High-temperature mechanical performance and microstructural insights of polypropylene fiber-reinforced fly ash–slag geopolymer mortars

Fly ash (FA)–ground granulated blast furnace slag (GBFS)-based geopolymer mortars are promising low-carbon alternatives to ordinary Portland cement for structures exposed to elevated temperatures. Polypropylene fibers (PF) are among the most widely used fibers in cementitious composites because of their effectiveness in controlling cracking and improving toughness. However, comprehensive experimental and predictive evaluations of how PF reinforcement and FA–GBFS composition jointly influence the high-temperature mechanical performance and microstructural evolution of such materials remain limited, highlighting the need for optimized mixture design for high-temperature structural applications. This study investigates the effects of replacing FA with GBFS and incorporating PF (0.3–0.9 vol%) on the fresh, mechanical, durability, and high-temperature performance of geopolymer mortars. Twenty mixtures activated with sodium silicate and 10 M NaOH were evaluated at ambient temperature and after exposure to 300–900 °C. The results indicate that increasing GBFS content and PF dosage reduced workability but improved mechanical strength and durability. The mixture containing 100% GBFS and 0.6 vol% PF achieved the highest compressive strength (62.9 MPa) and the best overall mechanical performance, while the highest flexural strength (7.03 MPa) was obtained with 0.9 vol% PF. Strength increased slightly up to 300 °C because of continued geopolymerization but decreased markedly beyond 600 °C owing to fiber decomposition and thermal cracking. FA-rich mixtures exhibited greater thermal stability, while microstructural analysis confirmed matrix densification at intermediate temperatures and void formation after severe heating. Machine-learning models outperformed MLR (R² = 0.702–0.714), with XGBoost providing the best overall performance (R² = 0.937 for flexural strength, 0.938 for compressive strength, and 0.987 for mass loss). SHAP analysis identified temperature as the dominant parameter. These findings provide practical guidance for designing PF-reinforced FA–GBFS geopolymer mortars for high-temperature applications and demonstrate the value of explainable machine-learning models for mixture optimization.