A logical way through the limits of quantitation in inductively coupled plasma spectrochemistry

The limit of quantitation (LOQ) is a crucial parameter in quantitative analysis. Besides the classical “10 sB approach”, other concepts have been previously described based on the limitation resulting from the calibration procedure, namely from the prediction bands and the uncertainty calibration. Because of the difficulty of obtaining a reliable value of the blank standard deviation, this approach was replaced by the study of the %RSD of the net signal as a function of the concentration. Recently described, the so-called accuracy profile method, accuracy being the contribution of both the trueness and the precision, allows the analyst to define a validity domain, the lowest point acting as a limit of quantitation. The aim of this work was to determine these various limits of quantitation, including that from the accuracy profile, to compare them, and to study the parameters that can influence their values. It was concluded that the LOQ deduced from the accuracy profile is a realistic one. Its value is mainly influenced by the calibration procedure, particularly when a weighting procedure is used. A 1/y weighting factor, y being the analyte line intensity, seems to be an efficient compromise for calibration.

► An analytical method is defined by its accuracy, i.e. both trueness and precision.
► A limit of quantitation should be based on an accuracy criterion and not on precision.
► The accuracy as a function of an analyte concentration is an accuracy profile.
► Profile-based tolerance intervals define a validity domain.
► The lowest concentration of this domain corresponds to a limit of quantitation.