Nanoscale modelling of ionic transport in the porous C-S-H network

To reduce the CO2 footprint of construction materials, concrete producers blend their cement with Supplementary Cementitious Materials (SCMs). SCMs such as fly ash or blast furnace slag are mostly the byproducts of other industries. And while SCMs are chosen to match the properties of the common Ordinary Portland cement (OPC), their addition to the cement recipe may alter the chemistry of the system. These changes may lead to different mechanical and transport properties of the cement-based structure which may, in turn, affect its long-term resistance to harmful external agents. Therefore, understanding the relationship between cement’s microstructure and the degradation mechanisms is key for optimizing the design of new cementitious materialsIn this context, chloride attack is the most common reason for steel rebars to corrode especially when exposed to external chloride (seawater, deicer salts
). At low w/c ratios (typically 13) pore solution, the C-S-H surface develops a negative surface charge density. Electrostatic interactions between the surface and the ions in the pore solution result in a redistribution of the ionic species in two layers of charge at the interface of the solid i.e. the Electrical double layer (EDL). In nanoscopic pores, the EDL is dominated by atomic phenomena which are thought to interfere with the mobility of ions and chloride in particular. In order to understand and quantify the surface effects and their influence on ionic transport, we firstly propose an atomistic model of the EDL formation based on the use of the Metropolis Monte Carlo algorithm. This model is used to compute the ionic distributions and electrochemical potentials of electrolytes at equilibrium in nanoscopic pores. These quantities constitute the main driving forces of ionic transport at the pore scale. The microstructure parameters including the surface charge density of C-S-H, the ionic strength (and pH) of the pore solution, and the pore size are equally investigated and their effect on chloride’s behavior quantified. Among the other parameters, the model also provides quantitative information on the effect of calcium ions which are thought to play a major role in the binding of chloride on C-S-H. The next step consists in using the calculated atomic-scale properties of the EDL in order to resolve the transport problem at the pore scale and compute microscopic diffusivities of chloride. This is achieved by using the molecular computations from the Monte Carlo (MC) engine in order to implement a modified version of the classical Poisson-Boltzmann system. The method is compared to the classical Finite element analysis of the Poisson-Nernst-Planck (PNP) equations and the data are discussed in the light of established experimental results in the literature.