In Phase I, particles from 17 different aggregates, 1.25-5 mm in size, were immersed in continuously agitated solutions at 38 degreesC: distilled water, Ca(OH)(2)-saturated solution, 0.7 M NaOH (measurement of K supply), and 0.7 M KOH (measurement of Na supply). These solutions were periodically analysed for K and/or Na up to 578 days. More alkalies were released in alkaline solutions than in
... [Show full abstract] lime-saturated solution, with lower values in water. After 578 days, the aggregates released between <0.01% and 0.19% Na(2)O(e), excluding the nepheline-rich aggregate tested (0.68%). This would correspond to a contribution to concrete from <0.1 to 3.4 kg/m(3) Na(2)O(e) (12.7 for the phonolite), based on an aggregate content of 1850 kg/m(3). In general, the feldspar-rich aggregates released significantly more alkalies. In Phase II, the water-soluble alkali content of mass concrete elements from many dams was measured using a hot water extraction method. The values obtained often largely exceed the soluble alkali content expected to be released by the cement used. These results thus also suggest that large amounts of alkalies were supplied with time by the aggregates, particularly by feldspar-rich ones.