further.
For historic developments, more ketone-based catalysts and the synthetic scope of the initially reported fructose-based catalyst please refer to a review by Frohn and Shi (Synthesis 2000, 1979. DOI). Recent developments can also be found in the subsequent literature section.