Title: Evaluation of Testing Methods Used for Assessing the Effectiveness of Mineral Admixtures in Suppressing Expansion due to Alkali-Aggregate Reaction

Author(s): M. A. Berube and J. Duchesne

Publication: Symposium Paper

Volume: 132

Issue:

Appears on pages(s): 549-576

Keywords: accelerated tests; alkali-aggregate reactions; alkalies; blast furnace slag; cements; concretes; fly ash; mortars (materials); silica fume; siliceous aggregates; tests; Materials Research

DOI: 10.14359/2047

Date: 5/1/1992

Abstract:
The effectiveness of one ground granulated blast furnace slag, two condensed silica fumes (high-silica/low-alkali, low-silica/high alkali), and three pulverized fly ashes (low-alkali/low-calcium, low-alkali/moderate calcium, high-alkali/high-calcium) have been evaluated in the presence of two very alkali-silica reactive aggregates from Canada, a siliceous limestone and a rhyolitic tuff. Mortar bars and concrete specimens were made with various admixture contents and different cements (high- and low-alkali), and tested for expansion with the accelerated mortar bar method (ASTM C 9-P214) and the concrete prism method (CAN/CSA-A23.2-14A). The mineral admixtures were also submitted to the pyrex mortar bar method (ASTM C 441). Based on the results, the ASTM C 441 test is not recommended for assessing the effectiveness of mineral admixtures in suppressing expansion due to alkali-aggregate reaction, unless account is taken of a number of modifications concerning mix design (admixture content, water/cement ratio, alkali content, etc.) and performance criteria. Pyrex does not behave like a natural aggregate. The results from ASTM C 9-P214, using a limit of 0.1 percent expansion at 14 days, are in agreement with those from the concrete prism method, which is the most recommended test procedure. However, when testing concrete, the alkali content of the mix must always be increased to 1.25 percent of the mass of cement (Na?2O equivalent), otherwise the test is not accelerated sufficiently and low expansion will be observed in the presence of reactive aggregates, even with no mineral admixtures. The long-term effectiveness of mineral admixtures against alkali-aggregate reactions (AAR), in particular silica fume, is presently questioned by a number of workers. Therefore, it is firmly recommended that conservative limits be used when conducting laboratory tests on concrete specimens, and that the tests be extended to at least two years. The mineral admixture under study should be used in amounts such that expansion never exceeds 0.04 percent in the long term (two years or more). A more conservative, and more recommended, performance criterion is to obtain expansion in the long term that is similar to that of a control made with a low-alkali cement and containing no mineral admixture