Measurement uncertainty in Total Reflection X-ray Fluorescence

Total Reflection X-ray Fluorescence (TXRF) spectrometry is a multi-elemental technique using micro-volumes of sample. This work assessed the components contributing to the combined uncertainty budget associated with TXRF measurements using Cu and Fe concentrations in different spiked and natural water samples as an example. The results showed that an uncertainty estimation based solely on the count statistics of the analyte is not a realistic estimation of the overall uncertainty, since the depositional repeatability and the relative sensitivity between the analyte and the internal, standard are important contributions to the uncertainty budget. The uncertainty on the instrumental repeatability and sensitivity factor could be estimated and as such, potentially relatively straightforward implemented in the TXRF instrument software. However, the depositional repeatability varied significantly from sample to sample and between elemental ratios and the controlling factors are not well understood. By a lack of theoretical prediction of the depositional repeatability, the uncertainty budget can be based on repeat measurements using different reflectors. A simple approach to estimate the uncertainty was presented. The measurement procedure implemented and the uncertainty estimation processes developed were validated from the agreement with results obtained by inductively coupled plasma - optical emission spectrometry (ICP-OES) and/or reference/calculated values. (C) 2015 Elsevier B.V. All rights reserved.