Experimental Determination, Correlation with Microanalyses, and Development of Simplified Prediction Models for Drying Shrinkage of Alkali-Activated Concrete

Abstract

The drying shrinkage of alkali-activated concrete (AAC) is a very complex process, and it warrants long-term testing. The present study was focused on the development of drying shrinkage prediction models for ambient-cured AAC. AAC with four distinct precursor combinations for experimental validation and investigation of the influence of mix proportions were used. The applied fly ash (FA) to ground granulated blast-furnace slag (GGBFS) ratio was varied as 100:0, 70:30, 60:40, and 50:50. Drying shrinkage strains were determined for hardened paste, mortar, and concrete specimens. The observed drying shrinkage behavior was correlated with the mineralogical, chemical, and morphological characteristics of corresponding paste specimens of alkali-activated binder (AAB). These characteristics were evaluated using X-ray diffraction (XRD), Fourier transform infrared spectroscopy (FTIR), and scanning electron microscopy coupled with energy-dispersive X-ray spectroscopy (SEM-EDS), respectively. The results showed that the drying shrinkage of hardened AAB paste and AAB mortar increases by 26%–32% and 13%–28% with GGBFS content, respectively. The qualitative and quantitative microstructural analysis suggested that this behavior could be ascribed to the formation of a calcium aluminosilicate hydrate matrix. Moreover, FA-based AAC exhibited maximum shrinkage, probably due to its comparatively high overall porosity with the addition of both coarse and fine aggregates. Using the experimental data, stepwise linear and multivariate nonlinear regression models were developed by modifying existing portland cement concrete (PCC) models to predict the drying shrinkage strains of AAC. Recommended linear and nonlinear regression models were selected based on the least deviation from the experimental value. The modified GL2000 model was found to be the best fit for shrinkage predictions in AAC because of its simplicity and comparable precision to experimental findings.