Study of the performance of a resolution criterion to characterise complex chromatograms with unknowns or without standards

Abstract

The search for best conditions in liquid chromatography is routinely carried out with information provided by chemical standards. However, sometimes there are samples with insufficient knowledge about their chemical composition. In other cases, identities of the components are known, but there are no standards available, and in other cases the identities of peaks in chromatograms taken under different conditions are ambiguous. Most resolution criteria used to measure separation performance cannot be applied to these samples. In this work, a global resolution function valid for all situations was developed based on automatic measurements of peak prominences (area fraction exceeding the line that joins the valleys delimiting each peak). Relative performance is evaluated against the peak purity criterion (which measures the area free of overlapping). Peak purity provides a truly comprehensive measurement of global resolution since the underlying signals for each compound are used. However, it is only accessible through in silico simulations. In contrast, peak prominences are not based on a comprehensive knowledge of the individual signals, and can be obtained from experimental chromatograms besides in silico simulations. Therefore, this criterion is suitable for direct evaluation of the resolution of chromatograms with high complexity. A comparison study was carried out based on the agreement of the gradients chosen as Pareto-optimal by both criteria, using information from standards of the 19 primary amino acids found in proteins. The developed global resolution function was applied with success to chromatographic fingerprints of medicinal herbs.