PROBLEMS INVOLVED IN THE CALCULATION OF THEORETICAL SENSITIVITIES OF FLAME ATOMIC-ABSORPTION MEASUREMENTS

Theoretical sensitivities cannot replace calibration. Nevertheless, th ey are important for standardless atomic absorption determinations, an d especially, for optimization of experimental conditions. Starting wi th fundamental processes involved in the interaction of electromagneti c radiation with the matter, a simple sensitivity formula was derived. S(A) = [dA(lambda(o))/dc(A)]o = K(n)(b/D)(rho/M(A))lambda(o)G The sen sitivity S(A) is the initial slope of the calibration curve absorbance A(lambda(o)) measured at the centre lambda(o) of the resonance line v s. concentration c(A) of the element to be determined (analyte A). The relative atomic mass M(A) of the analyte and the density rho of the s olutions need no further discussion here. The ratio (b/D) of the absor ption path length b to the dilution factor determines the sensitivity of determinations by AAS to a high degree and will be discussed in det ail. The numerical factor K(n) depends on the concentration unit and t he system of units (CGS or SI) used for the physical constant involved . The former were always used in earlier works in the field of atomic spectroscopy. In order to make a connection more easily between classi c (CGS) and modern (SI) formulations, the most important equations are given in both systems of units. The other important factors are combi ned in the group G-factors of dimension one. The comparison of the the oretical and experimental sensitivities may be transformed to the comp arison of theoretical and experimental G-values. Whereas the compariso n of sensitivities is always limited to an experimental set-up, the co mparison of G-values is released from this ''ballast'' and therefore w ill be preferred here. It is proposed that experimental G-values are c omputed from experimental sensitivities and used for compilations in t he future, because they are more precise and accurate than their cofac tors.