On the Beneficial Effect of ortho-Methoxy Groups in the Asymmetric Ring Opening of meso Epoxides with Silicon Tetrachloride Catalyzed by Chiral ortho-Methoxyphenyldiazaphosphonamide Lewis Bases

Abstract

We published in this journal a communication on the enantioselective opening of meso epoxides with silicon tetrachloride in the presence of a catalytic amount of chiral phosphonamides.1 In that paper we reported the asymmetric ring opening of cyclooctene oxide with SiCl4 in the presence of 10 mol % of chiral phophonamide base (R,R)-1 as a catalyst (Scheme 1). For this reaction, we described up to 83 % yield and enantiomeric excess higher than 99 %. This spectacular result attracted the attention of Professor Scott E. Denmark, who pioneered this research field,2 because it was found that cyclooctene oxide is one of the slowest acting epoxides in this catalytic system and produced the racemic chlorohydrin. However, Denmark et al. did not succeed in reproducing the results we disclosed and submitted a rebuttal to our paper.3 We were then contacted by the Editor and Prof. Denmark and prompted to elucidate the origin of such a striking discrepancy. At this stage, I asked my co-workers4 to check the experimental procedures and analytical conditions and to try to reproduce the described results. I was provided with experimental data and material that fully confirmed our original claims. On the basis of these results we sent a first answer to the rebuttal published in Angewandte Chemie, an answer that could not explain why our observation differed from Prof. Denmark's. One month later, we received a letter from the Editor who required further complementary analytical data and information. When I had these new data from my co-workers in hand, I noticed several inconsistencies with the analytical material they had previously provided. Therefore, I decided to reproduce personally and independently the whole experimental procedures from the very beginning. I was unable to observe the opening of cyclooctene oxide under the described conditions and to reproduce the analytical data (for example, NMR, retention times in chiral GPC and HPLC analyses) provided by my co-workers. Instead, I obtained results similar to those reported by Denmark et al. in their rebuttal.3 A cross-check undertaken by several colleagues led to identical results. To date, my co-workers4 have been unable to provide an appropriate scientific rationale for the non-reproducibility of the former results and analyses. Under these circumstances, although our communication presented several other meso epoxide openings with fair to high enantioselectivities, which require experimental verification, I wish to withdraw that communication.1 Other pieces of work based on related experimental results that were published elsewhere will also be withdrawn.5, 6