Multiple absorbance ratio correlation — a new approach for assessing peak purity in liquid chromatography

Abstract

The theory of multiple absorbance ratio correlation (MARC), a novel method for examining chromatographic peaks, is presented. The MARC hypothesis provides a method for examining chromatographic peaks for homogeneity, identity and purity, operating on multi-wavelength data sets which do not contain full spectra. Paraxanthine (PX) is an endogenous caffeine metabolite which is spectrally similar (r 0.99) to theophylline (TH) and which often coelutes with it. Multi-wavelength data, collected using an HP1040A diode-array detector, for a series of binary mixtures of TH and PX (0.7 < R s < 2.9; 0–50% PX), was analysed using MARC. The algorithm calculates a correlation factor between an internal (for homogeneity) or an external (for identity) reference 5-point vector and a similar 5-point vector acquired at sequential time intervals throughout the peak being investigated. For samples where the binary peak was homogeneous, limits of detection for PX of 1% (w/w) were obtained. This compared with limits of 5–10% (w/w) for the standard absorbance ratio (AR) method. MARC was found to be less sensitive to wavelength choice than AR. The sensitivity of MARC was independent of the chromatographic resolution.