The use of poly(sodium N-undecanoyl-L-leucylvalinate), poly(sodium N-undecanoyl-L-leucinate) and poly(sodium N-undecanoyl-L-valinate) surfactants as chiral selectors for determination of enantiomeric composition of samples by multivariate regression modeling of fluorescence spectral data.

Steady-state fluorescence spectroscopy was employed to investigate the use of chiral polymeric surfactants as chiral selectors in chiral analysis by multivariate regression modeling of spectral data. Partial-least-squares regression modeling (PLS-1) was used to correlate changes in the fluorescence spectral data of 1,1'-bi-2-naphthol (BOH), 1,1'-binaphthyl-2,2'-diamine (BNA), or 2,2,2-trifluoroanthrylethanol (TFA) in the presence of poly(sodium N-undecanoyl-L-leucylvalinate), poly(sodium N-undecanoyl-L-leucinate) or poly(sodium N-undecanoyl-L-valinate) as the enantiomeric composition of the chiral analytes was varied. The regression models produced from the spectral data were validated by determining the enantiomeric composition of independently prepared test solutions. The ability of the model to correctly predict the enantiomeric composition of future samples was evaluated using the root-mean-square percent-relative error (RMS%RE) of prediction. In terms of RMS%RE, the ability of the model to accurately predict the enantiomeric composition of future samples was dependent on the chiral analyte, the polymeric surfactant used, and the surfactant medium, and ranged between 1.57 and 6.10%. Chiral analyte concentrations as low as 5 x 10(-6) M were found to give regression models with good predictability.