Shrinkage and Related Properties of Alkali-Activated Binders Based on High MgO Blast Furnace Slag

Abstract: Concrete is the second most used material in the world just after water. A drawback is that it is mostly based on Portland cement, which has an extremely high carbon footprint reaching a staggering 900 kg/tonne. The carbon dioxide emissions related to the production of the Portland cement accounts for nearly 8 % of the global total. Consequently, the construction sector is engaged in an active search for sustainable alternatives. Over the past few decades, alkali-activated materials (AAMs) emerged as one alternative and attracted strong scientific and commercial interests. Many industrial by-products produced in large volumes can be used as precursors for the AAMs system. The most common include blast furnace slag, fly ash, mine tailings, metallurgical slags, and bauxite residues. So far, products based on ground granulated blast furnace slag (GGBFS) showed the best price/performance ratio. Still, there are a number of unresolved issues, which must be addressed to ensure the economical and safe full-scale utilisation of that material. The research work presented in this thesis focuses on alkali-activated concretes based on Swedish water-cooled high-MgO ground granulated blast furnace slag. The objective of this work was to identify experimentally factors that are controlling the shrinkage and the creep of concretes made with this type of GGBFS and to understand their influence on various physical and chemical properties of fresh and solidified systems. Liquid sodium silicate, powder sodium carbonate and a combination of both were used to activate the binder chemically. Two curing procedures were followed; laboratory curing and heat curing at 65°C applied for 24 hours. Various properties were determined including workability, setting time, hydration heat development, shrinkage, creep, efflorescence, carbonation, compressive strength, microstructure and phase composition. Additionally, the effects of the activator type, dose, binder fines, binder composition and curing regime were investigated. The results revealed that the particle size distribution of the binder as well as the activator type and its dosage have strong effects on the produced materials. Increasing the activator amount or decreasing the alkali modulus of the used sodium silicate activator improved the early-age compressive strength and accelerated the hydration reaction. Alkali-activated high-MgO slag concrete showed higher autogenous and drying shrinkage, as well as higher creep in comparison to the Portland cement-based reference concrete. The sodium silicate increased the slump, shortened the setting time, increased the compressive strength and shrinkage but lowered the creep in comparison with the sodium carbonate-activated mixes. Replacing 20% of the slag with fly ash and decreasing the alkali modulus of the sodium silicate activator increased the autogenous shrinkage but decreased the ultimate drying shrinkage. Application of a heat treatment produced in general a higher early age compressive strength, a lower VI later strength development, a more porous microstructure and a decreased ultimate measured shrinkage. Sealed curing decreased the ultimate shrinkage by up to 50%. Some of the produced mixes showed strong efflorescence. Two years of curing in laboratory conditions resulted in an extensive carbonation of some of the mixes. This weakened the silicate binding of the gel and produced a coarser porosity due to the decalcification of C-(A)-S-H. The heat-cured samples activated with sodium silicate were the most affected. Many mixes showed an extensive microcracking of the binder matrix. However, the within this study newly developed mixes were substantially less effected. These optimised mixes were based on a combination of sodium silicate and sodium carbonate activators, combined with a heat treatment and partial replacement of the slag with fly ash. The main hydration phase that formed was C-(A)-S-H, with gaylussite, calcite, nahcolite and hydrotalcite as secondary phases. The partial replacement of slag with fly ash resulted in a dominant formation of N-(A)-S-H and C-(A)-S-H.