The characterisation, improvement and modelling aspects of Frost Salt Scaling of Cement-Based Materials with a High Slag Content

The characterisation, improvement and modelling aspects of Frost Salt Scaling of Cement-Based Materials with a High Slag Content

Copuroglu, O.

Bijen, J. (promotor)
van Breugel, K. (promotor)

Civil Engineering and Geosciences

2006-05-09

Blast furnace slag cement concrete is used extensively in a number of countries. In comparison with OPC, it is particularly well known for its excellent performance in marine environments. One dis-advantage of slag cement is its vulnerability to scaling under the combined load of freezing-thawing and de-icing salts. The current investigation was triggered by positive observations regard-ing certain grinding agents used in slag cement production to improve frost salt scaling resistance. The investigation was aimed at explaining the cause of this improvement, at finding alternative methods to improve scaling resistance and at developing a model that would be suitable for the simulation of frost salt scaling behaviour. The investigation conclusions are essentially confined to high slag cement, particularly type CEM III 42,5/B which has a 67% granulated slag content. The w/c ratio of the paste, mortar and concrete specimens is generally maintained at 0,45. Carbonation, known as the critical parameter in frost salt scaling, constituted the key area of inter-est. From previous investigations it is known that carbonation increases porosity and coarsens the pore system in slag cement paste while it actually does the reverse in OPC paste. In the light of lit-erature a new hypothesis has been suggested that the transition zones, which are the weakest points in normal-performance cement-based materials, critically determine frost salt scaling resis-tance. These zones are even more indicative in the case of slag cement pastes because of the sig-nificant amount of transition zones that can be weakened by carbonation unlike with low-slag ce-ment or OPC pastes. In the present investigation it was observed that carbonation causes significant slag cement paste shrinkage. It was especially the transition zones between non-reacted slag particles and hydration products that were found to be affected. Consequently this process leads to the paste having a coarser pore structure thus making it prone to greater water uptake when compared to non-carbonated slag cement paste or OPC paste. The new hypothesis was supported by findings emerging from the ESEM study. It was observed that frost salt scaling attack generates cracks in the microstructure which adhere to slag-matrix interfacial zones. This was confirmed by nano-indentation tests which demonstrated that carbonation creates a significant number of weak zones in the slag cement paste. In the case of OPC paste the picture that emerged was quite different. Natural air carbonation influences the mineral characteristics of cement pastes. The XRD study re-vealed that both slag cement paste and OPC paste possess various types of carbonate minerals, namely: calcite, aragonite and vaterite. However, accelerated carbonation creates overwhelmingly stable calcite phases in both types of cement pastes which are subsequently transformed from me-tastable carbonates. This observation draws attention to the role played by Ca(OH)2 in the good scaling resistance of OPC or low-slag cement systems. A curing regime, especially curing in lime water, appears to be favourable for slag cement materials. However, when compared to the effect of carbonation, the influence that the curing water quality has on scaling resistance is minimal. The contribution made by prolonged water curing to scaling resistance could have been greater but, as it was, the curing periods were limited to 5 weeks in the interests of remaining realistic and practi-cal. The main goal of the project was to investigate the improvements in frost salt scaling resistance in-stigated by chemical grinding agents on the basis of the various positive results gained from the preliminary tests. The intention was to study the effects that the chemicals had on the cement paste microstructure in order to understand frost salt scaling resistance in slag cement concrete and so as to contribute to structural improvements in that area. A microstructural comparative study was carried out on slag cement pastes that contain alkanola-mines/hydrocarboxylates (the best performing ones) and diethylene glycolâbased (the worst per-forming example) grinding agents. The most notable difference was in the pore structure of the paste samples. Alkanolamines/hydrocarboxylates-based grinding agents were found to produce smaller pore sizes when compared to the ones containing diethylene glycol. This is consequently likely to give rise to higher carbonation resistance, lower water uptake and, eventually, to higher frost salt scaling resistance. However, the improvement achieved by alkanolamines / hydrocarboxy-lates is not sufficient to enhance the scaling resistance of the slag cement materials investigated in similar detail to OPC pastes. Another technique that was investigated was sodium monofluorophosphate (Na-MFP) surface treatment. Remarkable improvements in frost salt scaling resistance were achieved by applying a 10% Na-MFP solution to the surface of the carbonated slag cement paste and concrete. The scal-ing resistance improved by about 95% after 7 freeze-thawing cycles. Evidence was found pertain-ing to the reaction between Na-MFP and metastable carbonates in the carbonated slag pastes. The application appears to significantly increase the tensile strength of the carbonated slag cement paste which is extremely favourable in terms of scaling resistance. The study finally resulted in the development of a new integrated model. The model mainly takes into account the glue-spall theory and the hypothesis developed in this thesis and it runs on the Delft Lattice Model platform. The model successfully demonstrates the experimental observations and the crack patterns created by the scaling action. The glue-spall theory suggests that cement-based material surface scaling derives from external ice layer cracking due to further cooling. Cooling consequently generates tensile stress due to the shrinkage of ice and causes cracking when the stress exceeds the tensile strength of the ice. This theory can explain many phenomena including the pessimum effect. On the basis of this theory, the new integrated model proved to be capable of simulating two important experimental observa-tions. Under identical conditions the model can predict higher surface scaling at a 3% salt concen-tration level in relation to higher and lower values. The effect of ice layer thickness is furthermore found to be crucial with respect to frost salt scaling. Under identical material and environmental conditions the thicker external ice layer creates more damage than thinner ice layers. This observa-tion was also successfully demonstrated with the new integrated model.

frost salt scaling
slag
carbonation
microstructure
numerical modelling
CT02.30
Smart Sustainable Infrastructure
CT02.31.31
freeze-thaw resistance of BSFC concrete
Delft Cluster

http://resolver.tudelft.nl/uuid:c2511bac-c222-4cb2-9dc3-bed247c73d1b

90-9020622-1

Institutional Repository

doctoral thesis

(c) 2006 O. Copuroglu