AmericanChemicalSociety.com

During the summer of 2010 Sarah Border and Robert Parker worked on the Kekulene project.  Sarah tried many different approaches to the condensation reaction shown; she was able to reproduce the previous workers' best yield of 10% but was unable to improve on it.

One of the problems we had was that our stock of the starting ketoester 1 had been depleted, and attempts to make more were running into problems.  Previous students had run this reaction in a polar aprotic solvent, and workups that removed solvent also removed product.  Distillation also was less than satisfactory.  At the end of the summer, we finally developed a solventless approach that made the desired compound quite cleanly with little workup, just a distillation straight out of the reaction mixture.  We now have a large supply of clean starting material.

There are two problems we have identified with the above cyclization reaction.  The first is that the methyl ester generates methoxide in the reaction mixture, which can act as a nucleophile to trigger side reactions such as the Baylis-Hilman reaction.  We thought we could prevent this by using the tert-butyl ester as the starting material.  During the spring of 2010, Katerina Korch was able to make this ester in good yield by doing a reagentless Krapcho decarboxylation (polar aprotic solvent but no water or salt added).  Unfortunately, the desired reaction did not occur at all with the tert-butyl ester.

A second problem is that when treating the above mixture with base, there are many possible locations for deprotonation to occur, only one of which leads to the desired product.  We speculated that by making the enamine of the ketoester 1, we could possibly isolate the enamine corresponding to the correct site for reaction, or, at worst, limit the reaction to two sites.  In fact, during the summer of 2010, Robert Parker was able to make the desired enamine, but could not separate the two isomers from each other.  He was also unable to make the enamine react with cyclohexenone.  We are now trying to make the enol silyl ether of 1, again assuming that on treatment with fluoride, it will make the desired enolate and its isomer but not other enolates.