Selective extraction of water-soluble thallium fraction from contaminated drinking-water distribution networks: optimization of the procedure and extracts speciation

Metallic pipes in drinking water (DW) distribution networks are able to retain some elements over time that may be subsequently released into the water. In September 2014, researchers from the Department of Earth Sciences (University of Pisa, Tuscany, Italy) reported the presence of thallium in water samples collected from Pietrasanta DW wells (Lucca, Tuscany, Italy) at concentrations (up to 10 μg/L) much higher than US-EPA max contaminant level goal (2.0 μg/L). So far, there is no analytical sequential extraction procedure to selectively distinguish between the fraction released into DW, characterized by health and hygiene relevance, and the remaining insoluble one, strongly retained on the pipe inner surface. In 1992, Community Bureau of Reference (BCR) published an extraction procedure to be applied to sludge and sediments, in which four extracting solvents with increasing reactivity were sequentially made to come into contact with the sample. In particular, an acetic acid solution was used to dissolve the water soluble fraction at weakly acid pH during the first extraction step. The procedure, optimized for fine powdered solids that are dispersible in the extracting solvent by mechanical stirring, is not suitable for thick alternating layers, which could cover the inner surface of pipe core samples. Starting from the BCR method, we have developed an innovative procedure to determine the thallium water-soluble fraction retained by contaminated pipe core samples. The target fraction is solubilized in acetic acid solution by means of consecutive sonication treatments (at least, five) carried out after changing the aliquot of the extractant. All the working conditions (such as, solvent volume and concentration, sonication time) have been optimized while the number of extraction steps required to ensure the complete dissolution of the water-soluble fraction has been calculated. The obtained extracts were analyzed by ICP-MS and subsequently speciated by polarography with a dropping mercury electrode (DME) and by HPLC - ICP-MS, as well. The procedure applied to a number of pipe core samples collected from a thallium-contaminated DW network showed the presence of the water-soluble fraction in a percentage of 5-9 % of the total quantity detected, mainly coming from the release of Tl(I)-enriched layers.