Global least-squares analysis of large, correlated spectral data sets and application to chemical kinetics and time-resolved fluorescence

Publisher Website
Google Scholar
Abstract
A general, nonlinear least‐squares, direct‐search strategy for the analysis of the time‐dependence of single or multicomponent spectral data sets is described. The generic procedure is named component resolved (CORE) spectroscopy, and has previously been successfully applied to the special case of Fourier transform pulsed field gradient spin‐echo NMR data (FT‐PGSE). The main purpose of the CORE processing is to improve the quality of evaluated data through its intrinsic S/N enhancement, and to confidently allow studies on multicomponent data sets characterized by either or both of (a) extensive spectral overlap and (b) only minor differences in (for example) individual component kinetic decay rates. The generality of the CORE approach is here illustrated through examples from chemical kinetics and time‐resolved fluorescence. Unlike previous strategies for the same purpose, CORE can easily be adapted to a large variety of data set types, that can be of almost unlimited size. As a consequence of its direct approach to the underlying problem, the minimization approach is also perfectly stable. The generic CORE strategy also does appear to lend itself nicely to parallel processing, which would speed up the data processing by one or two orders of magnitude in future implementations.