Quantification of uncertainty in mercury wastewater analysis at different concentration levels and using information from proficiency test with a limited number of participants.

Measurement uncertainty is one of the most required parameters of analytical quality in environmental decision-making. Several approaches have been reported for estimating uncertainty in analytical measurements. The use of proficiency test (PT) is an alternative, not only for assessing the performance of individual analytical laboratories, but also for estimating the concordance of an analytical method and their measurement uncertainties. Here we develop a PT scheme for a limited number of analytical laboratories that took part in the analysis of total mercury in samples with complex matrix (chlor-alkali process wastewater effluent) in absence of references of the highest metrological hierarchy (e.g. primary method, primary CRM). Two in-house reference materials (IHRMs) were prepared at different levels of analyte concentration and the homogeneity required was verified for the intended use. A set of parametric and robust statistical tests were applied to evaluate the assigned values of each IHRM. Metrological compatibility assessment of PT results, evaluation of individual and global performance of the laboratories, and determination of uncertainty of the analytical measurements were evaluated in this restricted study scenario. Between-laboratory differences were found not significant (α=0.05). For both concentration levels we computed a relative standard uncertainty of 30% for the total Hg content.