7.7.1 Type A evaluation

Type A evaluation refers to the evaluation of a component of measurement uncertainty by a statistical analysis of measured quantity values obtained under defined measurement conditions. In EDXRF, k replicate measurements of CRMs with matrix similar to that of the samples can be used to evaluate the uncertainty of a population of results as its standard deviation σ(y):

As a general rule, the use of CRM having certified values with relatively small uncertainties is recommended. Otherwise, the uncertainty will be enlarged by the uncertainty of the certified value.

Type A evaluation is the simplest way of evaluating the uncertainty. However it is worth to notice that the nature of the replicate measurements must comprise all the sources contributing to the uncertainty of the method (assessment of reproducibility). Performing specific experiments aimed to the assessment of uncertainty due to a single component are of great value to evaluate the effect of each component on the method analytical performance. For example, replicate measurements of a single sample pellet are supposed to reflect the uncertainty due to instability in x-ray tube flux or electronic processing (repeatability in measurement). If replicate pellets of a sample are measured, the uncertainty due to deviations in sample preparation will be also taken into account. If the measurements are carried out in different days and by different operators, a more overall estimation of uncertainty is achieved (method reproducibility).

Any experiment designed to evaluate uncertainty shall be defined as to reflect both matrix type and expected mass fraction range of the elements of interest in the samples. If the mass fraction varies in a wide interval, then several CRMs covering the expected interval must be analysed.