The Mechanisms Governing the Robustness of Fresh Self-Compacting Concrete

This thesis deals with the robustness of fresh self-compacting concrete (SCC). As no external compaction is necessary, many problems caused by a poor compaction are evaded and a denser reinforcement or more complex formworks can be applied when using SCC. Although reduced construction times, less labour costs, and an improved pumpability encourage the usage of SCC, an inferior robustness – the capacity to tolerate small variations in the mix proportions, material properties and procedures – limits the general use or SCC. A more severe quality control of all constituent materials and a more experienced staff is needed.

SCC has a more complex mix design compared to vibrated concrete. Balancing in between a sufficient flowability allowing air bubbles to escape and a loss of cohesion due to segregation or bleeding, the impact of small changes on the fresh properties is much larger. This thesis aims to investigate the physics governing the robustness, and especially the role of the microstructure and particle clustering on the robustness. Due to the large number of influencing factors, not every aspect governing the robustness could be studied. By focusing on the fundamental mechanisms involved during the analysis of the experimental results, the results can be generalized to other situations or parameters. To study the particle clustering in cement paste more directly, the potential use of new experimental techniques such as ultrasound attenuation and Small Angle Light Scattering (SALS) was evaluated.

A concise overview of the fundamental mechanisms and interactions governing the microstructure of fresh SCC is given in the literature review (Chapter 3). The chemical reactions during the first hours of hydration, the forces and interactions in between particles, the working mechanisms of admixtures, and the impact of shear stresses are explained in more detail in order to obtain a better understanding of the microstructure.

Based on the experimental results, methods to improve the robustness of fresh SCC were developed. Dependent on the mechanism providing stability in SCC, another strategy should be used. As different applications have different workability demands, SCC with proper rheological properties customized to the application should be used.

When the yield stress is providing stability, the sensitivity of this yield stress to small changes is determining the robustness of the mixture. Because such mixtures generally have a low plastic viscosity, changes in this plastic viscosity due to small variations have no major impact. Based on the experimental program and the yield stress model (YODEL), possible solutions to enhance the yield stress robustness of SCC have been proposed. A higher maximum packing density results in a more robust yield stress and can be achieved with a higher superplasticizer dosage, which can be reached with a lower paste volume. An alternative solution is the combination of a lower paste yield stress and a slightly higher plastic viscosity, for example by adding a VMA combined with a higher superplasticizer dosage.

In mixtures with a low or zero yield stress and the paste plastic viscosity providing stability, the sensitivity of the paste plastic viscosity is determining the robustness of the global mixture. A small decrease in the amount of water can result in an unworkable, sticky mixture; a small excess in water can result in bleeding. The best approach to improve the robustness of such a mixture is to enhance the robustness of the paste plastic viscosity. Based on a set of experiments and considering the Krieger-Doherty model, specific measures have been proposed. A first method is to lower the paste packing density, for example by raising the water-to-powder ratio. Also lowering the paste plastic viscosity helps to enhance the robustness against small variations in the mix composition. Mixtures prone to an excessive stickiness can be improved by reducing the thixotropic buildup, for example by increasing the water-to-powder ratio, a cement replacement by fly ash, or reducing the content of silica fume.

In case the risk of bleeding is limiting the applicability of SCC, a different approach is needed. Bleeding is a result of an imbalance in the intermolecular forces. When steric repulsion forces dominate over attractive Van Der Waals forces for small colloidal particles, the colloidal network is broken and syneresis will manifest itself as a layer of water floating on top of SCC. To eliminate the origin of bleeding, two methods exist: reducing the steric repulsive forces by a reduction of the superplasticizer dosage (higher water-to-powder ratio, higher paste volume), or increasing the attractive Van Der Waals forces inside the fresh concrete, for example by a replacement of cement by silica fume.

An attempt was made to investigate particle clustering and the structural buildup more fundamentally using innovative experimental techniques. Due to the very broad particle size distribution and the non-transparent opaque nature of cement paste, monitoring the flocculation and deflocculation of the individual cement particles is not evident. Two techniques have been selected for a more thorough examination of their applicability: ultrasound attenuation and Small Angle Light Scattering (SALS).

When ultrasound amplitude losses are measured over a range of ultrasound frequencies, attenuation spectra can be used to characterize an opaque dense solution. As no evolution in the attenuation spectra was observed on samples of cement paste at rest, the flocculation at rest cannot be evaluated using the ultrasound attenuation technique. Different water-to-powder ratios caused variations in the measured attenuation spectra. However, in an apparatus in which the particle size distributions were calculated based on a theory taking into account the viscous attenuation losses and the scattering attenuation losses, no realistic particle size distribution could be generated. Using another apparatus, in which the particle size distribution was calculated based on a calibration with a powder with known properties, realistic particle size distributions were obtained. As paste mixtures varying in water-to-powder ratio all generated similar particle size distributions, no effects of differences of the water-to-powder ratio on the particle clustering could be observed using ultrasound attenuation.

When Small Angle Light Scattering (SALS) was used on diluted samples of a solution of cement in water, the influence of adding a superplasticizer on the scattering patterns could be observed. Because the SALS images were independent on the shear rate applied on the samples, the flocculation / deflocculation mechanisms in cement paste subjected to shear cannot be studied more in depth using this technique. A calculation of the particle size distribution based on the SALS images was not successful. Although this technique is useful in polymer science, its application on cement paste is less promising.