PARTIAL LEAST-SQUARES REGRESSION FOR MULTICOMPONENT KINETIC DETERMINATIONS IN LINEAR AND NONLINEAR-SYSTEMS

The potential of partial least-squares regression (PLS) for resolution of binary mixtures by kinetic procedures is evaluated, both in a line ar system as in the presence of a synergistic effect. Kinetic first or der curves for the variation of the fluorescence intensity at a single wavelength were simulated and the effect of instrumental noise, exper imental reproducibility, the analyte concentration ratio and the rate constant ratio was studied. Instrumental noise was modelled from measu rements with a commercial spectrofluorimeter. To simulate experimental reproducibility, rate constant values were randomly chosen from norma l distributions with the rate constant nominal values as mean and diff erent standard deviations. Synergistic effects were modelled by adding a multiplicative term in the reaction rate of the product formation. Results show that PLS is an excellent calibration method to resolve mi xtures by first-order kinetic procedures, without the need of any prev ious knowledge about rate constant values. It is also shown that it is possible to resolve mixtures in the presence of synergistic effects i f the chemical system fulfil some conditions, mainly the rate constant ratio be higher than three. To show the potentiality of PLS on an exp erimental case it has been applied to the kinetic-spectrofluorimetric resolution of mixtures of hydrazine and hydroxylamine using as reactiv e the 2-hydroxybenzaldehyde azine. This chemical system is known to be subject of an important synergistic effect.