Cement, as a construction material, has low thermal resistance, inherent fire resistance, and is incombustible up to a certain degree. However, the loss of its mechanical performance and spalling are its primary issues, and it thus cannot retain its performance in refractory applications. The present study explores the performance of geopolymer formulations that have excellent fire resistance properties for potential refractory applications. This study is unique, as it investigates advanced solid geopolymer formulations that need only water to activate and bind. Various solid geopolymer formulations with fly ash as a precursor; potassium hydroxide and potassium silicate as activators; and mullite and alumina as refractory aggregates were studied for their compressive strength at up to 1100 °C and compared with their two-part conventional liquid alkaline geopolymer counterparts. Advanced solid geopolymer formulations with mullite and alumina as refractory aggregates had mechanical strength values of 84 MPa and 64 MPa post-1100 °C exposure and were further exposed to ten thermal cycles of 1100 °C to study their fatigue resistance and post-exposure compressive strengths. The geopolymer sample with mullite as a refractory aggregate yielded 115.2 MPa compressive strength after the fourth cycle of exposure. This sample was also studied for its temperature distribution upon direct flame exposure. All the geopolymer formulations displayed a drop in compressive strength at 600 °C due to viscous sintering and then a rise in strength at 1100 °C due to phase transformation. X-ray diffraction studies revealed that the formation of crystalline phases such as leucite, sanidine, and annite were responsible for the superior strengths at 1100 °C for the alumina- and mullite-based geopolymer formulations.

The present study focuses on the mechanical performance of steel and polyvinyl alcohol fibers embedded in the geopolymer matrix. A high-strength geopolymer concrete with fly ash, slag and silica fume as precursors and sodium hydroxide and sodium silicate solutions as activators has been tested for its strength in compression and flexure. The influence of fibers on flowability, long-term shrinkage and sulphuric acid attack on the geopolymer concrete has also been studied. The dosage of fibers was maintained at 1%, 2% and 3% by volume, and fibers of length 13 mm have been used in the study. Results indicate that slag with 3% steel fibers by volume had a predominant influence on the strength development of steel fiber-reinforced geopolymer concrete, yielding a compressive strength of 107 MPa after 28 days. Blast furnace slag resulted in increasing the shrinkage of concrete due to rapid gel formation owing to the presence of calcium ions, although the fibers helped reduce the shrinkage to some extent. The strength of steel fiber geopolymer concrete was superior to PVA fiber geopolymer concrete; however, after an acid attack, the strength of steel fiber geopolymer concrete was reduced more than PVA fiber geopolymer concrete due to the enhanced corrosion resistance of PVA fibers.

Portland cement production accounts for roughly 8% of anthropogenic CO2 emissions, driving interest in low-carbon geopolymer binders. One-part (“just-add-water”) geopolymers, which replace hazardous liquid activators with a dry, pre-blended solid activator, are especially suited to field deployment where handling safety and logistics are decisive. However, their formulation space is combinatorially vast, and trial-and-error development cannot efficiently navigate it. This paper reviews one-part geopolymer science, presents a new comparative and interpretable ML analysis of a published 80-mixture one-part fly-ash/GGBS geopolymer dataset from twelve studies, and proposes an AI-assisted design framework. The ML demonstration targets 28-day compressive strength only. Under leave-one-source-out (LOSO) cross-validation—the appropriate test for a literature-pooled dataset—gradient-boosted trees achieved R2 = 0.61, well above a linear baseline (0.36), suggesting that non-linear structure transfers across studies; a random split gives a higher but less reliable R2 = 0.90 on only 16 test mixtures. Because fly-ash and GGBS contents are near-perfectly anti-correlated (r=−0.99), we model the precursor axis as a single slag-fraction descriptor; SHAP then identifies this precursor balance and the activator’s Na2O dosage as the dominant, mechanistically coherent strength levers. Demonstrated for strength, the framework offers a transferable route toward multifunctional low-carbon binders for protective and infrastructure applications.

Portland cement production accounts for roughly 8% of anthropogenic CO₂ emissions, driving interest in low-carbon geopolymer binders. One-part (“just-add-water”) geopolymers, which replace hazardous liquid activators with a dry, pre-blended solid activator, are especially suited to field deployment where handling safety and logistics are decisive. However, their formulation space is combinatorially vast, and trial-and-error development cannot efficiently navigate it. This paper reviews one-part geopolymer science, presents a new comparative and interpretable ML analysis of a published 80-mixture one-part fly-ash/ground granulated blast-furnace slag (GGBS, hereafter slag) geopolymer dataset from twelve studies, and proposes an AI-assisted design framework. The ML demonstration targets 28-day compressive strength only. Under leave-one-source-out (LOSO) cross-validation—the appropriate test for a literature-pooled dataset—gradient-boosted trees achieved R² = 0.61 (RMSE = 15.5 MPa; 95% bootstrap confidence interval on R², 0.44–0.75), well above a linear baseline (0.36), suggesting that non-linear structure transfers across studies; a random split gives a higher but less reliable R² = 0.90 on only 16 test mixtures. Because fly-ash and slag contents are near-perfectly anti-correlated ($r=-0.99$), we model the precursor axis as a single slag fraction descriptor; SHAP then identifies this precursor balance and the activator’s Na₂O dosage as the dominant statistical predictors of strength in this dataset, an ordering consistent with known activation chemistry; causal confirmation of these associations awaits the experimental validation stage of the proposed framework. Demonstrated for strength only, at paste level, the framework offers a transferable route toward multifunctional low-carbon binders for protective and infrastructure applications; the multifunctional extensions are proposed, but not yet demonstrated.