The objective of this study is to quantify the contributions of microstructure and molecular size of diffusing species to tortuosity, constrictivity and effective diffusivity. The microstructural effect is simulated with different sound, leached or carbonated cement pastes with varying water to cement ratios and limestone replacement filler replacements. Leached and carbonated samples were obtained by accelerated experiments: leaching by immersing samples in ammonium nitrate solution and carbonation by subjecting the samples to pure CO₂ at elevated pressure. To characterise the microstructural properties, Mercury Intrusion Porosimetry (MIP), and N₂-adsorption were used. The effect of molecular size is quantified with a recent developed diffusion setup allowing for simultaneous measurement of the diffusion of species with different molecular sizes. Previously developed models were also used to verify and give insights into the evolution of diffusion of degraded materials. In addition, a larger dataset from literature is used to evaluate a model which accounts for molecular size as well to predict diffusivity. Resutls show that because of the significant contribution of the molecular size of the diffusing species to the diffusion process, the constrictivity and thereby geometric factor may not be considered as intrinsic properties of the cement pastes. The geometric factor and/or constrictivity of cement pastes depends on the interrelationship of the molecular size of the diffusing species with the microstructure of the cementitious materials. A smaller diffusing species or/and a higher porosity of the sample results in a lower value of geometric factor. Interestingly, constrictivity is significantly influenced by the molecular size, but not the porosity.

@en

The role of capillary pores percolation and gel pores are investigated to explain the underlying differences between relative diffusivity obtained from different experimental techniques using microstructures generated from two different types of hydration model viz., CEMHYD3D (a voxel based approach) and HYMOSTRUC (a vector based approach). These models provide microstructures with different capillary pore connectivity for the same degree of hydration and the same porosity due to the underlying assumptions. In order to account for a C-S-H diffusivity at the micro-scale, a continuum micro-mechanics based model has been proposed. These simulations show that deperolation of capillary pores at around 20% of capillary porosity is essential in order to correctly predict diffusivity of cement paste with water-cement ratio by mass (w/c) in between 0.4 and 0.5. Furthermore from our analysis we present a viable postulate that the higher diffusivity measured by electric resistivity compared to other methods is due to differences in contribution from gel pores. For electrical resistivity measurement it is proposed that all gel pores are diffusive whereas for ion and tracer transport it is proposed that only nitrogen accessible gel pores are diffusive.

@en

In this paper, a newly developed lattice Boltzmann method based reactive transport model to simulate changes in microstructure of ordinary Portland cement paste due to calcium leaching is presented. The model takes three-dimensional digitized cement paste microstructure as input and is capable to capture an evolution of microstructure due to leaching, accounting for the dissolution of portlandite and corresponding increase in capillary porosity and the decalcification of C-S-H resulting in increase in gel porosity. The developed model has been applied to microstructures generated using two cement hydration models, CEMHYD3D and HYMSOTRUC, for three water-to-cement ratios. It was observed that the rate of leaching is directly proportional to ability of microstructure to transport calcium ions and higher fraction of percolated capillary pores result in higher rate of leaching. The model qualitatively reproduces experimentally observed changes in cement paste porosity and pore size distribution due to leaching. The quantitative validation of model at this scale is not possible by comparison of leaching obtained experiments and simulations which can be attributed to several factors including the differences in the scales of experiment and modelling study presented in this paper.

@en

Because of environmental and economic benefits, a fraction of cement is increasingly replaced by limestone fillers raising a question on to what extent limestone fillers affect the durability of cementitious materials. This work aims at understanding the effects of water/powder (w/p) ratio and limestone filler replacement on water permeability of cement pastes. A newly proposed technique using a controlled constant flow concept was applied to measure permeability of hardened cement paste samples following a factorial experimental design. It was observed that both limestone filler and w/p ratio significantly influence the water permeability. At a given w/p ratio, adding limestone filler made the microstructure coarser, especially for high w/p ratio. Nevertheless, if the comparison is based on a given water/cement (w/c) ratio instead of w/p ratio, the limestone filler replacement refined the microstructure in terms of capillary porosity and pore size distribution, resulting in permeability decreases of cement pastes. Furthermore, a modified Carmen-Kozeny relation was established which enables prediction of the permeability from capillary porosity and the critical pore diameter. @en

