Multiscale computational modelling of nano-silica reinforced cement paste: Bridging microstructure and mechanical performance

Abstract

This study presents a systematic model to optimise nano-silica utilisation in cement paste and predict its nonlinear behaviour. The model integrates a hydration model, which calculates the dissolution rate of each clinker mineral, with a thermodynamic model that simulates the hydration reaction and interaction between hydrates and nano-silica. The investigation considered different replacement levels of nano-silica (2 and 4% by weight of cement) to analyse the phase assemblages and porosity of nano-silica reinforced cement paste. The reaction between nano-silica and the hydrate (portlandite) was meticulously accounted for by incorporating the formation of calcium-silica-hydrate (C-S-H) with a realistic transition from jennite-type C-S-H (Si/Ca ratio of 0.58) to tobermorite C-S-H (Si/Ca ratio of 0.67). The coupled model was validated against the experimental results obtained from thermogravimetric analysis, X-ray diffraction, and scanning electron microscopy, ensuring its reliability. Subsequently, the model was used to compute the volume fractions of various phases including hydrates, unhydrated clinkers, pores, and unreacted nano-silica. A representative volume element (RVE) was formulated for the cement pastes with and without nano-silica using MATLAB. The RVE was further evaluated through finite element analysis using COMSOL Multiphysics, enabling the computation of homogenised material properties such as compressive strength, and this aligned well with experimental findings. In summary, the proposed systematic model provides a realistic prediction of the nonlinear behaviour of cement pastes with different nano-silica replacement levels. Its applicability extends to the optimisation of cement-based composites for diverse engineering applications.