A survey of the inorganic chemistry of bottled mineral waters from the British Isles

The inorganic chemistry of 85 samples of bottled natural mineral waters and spring waters has been investigated from 67 sources across the British Isles (England, Wales, Scotland, Northern Ireland, Republic of Ireland). Sources include boreholes, springs and wells. Waters are from a diverse range of aquifer lithologies and are disproportionately derived from comparatively minor aquifers, the most represented being Lower Palaeozoic (10 sources), Devonian Sandstone (10 sources) and Carboniferous Limestone (9 sources). The waters show correspondingly variable major-ion compositions, ranging from Ca-HCO3, through mixed-cation-mixed-anion to Na-HCO3 types. Concentrations of total dissolved solids are mostly low to very low (range 58-800 mg/L). All samples analysed in the study had concentrations of inorganic constituents well within the limits for compliance with European and national standards for bottled waters. Concentrations of NO3-N reached up to half the limit of 11.3 mg/L, although 62% of samples had concentrations <1 mg/L. Concentrations of Ba were high (up to 1010 mu g/L) in two spring water samples. Such concentrations would have been non-compliant had they been classed as natural mineral waters, although no limit exists for Ba in European bottled spring water. In addition, though no European limit exists for U in bottled water, should a limit commensurate with the current WHO provisional guideline value for U in drinking water (15 mu g/L) be introduced in the future, a small number of groundwater sources would have concentrations close to this value. Two sources had groundwater U concentrations > 10 mu g/L, both being from the Welsh Devonian Sandstone. The highest observed U concentration was 13.6 mu g/L. Solute concentrations in waters contained in glass bottles compared with waters in PET showed slightly though significantly higher concentrations of Al, Ce, Cu, La, Nd, Mn, Sn, W, Zn and Zr (rank-sum testing, p < 0.05). By contrast, Sb concentrations were significantly higher (p < 0.001) in samples contained in PET bottles. This accords with other studies that have recognised Sb contamination in water from PET bottles. However, in no cases did the concentration of Sb exceed or approach the national and European limit for Sb in natural mineral water/spring water (5 mu g/L), the highest observation being 1.35 mu g/L. Bottled water compositions were mostly similar in their major-ion characteristics to raw groundwaters from the equivalent aquifers in Britain, although concentrations of several trace elements (Al, Cd, Cu, Fe, Mn, Pb and Zn) were appreciably lower, in some cases by one or two orders of magnitude. The most likely mechanism for the reduction is use of aeration, settling and filtration to remove unstable constituents before bottling. The comparatively low concentrations of Cd, Cu, Pb and Zn are likely to be due to co-precipitation with/adsorption to precipitated metal oxides, although choice of resilient pipework (e.g. stainless steel) in bottling plants may also be a factor. Although for the most part the major ions in the bottled waters appear representative of the groundwater in their host aquifers, the results suggest that many of the trace elements have been modified significantly from natural compositions in situ. (C) 2010 Natural Environment Research Council. Published by Elsevier Ltd. All rights reserved.