Algorithms for time-dependent chromatographic peak areas: 1. Algorithms evaluated for fully resolved peaks

Abstract

This paper describes a preliminary study of an alternative method to fit chromatographic data. Peak-shaped responses are integrated to obtain unidirectional data for area versus time and an empirical model is used to fit the data. Results are reported for two similar models, each of which characterizes the total sigmoid function for each component in terms of the leading and trailing segments of two separate sigmoids. The use of leading and trailing segments of two sigmoids makes it possible to fit area chromatograms with large degrees of asymmetry. Performance of the models was evaluated by fitting them to data for single-component experimental and simulated data with degrees of asymmetry at 10% of peak height between 0.37 (fronting) and 7.33 (tailing). Results obtained using the area-based models are compared with results obtained by fitting a commercially available model for an exponentially modified Gaussian (EMG) to the peak data from which the area data were obtained. Areas computed using the two area-based models vary linearly with barbiturate concentrations over 100- to 200-fold concentration ranges and are more rugged to changes in peak shapes and asymmetry than the EMG model. Both area-based models give substantially smaller errors in peak area and smaller root-mean-squared fitting errors than the EMG model for peaks with moderate to high degrees of asymmetry.