Atomic emission detection limits; more than incidental analytical figures of merit! - A tutorial discussion of the differences and links between two complementary approaches—

Abstract

This article discusses the essentials of two apparently differing approaches for measuring, reporting and assessing detection limits in atomic emission, fluorescence, X-ray fluorescence and mass spectrometry: (i) the approach using the signal-to-noise ratio (“SNR approach”) and (ii) the approach using the signal-to-background ratio and the relative standard deviation of the background signal (“SBR- RSDB approach”). The paper contrasts the two approaches, reveals their differences, and emphasizes their connections and complementary character. Since both approaches yield exactly the same numerical values of detection limits, the question is raised whether the one approach has advantages over the other. It is noted that such advantages are not to be sought in the values of the detection limits themselves but in the value of the reportable and transferable information behind the bare values of the detection limits. It is argued that the value of this information may be, in principle, the same but actually is not so, the reason being the difference in tile “standard source of reference”. In the SNR environment, the transfer of “data behind detection limits” requires an actual, universal standard source for calibrations in terms of absolute units and the execution of such calibrations, which is hardly ever done in the spectrochemical field. This contrasts with the SBR-RSDB environment, where the radiant background of the source for which detection limits are reported serves as the “natural standard source of reference”. This makes the transfer of data between systems convenient and practicable, eases comparisons, assessments, and breakdown of detection limits, and, on the whole, greatly enhances the value of reported detection limits and upgrades them to more than incidental figures, with or without merit!