Guiding concepts relating to trace analysis

Abstract

Trace' is a term much used in colloquial language; it has many connotations and should not be ‘appropriated’ to denote ‘scales of working’ in chemistry. The concept “scope of analysis” is proposed instead. The quantities “amount of substance”, “mass” and “concentration” are critically considered and it is proposed that more use of a logarithmic presentation of concentration values should be made. The role of the “blank portion” of measured quantities with respect to trace determinations is pointed out; “analytical signal” and “analytical noise” are discussed. Linearization of the “calibration function” and elimination of the blank measures are required for extrapolation of calibration functions to low concentrations. The different calibration methods (σ,α,δ,ω) are presented; the importance of “rating” analysis and “gradual” analyses with coarsened scales is pointed out. The ‘type’ of information, the “informing power” provided by a procedure and the “information required” by the analytical problem are discussed. The “topological structure” of a procedure is shown to be decisive for its applicability to trace determinations. “Selectivity” and “sensitivity” are defined as metric quantities to be calculated from the “calibration matrix” of the procedure. High values of these two quantities are required for ‘trace work’. “Specificity” is mentioned only. Finally some remarks about the “limit of detection” and the “limit of guarantee of purity” are made to clear up common mistakes, made when applying these useful concepts. © 1973, Walter de Gruyter. All rights reserved.