Multicomponent analysis in clinical chemistry by use of rapid scanning fluorescence spectroscopy.

Abstract

To be useful in the clinical laboratory, multicomponent fluorescence analysis requires both the rapid measurement of the fluorescence intensity at a variety of excitation and emission wavelengths and the unambiguous reduction of the data by efficient algorithms. The Video Fluorometer, which exploits the multi-channel capability of a low-light-level television sensor to simultaneously acquire excitation and emission spectra, can meet the first requirement. For example, a complete set of emission and excitation spectra for perylene can be obtained in less than 2 s at concentrations of 10(-10) mol/liter. To meet the second need, we present two types of data-reduction strategies: (a)a least-squares fit to the data, with use of the spectra of previously determined compounds likely to be present; and (b)a determination of the eigenvalues and eigenvectors fo the fluorescence matrix, from which the number of components and the possible spectra of each can be estimated. Computer simulations of the least-squares fitting algorithms show that five strongly overlapping components can be determined in the presence of noise with an accuracy of better than 5%. Also, a fluorescent sample containing two species with very similar but unknown spectral properties can be resolved to obtain the spectrum of each.