Improved selectivity in multicomponent determinations through interference modelling by applying partial least squares regression to kinetic profiles

Abstract

Binary and ternary mixtures of cysteine, tyrosine and tryptophan were used to show how kinetic methods together with partial least squares regression (PLSR) can be used to overcome interferences in multicomponent determinations. Two methods were compared: an equilibrium method which uses the characteristic UV spectra of amino acids and a kinetic method which follows the reaction kinetics of the oxidation of the amino acids by hexacyanomanganate(IV). In both cases, multi-wavelength detection with a diode-array spectrophotometer was used. The results obtained from the two methods are critically compared. Lower concentrations of tyrosine and tryptophan can be determined by the equilibrium method because of their high absorptivities. On the other hand, the kinetic method, in conjunction with PLSR, can be considered an example of how selectivity can be improved. This allowed the resolution of mixtures of tyrosine and tryptophan in the presence of several complex spiked matrices (aqueous extracts of feed samples that interfere to different extents) without any previous separation.