Pitfalls in the Optimization of Conformer Populations to Maximize the Similarity between Predicted and Experimental Chiroptical Spectra.

Abstract

The conformational populations of pantolactone, epichlorohydrin, and N-acetyl-tryptophan methyl ester were investigated by using similarity analysis between their calculated and experimental chiroptical spectra. By performing the analysis on pantolactone using two different chiroptical methods, namely, vibrational circular dichroism and Raman optical activity, it was found that the optimal sets of conformers do not match between the two methods, indicating that the conformational populations obtained by optimizing the similarity between calculated and experimental spectra are unlikely to be more accurate than energy-based Boltzmann populations. Also, it was found for pantolactone, epichlorohydrin, and N-acetyl-tryptophan methyl ester that the similarity between calculated and experimental spectra would often not vary significantly if each of the populated conformers was discarded, one at a time. This observation indicates that more than one set of conformers can provide acceptable similarity between the predicted and experimental spectra. Therefore, the correct set of conformers cannot be accurately determined by similarity analysis.