EVALUATION OF CALIBRATION METHODS FOR ZEEMAN GRAPHITE-FURNACE ATOMIC-ABSORPTION SPECTROMETRY USING COMPUTER MODELING

Computer modeling was used to compare calibration curves and relative concentration errors for normal, linearized, and three-field Zeeman GF-AAS. The model assumed that either photon shot noise or the combination of photon shot and analyte fluctuation noise were limiting and that the sole source of nonlinearity was stray light. For absorbance, the calibration range and the relative concentration error for all three methods are almost identical. The difference is a reduced-sensitivity curve for three-field Zeeman, which offers a relative concentration error advantage in the concentration region where the most sensitive curve rolls over. For integrated absorbance, the sum of absorbances over the analytical peak, linearized Zeeman provides a significant relative concentration error advantage over the other methods at the high concentration end of the calibration curve. The calibration range is effectively extended by at least 1.5 orders of magnitude. This advantage arises from integration of absorbances which have a linear relationship to concentration. At high concentrations, absorbances computed for normal and three-field Zeeman are nonlinear with respect to concentration. Three-field Zeeman offers no advantage over normal Zeeman for integrated absorbance.