Dependence of compressive strength on phase assemblage in cement pastes: Beyond gel-space ratio - Experimental evidence and micromechanical modeling

Cement paste is a complex material, with a porous microstructure spread over several orders of magnitude, filled with water and gas, and a solid matrix composed of many different minerals - unreacted anhydrous products and crystalline or amorphous hydration products. The relationship between this complex phase assemblage and its mechanical properties is usually described through the 'gel-space ratio' descriptor, which links compressive strength to the ratio between the volume of hydration products to the sum of volumes of hydration products and capillary porosity. Surprisingly, little has been done on how the nature of hydrates impacts on the mechanical properties of cement paste. This is partly due to the difficulty to quantify the different hydrates in the system, and partly due to the absence of morphology homogenization models enabling one to relate the mineral assemblage to its mechanical properties. In this work, we generated experimental results in order to explore the 'mechanical efficiency' of C-S-H compared to portlandite, by hydrating C3S or C2S. The experimental results show that C-S-H, in the domain explored, is the first-order parameter explaining compressive strength. On this basis, a micromechanical model was developed, in which the smallest scale consists in solid C-S-H and total porosity (self-consistent scheme). Residual anhydrous cement and portlandite are added in a second homogenization step, as inclusions (Mori-Tanaka scheme). Additional experiments were performed to assess the role of hydrated calcium sulfoaluminates by hydrating synthetic clinkers or blends made of C3S, C3A and calcium sulfate. Using the same model, we show that monosulfoaluminate (and all non C-S-H hydrates) can be considered, as for anhydrous cement and portlandite, as inclusions in the C-S-H + total porosity matrix.