Method validation and measurement uncertainty estimation in trace element determinations in water samples using total reflection X-rays fluorescence spectrometry

Abstract

This study validated a spectrometric method to determine the concentration of 8 elements such as chromium, manganese, iron, nickel, copper, zinc, arsenic and lead in water samples using total reflection X-ray fluorescence (TXRF). Method validation was conducted to determine its linearity, precision, limit of detection (LOD), limit of quantitation (LOQ), reproducibility and robustness using a certified multielement standard. A broad linear range of concentration was achieved for the elements with high coefficient of determination, R² > 0.99, particularly the linearity of Fe and Zn extended up to 10,000 μg/L. Five replicate analyses reported high repeatability or precision with <5 % relative standard deviation. The LOD (0.88–2.37 μg/L) and LOQ (2.92–7.91 μg/L) inversely increased with the decreasing atomic number of elements. Highly reproducible results were achieved with the relative percent difference (RPD), <9 % between 2 different operators over a few weeks of testing. The spectrometer still maintained its robustness with <4 % RPD, even operating at elevated temperature at 32 °C. Method performance was verified through spiking tests in drinking water with 82.6 %–103.9 % recovery at 100–300 μg/L. Its high performance was also verified with element recoveries of 81.4 %–112.8 % at 300 μg/L in the complex matrix interference background of groundwater. The participation of proficiency testing for elements in waste water achieved satisfactory results with the z-score < 0.7. Instrument sensitivity was the main contributor of measurement uncertainty, ranging from 11.8 % to 12.8 % of the measured values. Therefore, TXRF is a fast, reliable and feasible technique for trace elemental analysis as demonstrated from its validation and verification data.