Curve Fitting to Chromatographic Peaks

in Frequency Domain

Pápai, Zsuzsa, Pap, Tamás L.

University of Veszprém, Department of Analytical Chemistry, 8201 Veszprém,

P.O. Box 158, Hungary

Curve fitting methods are very popular in the field of chromatography because the fitted function can be integrated in closed form or numerically to determine the peak shape parameters and the statistical moments. These characteristics are available for describing signal-time functions, but cannot inform us about the frequencies and phases of the signal constituents.

The Fourier transforms of chromatograms have been widely used for noise filtering, interpolation, curve resolution, etc.

In this work the Fourier transformed form - calculated by us - of the function [1] suitable for describing chromatographic peaks is presented. The variation of the real and imaginary parts of Fourier transform of the function [1] due to peak shape changes was examined. We worked out a mathematical process, during which Fourier transforms of the chromatographic peaks are calculated, then curves are fitted separate to the real and imaginary parts of the Fourier transform. The process is presented using real HPLC chromatograms.

This method is very useful to perform parameter estimation on the basis of the Fourier transform of the signal. The recursive parameter estimation is a spectacular means to identify different parts in the frequency domain, characteristic for a given element of the total signal.

Literature:

[1] Pap, T. L., Pápai, Zs., J. Chromatogr. A., 930, (2001) 53.