Modeling the compressive strength of metakaolin-based geopolymers based on the statistical analysis of experimental data

Abstract

The relationship between the compressive strength of metakaolin-based geopolymer samples and different processing conditions has been investigated for both potassium and sodium based geopolymer systems. Cubic geopolymer samples were prepared by mixing the slurry for 1 h in a thermostatic bath at 0 °C. >1200 samples have been tested to gather enough data to carry out a meaningful statistical analysis. All the data evaluation and model development have been carried out extensively using R. The variation of curing and aging time, curing temperature, SiO₂/Al₂O₃ and H₂O/Al₂O₃ molar ratios has been accounted for via the application of statistical models whose reliability has been suitably checked. Curing has been performed in a sealed container at 100% relative humidity. Aging has been conducted in a climate chamber kept at 75% of relative humidity using an oversaturated solution of NaCl. Curing time has proved a positive relationship with compressive strength while aging time does not show evidence of any significant effect. Curing temperature negatively affects compressive strength. Increasing the SiO₂/Al₂O₃ molar ratio results in an increase of the compressive strength within a certain range of values for the ratio; however, above a threshold (3.8 for the potassium-based and 3.4 for the sodium-based geopolymer system) the mechanical properties decrease. The H₂O/Al₂O₃ molar ratio displayed an inverse proportionality with the compressive strength except for the sodium-based geopolymer, where the mechanical properties initially increased. A further comprehensive and statistically sound model has been proposed that allows us to predict the strength of geopolymer samples as a function of process variables and their composition, ranging in a rather wide set of values.