Persistence of Tensile Strain Hardening in Engineered Cementitious Composites: Ageing, Variability and Implications for Cracked-State Durability—A Critical Review

Engineered cementitious composites (ECCs) accommodate tensile deformation through distributed microcracking, but their strain capacity is commonly assessed on small specimens at 28 days. This is a critical narrative review to determine whether strain hardening is retained with age, production variability, and exposure, and whether retained capacity provides crack control and claimed durability. The synthesis combines micromechanical aspects, multi-age tensile observations, test method effects, transport, and healing of the cracked state. There are categories of evidence, indicating the extent of investigation, but methodological reliability is evaluated independently. Existing reviews give background for an integrated assessment of retention and are not evidence that ageing has been neglected. The approximate 54% reduction in ultimate strain after three years indicated by primary hybrid-ECC results and the approximate 84% reduction after one year indicated by secondary results from a different PVA-ECC system demonstrate that the overall performance of hybrid ECCs is superior to that of conventional eco-friendly concrete. The fly-ash mixtures tested with a different curing history have about 2–3% absolute strain at 90 days. These results are a function of the mixtures and cannot be combined into a global ageing factor. Further reaction can alter the toughness of a matrix and fibre–matrix bonding; geometry, gauge length, and production variability make comparisons difficult. Transport can be restricted in tight cracks as compared to wide-cracked mortars; however, other factors such as crack number, deformation, and exposure also affect transport. Proposals for age-specific characteristic strain capacity and capacity–demand screening criteria are made for research. None sets persistence in the absence of direct proof of ageing tensile properties. For practical use, the statistical factors, crack-width limits, and service-life assumptions need to be validated.