Non-linear transformations of factors in the generalized standard addition method

Abstract

Experimental designs and non-linearly transformed factors (concentrations of analytes) have been applied in the Generalized Standard Addition Method (GSAM). The relationship betweenN measured analytical signals,R(1),...,Rsu(n), and concentrations of standard additions, Δc1,..., Δcn, ofn analytes has been approximated by the following polynomial models: The regression coefficientsB in models (1) are calculated from thesimple formulae1, 3, 4.Functionsfi(l) (e.g. linear, parabolic, exponential, hyperbolic or logarithmic) are selected on the basis of the experimentally determined relationshipsR(l) (l=1, ...,N) vs.ci (i=1, ...,n). An example concerns the flame emission photometric determination of Na and Ca. 22 factorial and uncomplete second-degree modelsR(l) have been applied expressed by linear and hyperbolic functionsfi(l). The results of determination of Na by GSAM reveal significant improvement of accuracy as compared with the conventional single-component standard addition method.