Determination of measurement uncertainty for the determination of triazines in groundwater from validation data

Laboratories are increasingly urged to submit full uncertainties of their a nalytical results rather than only standard deviations. The determination o f measurement uncertainties in compliance with the Guide to the Expression of Uncertainty in Measurement (GUM) is demonstrated using the validation ap proach explicitly endorsed by the recent edition of the EURACHEM guide for the determination of measurement uncertainty. Measurement uncertainty was s plit into uncertainty of the sample mass, uncertainty of the concentration of the stock standard solution, uncertainty of the calibration and uncertai nty connected to within- and between-series precision. Uncertainties of sam ple mass and of the concentration of the stock standard solution were 0.26 and 1.14% for all analytes, which is negligible compared with the contribut ions of precision and calibration. Uncertainty of calibration was estimated from the calibration graph. Relative uncertainty of calibration was found to be strongly concentration dependent and to be the main uncertainty contr ibution below 0.2 mug L-1. Precision was split into within-series and betwe en-series standard deviation, which dominate the combined standard uncertai nty at higher concentrations. The results obtained from these calculations are compared with results for a certified reference material and with the p erformance in an interlaboratory comparison. It was found that all results agreed within their uncertainty with the target values, showing that the es timated uncertainties are realistic.