Study on the effect of selected parameters on the alkali-silica reaction of aggregate in ground glass fiber and fly ash-based geopolymer mortars

In this study, the effect of selected parameters on the alkali-silica reaction (ASR) of aggregate in ground glass fiber (GGF) and fly ash-based geopolymer mortar was investigated. This work focuses on three main objectives: 1) to study the effect of the binder composition, 2) to study the effect of the ASR accelerating environment, and 3) to study the effect of the chemical/mineralogical composition of aggregate, on the ASR behavior of geopolymer mixtures. To study the effect of paste composition on the ASR behavior of GGF-based geopolymer mortars were compared to a fly ash-based geopolymer mixture and an ordinary Portland cement mixture. To investigate the effect of curing condition, mortar bars were cured in two ASR accelerating environments; mortar bars were either immersed in a high alkali solution or exposed to a high relative humidity environment. To evaluate the effect of the chemical composition of aggregate, two types of ASR-reactive aggregate, one mainly composed of glassy silica and one mainly composed of limestone with chert impurities were used for the ASR study, and the results were compared to a mixture with non-reactive limestone aggregate. To evaluate the ASR behavior of geopolymer mixtures, length expansion, and the change in dynamic modulus of elasticity (DME) of mortar bars that were exposed to either of the ASR accelerating conditions were monitored for ten weeks. Based on the results obtained from this study, both geopolymer mixtures showed lower levels of expansion in comparison to the OPC mixture. However, despite the lower expansion level, a significant loss in DME was observed for fly ash-based geopolymer specimens made with siliceous aggregate and cured in a high alkali environment. Finally, curing the specimens in a high relative humidity environment did not seem to be an effective procedure in the evaluation of the ASR behavior of aggregate, as this method failed to discriminate the ASR-reactive aggregate used in this study. (C) 2020 Elsevier Ltd. All rights reserved.