Evaluation of chemometric methods for determining the number and position of components in high-performance liquid chromatography detected by diode array detector and on-flow 1H nuclear magnetic resonance spectroscopy

Abstract

Exploratory data analysis was performed on DAD-HPLC and on-flow HPLC-NMR data sets to determine the number of components (rank) in the mixture, and their elution times. The performance of principal component analysis, correlation plots, Euclidean distance, derivative plots, orthogonal projection approach, SIMPLISMA (SIMPLe-to-use Interactive Self-modelling Mixture Analysis) and evolving principal component analysis were compared on a mixture of three regio isomers of dihydroxynaphthalene. It was shown that derivatives and SIMPLISMA performed better with HPLC-NMR data than the other methods. OPA and evolving PCA performed best in DAD-HPLC. The relative performance of each chemometric method differs for each analytical technique, suggesting that criteria are dependent on instrumental method.