Influence of the slag content on the hydration of blended cement

International audienceSlag is increasingly used for replacing a part of clinker. The addition of slag modifies the hydration process and so the physico-chemical properties (such as porosity, transport properties) of cementitious materials. Indeed the evolution of the physico-chemical properties are determined by the hydration kinetics of the anhydrous phases and the composition of the hydrated phases. The hydration process of slag blended cement is still under investigation due to the complex interactions between clinker and slag. The aim of this work is to link the evolution of slag blended cement properties (such as bound water, chemical shrinkage, mineralogy, porosity) to the hydration state. The latter was defined by the degree of hydration of the cementitious material.For this study, three commercial cements with variable amounts of the same slag (0%; 61%; 80%) were considered. Cement paste samples hydrated from one week to one year were characterized. $^{29}$Si MAS NMR experiments were performed in order to quantify the consumption of the anhydrous phases. A degree of hydration of clinker, slag and cement could be deduced from these experiments. The results are compared with the evolution of the bound water content (measure by thermogravimetric analysis which is strongly correlated to the global degree of hydration). For each measurement of the degree of hydration, the porosity was determined (total free water porosity measurement) and the mineralogy was described (thermogravimetric analysis (TGA) and X-ray diffraction (XRD)).Measurement of the degree of hydration shows that the slag used in this study is less reactive than other slags which are characterized in the literature. On the one hand the low reactivity of slag explains why the chemical shrinkage and bound water are much more important for a Portland cement than for slag blended cement. On the other hand, for fully hydrated cement, the chemical shrinkage and bound water are more important for slag blended cement than for Portland cement. The consumption of portlandite by the hydration of slag is less important than its production by the clinker even after one year of hydration. The slag degree of hydration is too low to underline an impact of the slag hydration on the porosity and the bound water content.A simplified hydration model, based on mass balance, was applied to determine the mineralogical composition of hydrated slag blended cement according to a specific degree of hydration. Stoichiometric equations used in this model were based on the hydrates compositions as detected by XRD or available in the literature. The model considers the consumption of portlandite by the hydration of slag and the incorporation of aluminum in the C-S-H. The results are used to estimate a volume of chemical shrinkage and bound water. The model has given promising results concerning the porosity and chemical shrinkage estimation. However the portlandite content predicted by the model is too low when it is compared to experimental measurements