Measurements recovery evaluation from the analysis of independent reference materials: analysis of different samples with native quantity spiked at different levels

摘要

Measurement uncertainty evaluation involves combining uncertainty components reflecting all relevant random and systematic effects: the precision and trueness uncertainty components, respectively. Typically, trueness is assessed through the analysis of various materials with known reference value, such as certified reference materials (CRMs) or spiked samples, from which it should be decided about the relevance and the need to correct measurement results for systematic effects. Algorithms proposed so far to assess systematic effects are only applicable to the analysis of the same reference material type or assume that some uncertainty components affecting evaluations are negligible or constant. This work presents detailed algorithms for the assessment of systematic effects, through the determination of recovery and the respective recovery uncertainty, applicable to the analysis of various independent reference materials, such as CRMs and spiked samples with native analyte. These algorithms are applicable to cases where native analyte and/or spiking values are associated with relevant and significantly different uncertainties allowing for a reliable assessment of systematic effects and measurement uncertainty for these complex cases. This methodology was successfully applied to the quantification of Na, K, Mg, Ca, Cr, Mn, Fe and Cu in water samples from two proficiency testing schemes, by ICP-OES, where recovery was estimated from the analysis of samples with different native concentrations and spiked at different levels. The relative expanded uncertainties of the measurement results ranged from 28.9?% to 3.9?% and are fit for the monitoring of environmental water samples in accordance with criteria set in the European Union legislation.