Transport in concrete with new CO₂ reduced cements - Reactive Transport Model for Durability Estimations

Concrete is the most used building material in the world due to its cheap price and desired properties such as strength, workability and durability. Concrete is a composite material obtained by mixing cement, water, aggregates and small amounts of chemical additives. Following the mixing, a (fast) chemical process called hydration is initiated in which the different cement clinkers react with water and form different hydration products, such as calciumsilicate-hydrate (C-S-H), ettringite and monosulfate.

The hydration process can continue slowly over the lifetime of concrete structures and is influenced by the so-called service environment, which is the environment surrounding the concrete. In aggressive service environments, for instance, in the case of marine exposure, the altering of hydrated cement can lead to the end of its service life.

Most of the altering processes are related to moisture and gas transport within the pores of the concrete. This thesis was part of a project aiming to provide the basis for a green conversion of cement and concrete production in Denmark. The thesis contributes to a better understanding of long-term consequences of implementing new CO2-reduced cement in concrete through experimental investigations and reactive mass transport modeling.

CO2-reduced cement refer to bindersystems where amounts of the Portland cement is replaced with supplementary cementitious materials such as calcined clay and limestone filler, in order to decrease the CO2 emission from the production of concrete. Three major areas of focus related to moisture and gas transport in cement-based systems were covered.

First focus area was the explanation of vapor transport through a partially saturated capillary pore. In unsaturated porous materials, the presence of trapped fluid in a capillary pore increase the overall rate of vapor transport rather than blocking it. This is simply due to the fact that the trapped liquid can be considered as a bridge shortening of the apparent distance the vapor needs to travel through the pore system.

A simple approach was proposed describing the process with a single equation using the gradient of the chemical potential as the driving force for diffusion, in this case, equal to the gradient of the relative humidity. The model was established without the use of fitting parameters. The proposed method was compared with, experimental results for isothermal vapor transport through a partially saturated cylindrically symmetric capillary tube of variable cross-sectional area, with excellent agreement.

The study concluded that the enhancement of vapor transport in the presence of trapped pore liquid can be modeled using the geometry and the relative humidity at the gas-liquid interface. It is nearly impossible to directly apply this model to cement-based materials since it is not possible to accurately describe the geometry and boundary conditions at each gas-liquid interface, due to the complex microstructure of the pores of hydrated cement.

However, the study provided insight into one important pore-scale vapor transport mechanism in unsaturated systems. Furthermore, the study also demonstrated the conceptual advantage of using the chemical potential in modeling vapor transport at the pore-scale. Secondly, liquid water and water vapor transport properties relevant to concrete durability modeling were investigated.

Moisture properties of ten different CO2-reduced cement-based binder-systems containing supplementary cementitious material in the form of fly ash, calcined clay, burnt shale and gray micro-filler, where investigated using four different experiments being, (i) sorption tests (moisture fixation), (ii) cup tests in two different relative humidity intervals, (iii) drying tests, and, (iv) capillary suction tests. An inverse approach was developed for determining a separate description of the effective diffusion in the liquid phase and in the vapor phase as a function of the saturation degree, relevant for the proposed continuum moisture transport model adopted.

The inverse approach in all essential parts uses the measured mass change over time, as obtained from all the experimental investigations, to inversely obtain the effective diffusion parameters. Due to the use of multiple diffusion experiments both in absorption and desorption, covering different relative humidity intervals, the moisture properties obtained with the proposed inverse analyses method provide a good description valid for a wide range of cases of the moisture transport for the ten different binder-systems.

The proposed method does not explicitly account for the effect of pore structural change over time, but it provides a good description of moisture properties at the conditions the samples were tested at. Accurate models for moisture transport is crucial for more involved durability models also including ionic and gas transport features.

Thirdly, a multi-phase reactive mass transport framework for durability estimation of cement-based materials was developed. The goal was to include the gas phase alongside the implemented liquid phase and solid phase developed in earlier works. The addition of the gas phase into an earlier developed framework includes the description of gaseous transport in the air-filled space and chemical gas-liquid interaction.

The governing system of equations includes a modied version of the Poisson-Nernst-Planck system of equations including gaseous transport in the air-filled space, ionic transport in the liquid phase, electro-migration of ionic species, two-phase moisture transport consisting of water vapor and liquid water, and sorption. A stringent model for mass transport also including for gaseous constitutes contributes to a better understanding of the true mechanisms responsible for the altering of hydrated cement-based materials in unsaturated conditions as well as in cyclic drying-wetting conditions.

Furthermore, the inclusion of gaseous constituents into the framework enhances the model's accuracy and enables the investigation of combined effects of different degradation processes acting simultaneously, for example, carbonation and chloride ingress. Different numerical examples are included demonstrating applications of the above mentioned added futures.