Flow Loss in Superplasticized Cement Pastes

Abstract

Superplasticizers are polymeric dispersants widely used in concrete technology. Their last generation, polycarboxylate ethers-based comb copolymers (PCEs), are the most efficient ones providing excellent dispersion as well as enhanced workability, durability, and strength development. One of the most relevant problems of superplasticized concrete is the flow loss over time. In practice, an unexpected flow loss could lead to serious problems of placement and finishing of concrete. Such problems are usually improperly solved by re-dosing the PCE or – even worse – by adding water on the job site. The extent to which flow loss occurs depends – among several factors – on the type and dosage of the admixture. Flow loss in superplasticized systems has been mainly explained in qualitative and comparative ways over the past years. This is due to the intrinsic complexity of the underlying mechanism that must involve a change in the agglomeration degree as a result of minor advancements of the cement hydration. Specifically, increasing specific surfaces are expected to modify over time the dispersion degrees given by PCEs that themselves modify hydration kinetics. From an experimental point of view, the lack of a reliable methodology for measuring specific surfaces must have additionally discouraged researchers from attempting to understand the phenomena of flow loss in quantitative terms. A main initial focus of this thesis has therefore been to establish an ensemble of reliable analytical methods to study fresh cement pastes to obtain the data that has been missing in the literature to confidently attack the question of flow loss mechanisms. The evolution of the flow properties was investigated on cement pastes with PCEs, specifically examining how flow loss is modified in relation to the molecular structure of polymers. The dosages were also varied to cover a broad range of initial yield stress values. This allowed to properly account for changes in particle size, specific surface area, and adsorbed polymer over time, making possible an accurate rationalization of the yield stress evolution. A strong correlation between yield stress evolution and hydration kinetics expressed as heat rate after the onset was established. The onset time marks the clear transition from the induction to the acceleration period after which the yield stress grows exponentially. For the first time, it was possible to demonstrate that two distinct processes – reversible and irreversible – occur before and after the onset, showing different dependencies on hydration kinetics that cannot be described by a single expression. The increase of the yield stress after the onset depends on the molecular structure and dosage of PCEs and the increase of the heat rate. After the onset, hydration rate and specific surface changes are proportional one to the other. This implies that the mechanistic origin of flow loss lies in the formation of additional specific surface at the early stages of hydrates nucleation and growth rather than in the heat rate per se. Introducing the relation between yield stress and specific surface, it was possible to quantitatively account for the exponential increase of the yield stress through a combination of two main factors: a change in particle size and a change in surface coverage, both induced by the polymers. An essential and original aspect of this result involves having established the exponential dependence of yield stress on the amount of adsorbed polymer per surface unit of cement. The physical interpretation of the flow loss in relation to the molecular parameters of PCEs and cement composition reached in this thesis can be considered as one of the holy grails of research and development on chemical admixtures. The ability to properly tune concrete fluidity over time is of major practical relevance, facilitating the implementation of new generations of blended cements and the development of robust admixtures formulations.