Effect of seawater on the reactivity, microstructure, and strength of Roman-inspired lime-pozzolan binders

Abstract

Roman marine concretes were frequently produced using seawater, yet the influence of seawater on lime-pozzolan reactions remains poorly quantified. In this work, pastes and mortars prepared with slaked lime and a natural pozzolan from the Phlegraean Fields were mixed with either Mediterranean seawater or tap water at lime-to-pozzolan ratios of 1:2 and 1:3. Reaction kinetics, hydrate assemblage, pore solution chemistry, pore structure, and mechanical performance were evaluated using calorimetry, TG-DTG, XRD, SEM-EDS, mercury intrusion porosimetry, and compressive strength testing. Seawater-mixed pastes released more heat and consumed portlandite more rapidly than their tap-water counterparts, and the corresponding mortars developed higher strength. Cl-hydrocalumite phases were identified in seawater mixtures, whereas mono-carboaluminate was observed only in the tap-water systems. The pore solution contained lower concentrations of chloride (up to ~74% less, depending on the extraction method) and sulfate (~82–93% less depending on the extraction method) than expected from simple dilution; the apparent chloride reduction (about 5–74%) is method-dependent, and the chloride-rich AFm identified by XRD and SEM-EDS provides independent evidence that part of the chloride was incorporated into solid phases. Seawater also refined the pore network, reducing the modal pore diameter by about 30% and producing a bimodal pore-size distribution. Despite the increase in pozzolanic bound water, the total bound water changed little during hydration, suggesting that strength development was associated primarily with pore refinement rather than with an increase in total hydrate content. The results indicate that seawater influenced both the chemistry of hydrate formation and the evolution of the pore structure in lime-pozzolan binders.