"Peak-shape-learning Mode" For Resolution And Quantification Of Overlapping Chromatographic Peaks

Chromatograph and hyphenated chromatographic instruments have become efficacious means for qualitative and quantitative analysis of complex samples.1-4 However,it is sometimes difficult and time-consuming to search satisfying separating conditions for complex systems.What’s more,when two or more components share very similar physical and chemical properties,physical separation of these components can’t be completely accomplished even under an optimized separation condition,and then overlapping peaks will occur,which is a big obstacle to the qualitative and quantitative analyzes.Thus it’s imperative to propose powerful theories and methods to resolve overlapping peaks.Many theories and methods have been developed for this problem.Perpendicular method and tangent method are two of the most popular methods.They can be used in chromatography workstation to resolve overlapping chromatographic peaks.However,the scope of these two methods is small.Once the two methods are used beyond their range,the quantitative results of overlapping peaks will have large error.To meet the need for accurate analysis of overlapping peaks,a lot of chemometrics methods have been proposed,such as curve fitting,the Fourier transform,wavelet transform,and so on.The process of these methods is complex and time-consuming.And the selection of function and model parameters has great influence on the result of the resolution of overlapping signals even lead to wrong results.In this paper,a novel method,peak-shape-learning mode,was proposed to resolve and quantify overlapping peaks.Both simulated and experimental data were investigated with this method.And the results were compared with these obtained by translation modified iteration(TMI).It can be found that the quantitative precision and peak-shape-learning mode is superior to that of TMI.And the scope of application of peak-shape-learning mode is wider than that of TMI.