Methodology development and optimization for direct quantification of total selenium concentration in mine water by total-reflection X-ray fluorescence spectrometry

Methodology for direct quantification of total selenium (Se) concentration in mine water by total-reflection X-ray fluorescence (TXRF) spectrometry was developed and optimized. Inductively coupled plasma - mass spectrometry (ICP-MS), as a recognized standard test method, was used as a benchmark for results verification. Preliminary measurements with aqueous multi-element standard solutions showed that Se concentration can be accurately determined down to 10 ng/g by using a single-drop of solution (10 mu L) and a measurement time of 600 s. Quantification of lower Se concentration below 10 ng/g was subsequently optimized, also with aqueous multi-element standard solutions. Improvement in detection limit was accomplished by increasing amount of sample deposition as well as increasing measurement time. TXRF analysis of SRM1640a, Trace Elements in Natural Water, was also performed for accuracy check. The developed methodology was applied to analyze five raw mine water samples (MW-1, MW-2, MW-3, MW-4 and MW-5) collected from anonymous mine site locations. TXRF analysis was performed on the as-received samples directly without filtration, dilution and digestion. Samples were prepared as single-drop specimens and analyzed using a measurement time of 600 s. Se and fifteen other elements were detected in varying degree in the as-received samples. Se concentrations quantified by TXRF were 256.5 +/- 22.5, 88.4 +/- 1.4 and 61.3 +/- 2.8 ng/g for mine water samples MW-1, MW-2 and MW-3, respectively, compared to 255.3 +/- 5.6, 87.3 +/- 2.6 and 59.7 +/- 3.5 ng/g by ICP-MS. No clear Se-K alpha peak above background was observed for samples MW-4 and MW-5. Increasing amount of sample deposition by preparing triple-drop specimens and using a longer measurement time of 1800 s, allowed a Se-K alpha peak to emerge above the background for MW-4, but not MW-5. Se concentration in MW-4 quantified by TXRF was 4.9 +/- 0.5 ng/g, compared to 6.7 +/- 0.7 ng/g by ICP-MS, which represents an agreement within three standard deviation values between the two techniques. For sample MW-5, even after using triple-drop specimens and increasing the measurement time to 1800 s, Se concentration could not be determined because a Se-K alpha peak was not detected by TXRF, compared to the Se concentration of 0.83 +/- 0.1 ng/g obtained by ICP-MS. Overall, our analysis results showed that the TXRF technique can be used for direct analysis of total Se concentration in raw mine water samples down to ng/g range, providing high-quality, accurate and repeatable analysis comparable to ICP-MS while requiring much simpler sample preparation.