Effect of slag on expansion due to alkali-aggregate reaction in concrete

The efficacy of ground granulated blast furnace slag in controlling expansion dire to alkali-aggregate reaction (AAR) was evaluated using both concrete prism and accelerated mortar bar testing. Six aggregates were included in the program; four of these ic ere classified as alkali-silica reactive (siliceous limestone, sandstone, greywacke, and granite), one as alkali-carbonate reactive (dolomitic limestone), and one as nonreactive (dolostone). The partial replacement portland cement with slag in the range of 25 to 65 percent was effective in retarding the rare of expansion and limiting the ultimate expansion at 2 years in concrete prisms cast with all four alkali-silica reactive (ASR) aggregates. However, the minimum level of slag required to control expansion to an acceptable level (e.g., 0.04 percent at 2 years) was found to vary depending on the nature of the aggregate and the amount of alkali present in the concrete. There appears to be a reliable correlation between the expansion of mortar bars after 14 days storage in 1 M NaOH at 80 C and the expansion of concrete prisms after 2 years storage over water at 38 C. The accelerated mortar bar test appears to be an appropriate tool for determining the minimum "safe" level of slag required Sor a particular reactive aggregate source. The use of slag also reduced the initial rate of expansion of concrete prisms containing the alkali-carbonate reactive (ACR) aggregate. However, the expansion at 2 years exceeded 0.20 percent for all slag levels rested (25 to 65 percent) and this final expansion was actually greater for concretes with 25 to 50 percent slag than for the control concrete. The expansion of mortar bars was less than 0.10 percent at 14 days for all the mixes tested which confirms the unsuitability of this test for detecting ACR aggregates.