Alkali activated metakaolin with high limestone contents – Statistical modeling of strength and environmental and cost analyses

Highlights

• Alkaline binders of metakaolin with high limestone contents were statistically studied.

• The use of limestone reduced the water demand and the amount of alkalis.

• These are advantageous in strength, environmental impact and cost relative to Portland cement.

• The limestone load and Na/Al and Si/Al ratios, and their interactions, affected the strength.

• Optimized predicted and corroborated binders of 60%limestone reached ∼63  MPa at 28-day.

Abstract

Alkaline binders of metakaolin (MK) and limestone (LS) are promising sustainable cementitious new alternatives; however, systematic studies are required to understand them to take them to industrial level. This study presents the modeling and optimization of strength using the surface response method of such binders with up to 80%LS and molar ratios of Na/Al = 0.5–1.86 and Si/Al = 1.81–3.13. A quadratic model (R-squared = 94.31%) indicated an optimal formulation of 32.32%LS, Na/Al = 1.34 and Si/Al = 2.96 with a predicted, and experimentally corroborated, 28-day strength of 70 ± 3.9 MPa, which was higher than references of 100%MK and blended Portland cement (BPC). The microstructural features were in agreement with the high strength. The analyses of cost, CO₂-emissions and energy demand indicated pastes of 100%MK are better than BPC only regarding CO₂-emissions, while blends with %LS>32.12% were better in all indicators relative to both of the former. As limestone is abundant, cheap, reduces alkali demand and do not require calcination, these binders are actually sustainable.

Introduction

Geopolymers are alkali activated binders with a three-dimensional inorganic polymeric structure [[1], [2], [3]] formed by alkaline activation of an aluminosilicate precursor such as metakaolin (MK) [[4], [5], [6], [7]]; these can have similar or superior mechanical properties compared to Portland cement [8,9]. The purity of the precursor MK [7,[10], [11], [12], [13], [14], [15]] and the overall chemical composition of the alkaline system have strong influences on the properties [6,16,17]. Compressive strengths of 23–85 MPa have been reported for MK-based geopolymer [6,7,11,14,[17], [18], [19], [20]] for molar ratios Na/Al of 0.9–1.4 and Si/Al of 1.8–3.1; such molar ratios are commonly kept fixed by adjusting the concentration of sodium silicate and sodium hydroxide relative to the chemical composition of the MK, which resulted in the use of high amounts of alkaline activator. For example a blend of Na/Al = 1.4 and Si/Al = 3.1 can reach 72 MPa after curing at 50 °C for 24 h and then at 20 °C for 27 days [7]; the activator requirements were high, of ∼25%Na₂O relative to the mass of precursor, which required about 48% anhydrous sodium silicate of module Ms = 3.27 plus 18%NaOH relative to the mass of MK.

The literature shows debate about the environmental footprint of the geopolymers [8,[21], [22], [23], [24], [25], [26]] due to the strong environmental impacts of the of sodium silicate [[27], [28], [29]] and sodium hydroxide [22,27,30,31]. Even though aluminosilicate minerals are among the most abundant materials worldwide [[32], [33], [34], [35]], kaolinitic clays need calcination at 700–850 °C to make them reactive (MK) [[36], [37], [38]]; this increases the environmental impact and cost on the final product, so the manufacture of MK-based geopolymers is sometimes questioned [26,39].

On the other hand, limestone (LS) represents an attractive raw material for the manufacture of low environmental footprint binders as it is widely and readily available and as its handling can be readily scaled using the infrastructure of the Portland cement industry [35]. LS is used as a partial substitution of Portland cement of up to 35% according to international standards [40,41], although current studies [35,[42], [43], [44]] have suggested that cements with up 50% mass LS are suitable in terms of strength, durability, practicability and environmental impact. In those studies, LS seems not to be a totally inert component since it reacted with aluminum phases, from of Portland cement or from supplementary materials, improving the microstructure, early strength and durability of the resulting product [42,43].

Most recently, LS has been reported as a supplementary precursor on MK-based geopolymer cements [[45], [46], [47], [48], [49], [50]]. The addition of 10%LS in pastes [46,47] improved strengths reaching up to 20 MPa, while higher %LS was detrimental; the authors considered the LS as an inert filler that enhanced the microstructure of the hardened binders. Another study [50] on pastes found also a favorable on the strengths by adding 20%LS reaching up to 45 MPa but higher %LS was also deleterious; the authors discarded the possibility of a simple micro-aggregate effect of the LS, as its dissolution was observed participating in the geopolymeric gel structure improving the strength. A study in pastes of MK-LS activated with NaOH [48] reported 7 MPa as the best strength with 50%LS; the authors indicated that the LS enhanced the release of Al and Si ions from MK and the leaching of Ca²⁺ from the LS was observed.

Furthermore, there are also reports of LS as a supplementary precursor in blends with other precursors of a chemical composition different to the aluminosilicates, i.e in blends with blast furnace slag [[51], [52], [53]], blast furnace slag-fly ash [54] and waste silica soda lime glass [55] with interesting effects. Moreover, LS has been recently reported in novel alkaline binders as the only precursor activated with waterglass [56] forming reaction products of type calcium silicate hydrates.

In all cases above mentioned for alkali activated MK, the mixtures of pastes were designed by fixing the amount of alkaline activator relative to the mass of the precursor of 100%MK; however, the amount of alkaline activator was not adjusted when the LS was added, which did not reduce the high amounts of alkalis in the binders, so the cost and environmental impact were not significantly reduced. High amounts of alkali have two sides, on the one hand, can promote a greater dissolution of MK favoring the strength, but on the other hand it can result in an excess of unreacted free alkalis that produce efflorescence and brittleness in the hardened binders [57], which compromises the durability.

The alkali activated blends of MK-LS appear to be suitable alternative cements in terms of strength, durability, practicability, scalability and environmental impact; nonetheless, those criteria must be systematically studied in greater depth in order to take them to the industrial level. The experimentation based in statistical methods, such as the response surface methodology, is a robust route to obtain systematized information to ensure the quality and the desired properties in the materials [52,58,59]. Such methodology allows to study the behavior of the independent variables as well as to optimize the response variable according to desired specifications. We hypothesized that throughout the use of statistical methods, it is possible to design high performance alkali activated blends of MK-LS using high loads of limestone, reducing the amount of alkaline activators while reducing also the environmental impact and costs. Based on previous experience in the laboratory, this work presents a systematized study in pastes of alkali activated MK-LS blends incorporating up to 80% limestone, with statistical modeling and optimization of the compressive strength. In addition, CO₂ emissions, energy consumption and cost of production are comparatively analyzed for the alkali activated MK-LS blends and a reference of a blended Portland cement (BPC).