Evaluation of the Effect of Revising the Estimation Procedure of Uncertainty in Dioxin Analysis

Abstract

: Procedure for estimating uncertainty in the analysis of dioxins was revised, and differences in the results obtained between the former and revised procedures were investigated. In the former procedure, uncertainty was estimated from fluctuations in the actual sample data such as the recovery rate of the sampling spike (for a gas sample), duplicate sampling (for a water sample), and replicate tests (for a soil/ash sample). As such, the former procedure differed depending on the sample medium. In the revised procedure, the author estimated uncertainty according to a bottom-up approach. The uncertainty determined in a gas sample and an ash sample using the revised procedure tended to be larger than that obtained by the former procedure. In a water sample, the uncertainty determined using the revised procedure was smaller than that obtained using the former procedure in low concentration range. These results suggest that (1) the uncertainty using the bottom-up approach may become larger than the actual variation when the maximum estimate of an uncertainty factor is assigned (in this study, the variation in the peak area ratio from GC/MS measurements), and that (2) differences in uncertainty determined by different estimation methods readily occur, particularly in the low concentration range.