AN EXPERIMENTAL-DESIGN APPROACH TO SELECTING THE OPTIMUM LIQUID-CHROMATOGRAPHIC CONDITIONS FOR THE DETERMINATION OF VITAMINS B-1, B-2-PHOSPHATE, B-3, B-6 AND C-PHOSPATE IN EFFERVESCENT TABLETS CONTAINING SACCHARIN AND SUNSET YELLOW FCF

The ion-pair liquid-chromatographic separation of the water-soluble vi tamins thiamine mononitrate, riboflavin phosphate, nicotinamide, pyrid oxine hydrochloride, ascorbic acid, saccharin, and the dye Sunset Yell ow FCF (E110) has been examined for the analysis of the compounds in e ffervescent tablets. A Draper-Lin small composite design was used to s tudy the impact on the compounds' retention times and peak-widths (the response variables) of four different mobile phase variables: percent age of methanol as organic modifier, the concentrations of hexanesulfo nate as ion pairing reagent and of triethanolamine as competitive base , and pH. The interactions of these variables were also studied. The p roposed design enabled derivation of multiple linear regression equati ons to model each compound's retention time and peak-width at half-hei ght. The statistical reliability of the regression models was establis hed by comparing predicted and experimental values. By introducing the regression models into a spreadsheet program (Excel 5.0), retention t imes and peak-widths for each compound were calculated at fixed mobile phase pH. The values of all the other combinations of the three mobil e phase variables were changed in increments of two units within their examined boundaries, resulting in 225 different rows. For each combin ation the compounds' calculated retention times and their correspondin g peak-widths were sorted in increasing order and the resolution betwe en successive peaks was calculated. The minimum effective resolution ( R-S min) between each pair of peaks and the maximum retention time (t( R) max) in each row were then selected and used to construct contour p lots indicating the location of zones of mobile phase parameter combin ations where R-S min > 1.5 and the analysis was rugged, and where t(R) max values were minimum. Their common regions resulted in optimum chr omatographic separations. Examples are shown of chromatographic separa tions obtained using mobile phase combinations which were the best com promise of the three criteria selected.