Noise studies for detection limits for some electrothermal atomic absorption determinations and calculation of the optimal detection limit from one atomization

Abstract

A simple statistical calculation, confirmed experimentally, shows that in Zeeman-effect electrothermal atomic absorption spectrometry the value of the integrated absorbance [graphic omitted] measured during a time t in the absence of neutral atoms can be broken down into two terms according to [graphic omitted] In this relationship mi is a constant, different for each atomization indicated by the index i, and x is a random value whose standard deviation can be calculated from the standard deviation and the number of measurements of the instantaneous absorbance. The calculation of mi allows one to decrease the error due to the integration of the blank signal and, therefore, to decrease the detection limit. If a non-negligible variation of the mi value is induced in the atomization step, calculation of mi after this step is more advisable. For the tested apparatus, the proposed post-atomization correction improves the accuracy of measurements and reduces the instrumental detection limit (IUPAC). An improvement by a factor of 5 is obtained, for example, for the detection limit of aluminium with an integration time of 3 s. Moreover, it has been shown for manganese and aluminium that the value of the detection limit obtained after correction is very similar to that calculated from a single atomization using the following equation resulting from statistical approximations: [graphic omitted] where s is the standard deviation of the instantaneous blank absorbances, Ts the sampling rate, tint the integration time, m the mass of metal injected and QA the corresponding measured integrated absorbance. The use of this equation, which is based on the measurement of the real experimental noise from one atomization, can offer an important saving of time.