Does understanding metrological traceability in analytical measurement require going “back to basics”?

Traceability of a measurement result to a common reference is termed ‘metrological traceability’ in the International vocabulary of metrology, VIM [1], entry 2.41. It is a concept of basic importance because metrological traceability is a precondition to ‘metrological comparability of measurement results’ (entry 2.46 in [1]), and comparing measurement results is a major goal of measurement. Comparing measurement results, which are not traceable to a common reference, is a meaningless operation. Sadly enough, the term traceability was not really well understood in the analytical measurement community until about 20 years ago (the first Workshop on ‘‘comparability’’ and ‘‘traceability’’ was held at IRMM in Geel, Belgium in November 1992). Pictures of ‘metrological traceability’ of analytical measurement results have been shown in the form of a ‘metrological traceability chain’ (entry 2.42 in [1]), for instance in an IUPAC study released in 2011 [2]. Yet, the concept and associated term metrological traceability are still giving rise to complicated reasonings, confusing descriptions, and erroneous conclusions. Is it indeed so difficult? Let us try to improve understanding of ‘‘metrological traceability.’’ Usually, a measurement involves the measurement of a ‘certified reference material, CRM’ (entry 5.14 in [1]), i.e., a ‘reference material, RM’ (entry 5.13 in [1]) in which a known quantity or property value with associated uncertainty is embodied. To calibrate the measurement, we like to choose a CRM whose certified value is as close as possible to the expected measured value (entry 2.10 in [1]) of the test sample. We even like the CRM to have a chemical structure (‘‘matrix’’) as similar as possible to the test material structure in which the analyte is embodied. Note: an RM, hence also a CRM (a special case of an RM) can only serve one purpose in one and the same measurement: either being used as a calibrator, or as a trueness control material (entry 5.13 Note 2 in [1]). In our discussion here, we only consider the first of these two functions: using it as a calibrator. What then is a ‘‘reference’’ needed for the definition of metrological traceability? The answer is given in Note 1 of entry 2.41 in VIM [1]: