Combining method validation data with additional information for measurement uncertainty evaluation - determination of melatonin content in tea.

Abstract

The evaluation of measurement uncertainty involves combining available experimental data with additional knowledge on the impact of sources of uncertainty on the measured value to quantify combined uncertainty. This uncertainty should express the impact of relevant random and systematic effects on the measured value to allow an objective interpretation of the performed quantification. This research discusses how to assess and account for the impact on measurement uncertainty evaluation of uncontrolled high temperature of sample preparation and of supporting experimental validation data on a single calibration of the instrumental method of analysis used. After the detailed inspection of the measurand and measurement process, the various uncertainty components were quantified from experimental data and/or additional information. This research illustrates that experimental limitations can be overcome by a detailed and quantitative understanding of the measurement process. Melatonin content in tea was quantified between 3.7 g kg^-1 and 196 g kg^-1 with a relative expanded uncertainty between 28% and 40%, lower than the target or maximum admissible relative expanded uncertainty of 40%. The use of HPLC-FD calibration several days before sample analysis is the major uncertainty component responsible for up to 82.7% of the measurement uncertainty.