Automated Peak Recognition from Photodiode Array Spectra in Liquid Chromatography

Publisher Website
Google Scholar
Abstract
The use of multivariate computation techniques for the treatment of data from multi-wavelength photo diode-array detectors is an important step in computer-guided optimization strategies in HPLC. The treatment should provide the user with data on the identity and retention times of the individual solutes. These data are essential for those optimization strategies that can be classified as “interpretive”-methods, since peak recognition of both pure peaks and overlapping peak systems is a prerequisite for these methods. The peak recognition is based on comparison of spectra using their correlation coefficient. Three multivariate techniques used for deconvolution of overlapping peaks and determination of spectral information on the individual components are reviewed in this paper: Multi - component analysis, Target Factor Analyhsis and Iterative Target Transformation-Factor Analysis. The nature of the information and the quality of the results of the different multivariate techniques strongly influences the final result of the optimization procedure.