Optimization of Selectivity in High-Performance Liquid Chromatography Using Mixture-Design Statistical Techniques: Overview and Software for Data Analysis

Abstract

Optimization of separations in high-performance liquid chromatography (HPLC) is dependent on several factors: the proper choice of variables; a definition of criteria for a suitable separation; a rational strategy for examining the effects of these variables; and an efficient analysis of the data using manual methods or computer software. The selection of variables and, strategy for optimization is discussed in this paper, and a detailed description of a versatile software package for data analysis is presented along with provisions for obtaining the actual FORTRAN source code. The original mixture-design statistical software has been expanded and enhanced to accomodate data from both isocratic and gradient elution separations, single or multiple stationary phases, and other variables, such as temperature or pH. The software can be used to predict a single optimum set of conditions for a particular separation, and also to calculate the actual separation quality under any set of conditions bounded by the original experimental data. Examples of these functions are illustrated.