Computer assisted optimization of liquid chromatographic separations of small molecules using mixed-mode stationary phases

Mixed-mode stationary phases are gaining adepts in liquid chromatography (LC) as more and more applications are published and new commercial columns appear in the market ought to their ability to retain and separate analytes with multiple functionalities. The increased number of adjustable variables gives these columns an enhanced value for the chromatographer, but, on the other hand, it complicates the process of developing satisfactory separations when complex samples must be analyzed. Thus, the availability of computer assisted methods development (CAMD) tools is highly desirable in this field. Therefore, the first specific tool for the CAMD of LC separations in mixed-mode columns is presented. The tool consists in two processes. The first one develops a retention model for peaks in a predefined experimental domain of pH and buffer concentration. In this domain, the retention as a function of the proportion of organic modifier is modeled using a two-stage re-calibration process departing from isocratic retention data and then, from gradient elutions. With this two-stage approach, reliability is gained. In the second process, the model is finally interpolated and used for the unattended optimization of the different possible elution modes available in these columns. This optimization process is driven by an evolutionary algorithm. The development and application of this new chemometrics tool is demonstrated by the optimization of a mixture of neutral and ionizable compounds. Hence, several different types of gradients were generated, showing a good agreement between simulated and experimental data, with retention time errors lower than 5% in most cases. On the other hand, classical CAMD tools, such as design of experiments, were unable to efficiently deal with mixed-mode optimizations, rendering errors above 30% for several compounds. (C) 2012 Elsevier B.V. All rights reserved.