For Category I assays or Category II tests, accuracy can be determined by conducting recovery studies with the appropriate matrix spiked with known concentrations of elements. It is also an acceptable practice to compare assay results obtained using the AA procedure under validation to those of an established analytical procedure. In standard addition methods, accuracy assessments are based on the final intercept concentration, not the recovery calculated from the individual standard additions.

The analytical procedure should be assessed by measuring the concentrations of six independently prepared sample solutions at 100% of the assay test concentration. Alternatively, three replicates of three separate sample solutions at different concentrations can be used. The three concentrations should be close enough that the repeatability is constant across the concentration range. If this is done, the repeatability at the three concentrations is pooled for comparison to the acceptance criteria. If validating a procedure by the method of standard additions, the precision criterion applies to the final experimental result, not the accuracy of the individual standard addition levels.

The effect of random events on the analytical precision of the procedure should be established. Typical variables include performing the analysis on different days, using different instrumentation, or having the method performed by two or more analysts. As a minimum, the analytical procedure should be assessed by performing the repeatability test in any of the conditions previously mentioned (totaling 12 measurements).

The procedure must be able to unequivocally assess each analyte element in the presence of components that may be expected to be present, including any matrix components.

The limit of quantitation (QL) can be estimated by calculating the standard deviation of NLT six replicate measurements of a blank solution, divided by the slope of a standard curve, and multiplying by 10. If validating a procedure using the method of standard additions, the slope of standards applied to a solution of the test material is used. Other suitable approaches can be used (see 1225).

A measurement of a test solution prepared from a representative sample matrix spiked at the estimated QL concentration must be performed to confirm accuracy. If validating a procedure using the method of standard additions, the validation criterion applies to the final experimental result, not the spike recovery of the individual standard addition levels.

A response curve between the analyte concentration and absorbance is prepared from NLT five standard solutions at concentrations encompassing the anticipated concentration of the test solution. The standard curve is then evaluated using appropriate statistical methods, such as a least-squares regression.

For experiments that do not yield a linear relationship between analyte concentration and AA response, appropriate statistical methods must be applied to describe the analytical response.

Range is the interval between the upper and lower concentrations (amounts) of analyte in the sample (including these concentrations) for which it has been demonstrated that the analytical procedure has a suitable level of precision, accuracy, and linearity. Range is demonstrated by meeting the linearity, precision, and accuracy requirements.

The reliability of an analytical measurement is demonstrated by deliberate changes to experimental parameters. For AA this can include but is not limited to sample preparation steps and heating programs, including atomization hold time or atomization temperature. Exercise caution when changing fuel and oxidant gas flows and burner hardware, because this could potentially create a flash-back condition.