Traceability and measurement uncertainty in sample preparation (W5)

Full text: Very few chemical measurements are being made directly on the object of interest and sample preparation is thus the rule rather than the exception in daily practice. Unfortunately the operations undertaken in the course of sample preparation are prone to rendering a sample useless for the purpose of interpreting a measurement performed on it, as it might not represent the original and relevant status any longer. Sample preparation along with sampling itself constitutes therefore a procedure that leads to a loss of representation of the original specimen or population. On the other hand it is also not sufficient to confine aspects of traceability and measurement uncertainty to the ultimate measurement, as the key purpose of measuring is to supply adequate data for some kind of decision, be it in production, in health, in the environment, or indeed in any other circumstance. These considerations have led to severe confusion in the community as to what traceability really means in chemistry. CITAC and EURACHEM have only recently issued a preliminary document that clarifies these issues and gives a firm handle on the future development of quality assurance in analytical chemistry. In this talk it will be attempted to outline the general ideas and procedures that lead to traceability of analytical chemical results accompanied by valid statements of their uncertainty. It will be argued that the central element in achieving these goals is a well-designed validation study that frequently goes beyond those requirements currently laid out in official documents. (author)