|Abstract or Summary
- A procedure for evaluating the precision of atomic absorption
measurements is presented and applied to 21 elements under varying
instrumental conditions. The experimental relative standard deviations
(RSD's) in absorbance obtained from repetitive measurements made
with a high resolution voltmeter and on-line computer, are compared
to those predicted by a recently proposed theoretical equation. Good
agreement between theory and experiment is obtained which indicates
the usefulness of the theoretical equation and evaluation procedure.
The study reveals that, under the conditions used for atomic absorption
measurements for the 21 elements, the noise associated with the
lamp (i.e., signal shot noise and source flicker noise) and with the
flame (i.e., flame transmission flicker noise) limit the precision at
small absorbances (i.e., A < 0.1 ). Flame transmission flicker noise
generally increases with the flame absorbance.
Over most of the analytically useful range of absorbances (i.e.,
0.1 > A > 1.0-1.5) analyte absorption flicker limits the precision to
a value near 1% RSD for one second integration periods. The precision
can be improved by a factor of two to three for any absorbance region
with ten second integration periods, although not much better than this
with any longer integration periods. Precision at high absorbances
(i.e., A > 1.0-1.5) is limited by emission noises (analyte and background).
For longer wavelengths, the analyte emission noises become
The limiting noises' frequency dependence is obtained with noise
power spectra and analyte absorption, emission, and fluorescence
flicker are shown to all have a similar 1/f (inverse frequency)
character, as well as a value of about 1% for the RSD for Cu. Abbreviated procedures for evaluation of noise in atomic emission (AE) and atomic fluorescence measurements are presented. Other studies into
the source of analyte absorption noise are presented, and the conclusion is drawn that this flicker noise is probably due to nebulization fluctuations, non-fundamental in nature.
The study of these various noise sources' dependence on the
instrumental parameters (e.g. burner position, slit height and width,
flame type and stoichiometry, and lamp current) suggest how to optimize
conditions or concentrations so that an analysis can be carried out
with maximum precision.