This paper provides a comprehensive overview of existing experimental and modelling approaches to determine effective diffusion coefficients of water saturated ordinary Portland cement-based materials. A dataset for diffusivity obtained from different experimental techniques have been presented for cement paste, mortar and concrete. For cement paste at low porosities, diffusivity reported by different authors varies up to a factor of five and electrical resistivity measurements for low capillary porosity are up to one order of magnitude higher compared to other techniques. Experimental data of mortar and concrete reveals predominant influence of increasing tortuosity due to aggregates and limited influence of interface transition zone. Hence, a particular emphasis has been placed on assessing predictability of diffusivity models for cement paste on a larger dataset collected in this paper. It has been observed that all predictive models have similar level of accuracy and fail to predict electrical resistivity data at low capillary porosity as these models are not calibrated using electrical resistivity data.

@en

Underground concrete structures in radioactive waste disposal have the potential to be subjected to a high hydrostatic pressure and the surrounding environment may contain a high dissolved CO₂ concentration. Therefore, a combination of diffusion and advection should be taken into account when one considers the carbonation mechanism. This study aims at developing a model to predict the evolution of the microstructure and transport properties of hardened cement pastes due to carbonation under accelerated conditions in which a pressure gradient of pure CO₂ is applied. The current model is improved from the preliminary model in terms of extension to limestone cement paste and accounting for the transport of moisture. The proposed model is based on a macroscopic mass balance for CO₂ and moisture in both gaseous and aqueous phases. A simplified solid-liquid equilibrium curve is used to relate the Ca content in aqueous and solid phases. Besides the prediction of the changes in porosity, diffusivity, permeability, and saturation degree, the model also enables prediction of the carbonation degree, portlandite content, and CO₂ uptake. Verification with experimental results from accelerated carbonation tests shows a good agreement.

@en

Calcium leaching might be a significant degradation process in concrete and reinforced structures with an anticipated long-term service life such as nuclear waste disposal systems or large hydro structures (dams, bridges, water tanks). The leaching process is extremely slow under environmental conditions, which fosters the use of accelerated experimental approaches such as leaching in an ammonium nitrate (NH₄NO₃) solution. In this paper, we develop a one-dimensional diffusion-based transport model with the purpose to predict the changes in microstructure and transport properties of saturated cement pastes in contact with a NH₄NO₃ solution. The model helps to better understand the transient state of leaching which is difficult to observe by experimental work. The main new elements in this model are change in model configuration with extended solution domains; ability to predict the spatial profiles of diffusivity and permeability; including the effect on solubility of the spatial-temporal evolution of nitrate concentration; and including the effect of limestone addition to the cement paste of leaching kinetics. This model is based on macroscopic mass balances for Ca in aqueous and solid phases which are linked together by applying a variable solid-liquid Ca equilibrium curve. Besides the prediction of the leached depth, porosity increase, portlandite and C-S-H contents, and the amount of leached Ca, the model also enables to estimate the variation of permeability and diffusivity over the domain at different immersion periods in NH₄NO₃ solution. The model is verified by accelerated leaching experiments in 6 mol/l NH₄NO₃ solution on CEM I cement pastes with/without limestone fillers. Verification with experimental results shows a good agreement.

@en

Material microstructure defines its physical and mechanical properties. With this idea, different models have been developed to effectively determine the physical and mechanical properties of a given microstructure. When the microstructure alters as a consequence of chemical processes, coupling of transport with geochemical models is essential for further extension of predictive capabilities. In order to simulate alteration of the microstructure of cement paste during chemical degradation, a new threedimensional pore scale reactive transport code has been developed. The model has the capability to account for reactive transport processes in two types of pore space representations: (1) an explicit representation of pores and solid phases for capillary pores and (2) a continuum representation for unresolved pores (pores with sizes below the spatial discretization) embedded in the solid phases. Multicomponent transport is modelled using a lattice Boltzmann (LB) method. The transport model is coupled with an established geochemical solver (iPHREEQC). Due to dissolution and precipitation reactions, both the explicit pore and the continuum domains are updated using, respectively, static update rules and constitutive relations. The model is applied to a virtual cement microstructure under the process of calcium leachinig. For calcium leaching in cementitious systems we further show the results where the chemistry is represented by a abstracted chemical model. This model abstraction allows for numerically more efficient simulations which allows for parametric studies. Finally, comparison between the simulated results and an experiment carried out under accelerated conditions is presented. @en