Advances in Research

Freeze–thaw deterioration of cement-based materials is commonly assessed through mass loss, relative dynamic modulus, strength retention, or visual scaling, yet these macroscopic responses are downstream manifestations of processes that originate in a heterogeneous and evolving pore system. This critical narrative review examines the coupling between pore architecture, moisture state, ice formation and redistribution, microcracking, and macro-mechanical deterioration in cement paste, mortar, conventional concrete, high-strength and ultra-high-performance concretes, and selected modified systems. Literature published principally from 2000 to 17 July 2026 was considered, while earlier foundational studies were retained where essential to the mechanistic framework. The evidence indicates that total porosity alone is an inadequate state variable for frost damage. Freeze–thaw response depends more fundamentally on the joint distribution of pore and throat sizes, pore connectivity, entrained-air geometry, degree of saturation, freezable-water fraction, interfacial transition-zone quality, and the emergence of connected microcracks. Hydraulic pressure, crystallisation-related stress, cryosuction, osmotic effects in saline pore solutions, and thermomechanical mismatch are best interpreted as interacting processes rather than mutually exclusive theories. Across experiments, reductions in dynamic stiffness and tensile-sensitive properties often precede conspicuous mass loss or large compressive-strength reductions, reflecting the progressive loss of load-transfer continuity before gross material removal. Evidence from mercury intrusion porosimetry, low-field nuclear magnetic resonance, X-ray computed tomography, calorimetry, microscopy, and nanoindentation shows broadly consistent pore coarsening and connectivity growth, but quantitative relationships remain method-dependent because each technique interrogates different pore domains and introduces distinct preparation or inversion assumptions. The review therefore proposes a coupling framework centred on moisture-conditioned pore topology and crack connectivity rather than cycle count or porosity in isolation. Future progress requires harmonised freeze–thaw conditioning, multimodal in situ characterisation, three-dimensional topology metrics, uncertainty-aware cross-scale models, and validation against field moisture and temperature histories.