Abstract Alkali-silica reaction is a major durability issue of concrete structures in many countries. The main drivers considered for reaction are the aggregate type, moisture, and alkalis from the binder. However, it has been shown that aggregates themselves can also contribute to the free alkalis in pore solution of concrete. To account for that in standards, the real alkali contribution from aggregates needs to be understood. This investigation covers the alkali release from one Norwegian “low releasing” aggregate and one Canadian “high releasing” aggregate in alkaline solution according to the RILEM AAR-8 method. Besides the entire sample (0–4 mm), different single fractions, grading curves of filler and varying solid/liquid ratio were investigated. Fineness of particles and the specific surface area is shown to be the dominant factor for alkali release. The mineralogical composition played a minor role in the investigated aggregates. When summing up the measured alkali release from the four single fractions (2–4, 0.5–1, 0.125–0.25 and < 0.125 mm), the values were in all cases (both aggregates and cations) considerably higher than the measured alkali release from the created “SUM” sample comprising the same fractions and the entire AAR-8 sample. Increasing the amount of filler fines four-fold in the same amount in solution, led to approx. a doubling of the alkali released. Depletion of the added Ca(OH) 2 in the alkaline solution is suggested as a potential explanation for the limited alkali release.
Danner et al. (Sat,) studied this question.