The results of phase 1 of an interlaboratory test, coordinated by the RILEM TC 267-TRM “Tests for Reactivity of Supplementary Cementitious Materials” showed that the R³ (rapid, relevant, reliable) test method, by measurement of heat release or bound water, provided the most reliable and relevant determination of the chemical reactivity of supplementary cementitious materials (SCMs), compared to other test methods. The phase 2 work, described in this paper aimed to improve the robustness of the test procedure and to develop precision statements for the consolidated test procedure. The effect of the pre-mixing and mixing conditions, and the impact of the mix design on the test method robustness were assessed and fixed for optimal conditions to carry out the R³ heat release test. The effect of the drying step was evaluated to define the R³ bound water test procedure in more detail. Finally, the robustness of the consolidated final test methods was determined by an interlaboratory study to define the precision statements.

A primary aim of RILEM TC 267-TRM: “Tests for Reactivity of Supplementary Cementitious Materials (SCMs)” is to compare and evaluate the performance of conventional and novel SCM reactivity test methods across a wide range of SCMs. To this purpose, a round robin campaign was organized to investigate 10 different tests for reactivity and 11 SCMs covering the main classes of materials in use, such as granulated blast furnace slag, fly ash, natural pozzolan and calcined clays. The methods were evaluated based on the correlation to the 28 days relative compressive strength of standard mortar bars containing 30% of SCM as cement replacement and the interlaboratory reproducibility of the test results. It was found that only a few test methods showed acceptable correlation to the 28 days relative strength over the whole range of SCMs. The methods that showed the best reproducibility and gave good correlations used the R³ model system of the SCM and Ca(OH)₂, supplemented with alkali sulfate/carbonate. The use of this simplified model system isolates the reaction of the SCM and the reactivity can be easily quantified from the heat release or bound water content. Later age (90 days) strength results also correlated well with the results of the IS 1727 (Indian standard) reactivity test, an accelerated strength test using an SCM/Ca(OH)₂-based model system. The current standardized tests did not show acceptable correlations across all SCMs, although they performed better when latently hydraulic materials (blast furnace slag) were excluded. However, the Frattini test, Chapelle and modified Chapelle test showed poor interlaboratory reproducibility, demonstrating experimental difficulties. The TC 267-TRM will pursue the development of test protocols based on the R³ model systems. Acceleration and improvement of the reproducibility of the IS 1727 test will be attempted as well.