AI-driven Self-healing of Geopolymer Composite for Extreme Environments

Abstract

Geopolymer composites are increasingly regarded as promising materials for infrastructure operating in extreme environments, including high-temperature, high-pressure, chemically aggressive, and multi-hazard conditions. However, passive durability alone is not sufficient to ensure long-term performance under such coupled extreme conditions. Once cracking occurs, transport pathways can rapidly form for fluids, ions, and reactive species, leading to permeability increase, loss of sealing integrity, structural degradation, and premature failure. This Perspective argues that the central challenge is therefore not simply to improve the intrinsic stability of geopolymers, but to design them for functional resilience through self-healing under coupled thermal, chemical, and pressure-driven stresses. We further contend that many current self-healing strategies remain overly generic and often fail to account for the distinct demands of different extreme environments. To address this gap, we propose a shift from empirical trial-and-error and forward property prediction toward physics-informed, mechanism-aware AI frameworks capable of inverse design, multimodal data integration, and digital-twin-assisted adaptation. Such an approach could enable geopolymer composites that are not only durable, but also damage-tolerant, recoverable, and better suited for long-term service in extreme environments.