In order to develop an enantioselective version of the cyclisation described above, it was necessary to use chiral catalysts. For this reason, a variety of chiral primary amines were evaluated instead of achiral pyrrolidine.
Primary amines were found to be more efficient aminocatalysts. It was observed that a cyclohexyl diamine derivative provided the best results in terms of enantioselectivity and further catalysts were designed, built and tested to achieve the highest performance. Accordingly, several aryl substituents were introduced on the catalyst structure instead of the phenyl rings, such as 3,5-bis(trifluoromethyl)phenyl, amongst others. An inverse relationship between reactivity and enantioselectivity was typically found with these catalysts, which means that the most selective catalyst was also the least reactive. The highest levels of enantioselectivity achieved at this point was above 56% ee, but the yield of the product was still poor.
Intensive exploration of the reaction parameters included the use of different solvents, concentrations, temperatures, reaction times, additives and monodentate to multidentate phosphine ligands. These experiments led us to discover that most solvents did not modify the enantioselectivity of the process, but also did not improve the yield. Nevertheless, the use of multidentate phosphine ligands did actually improve the reactivity and did not erode the enantioselectivity of the process.
A significant amount of chiral and achiral phosphines were studied to further improve both selectivity and reactivity, and a bidentate phosphine was identified as optimal ligand. It was observed that this particular phosphine led to optimal results in terms of both reactivity and selectivity.
With the optimal conditions in hand, the scope of the new reaction was explored to develop the methodology further.
