| dc.description.abstract | The rheological behaviour of cementitious pastes has been studied by various means. Six different cements have been studied in main parts of the work and all of them have been characterized according to the Rietveld method in order to determine the exact content of minerals. Easily soluble alkalis were measured by plasma-emission- spectroscopy of the fluid filtered from paste. Three types of plasticizers namely naphthalene sulfonate formaldehyde condensate (SNF), lignosulphonate and polyacrylate grafted with polyether (PA) have been used throughout the work. The influence of the plasticizer type on the rheological properties of the cementitious pastes, their adsorption characteristics and their effects on heat of hydration of the pastes has been studied. Limestone has been used as a nonreactive model material for cement in some parts of the work.
All rheological measurements were performed with a parallel plate rheometer. Rather than describing the shear stress-shear rate flow curve with the usual Bingham model resulting in plastic viscosity and yield stress, the area under the curve (Pa/s) was used as a measure of “flow resistance”.
The effect of silica fume and limestone on the rheology of cementitious pastes
The rheological behaviour of cementitious pastes, with the cement being increasingly replaced by densified and untreated silica fume (SF) or limestone was studied. Three plasticizers were investigated namely two types of polyacrylate (PA1 and PA2) and SNF. PA2 proved to be the most efficient plasticizer of the three while PA1 and SNF provided comparable results.
The flow resistance was found to increase with increasing silica fume replacement when SNF and polyacryalte (PA1) were added as plasticizers which was explained by ionization of the silica fume surface and possible bridging with polyvalent cations like calcium. The flow resistance decreased, however, with increasing silica fume replacement when the second and more efficient type of polyacrylate (PA2) was utilized which was believed to occur since the cement pastes were better dispersed by PA2 than SNF and PA1. The silica fume particles could thus pack between the cement grains and displace water. An alternative explanation for reduced flow resistance with increasing silica fume replacement could be a ball-bearing effect of silica spheres.
There was found a trend of increasing gel strength with increasing silica fume replacement of cement even though the pastes seemed to be dispersed by PA2. Cement pastes with densified SF developed lower gel strengths than pastes with untreated SF. This phenomenon was attributed to more grain shaped agglomerates with lower outer surface in densified SF compared to dendritic agglomerated in untreated SF. Decreasing gel strength was found for pastes with increasing limestone filler replacement. Thus silica fume may be advantageous as stabilizing agent for self-compacting concrete preventing segregation upon standing due to a more rapid gel formation.
Effect of cement characteristics on flow resistance
Rheological experiments were performed on pastes prepared from 4 cements originating from the same clinker, but ground to different finenesses (Blaine). The results showed that the flow resistance increased exponentially with increasing Blaine number. No correlations between single cement characteristics such as Blaine, content of C3A, cubic C3A (cC3A) and C3S with the flow resistance were however found when cements from different clinkers were used. This finding indicates that cement should not be treated as a univariable material. However, the combined cement characteristic (Blaine•{d•cC3A+[1-d]•C3S}) was found to correlate with flow resistance, where the factor d represents relative reactivity of C3A and C3S. The flow resistance was found to be either a linear or exponential function of the combined cement characteristic depending on plasticizer type and dosage. Correlations were found for a mix of pure cement and cement with fly ash, limestone filler (4%), as well as pastes with constant silica fume dosage when the minerals were determined by XRD.
Influence of cement and plasticizer type on the heat of hydration
The initial heat of hydration peak was measured for the 6 main cements with 0.32% SNF, lignosulphonate and PA2 by cement weight. Correlations were attempted between the maximum heat of hydration rates of the initial peaks with various cement characteristics. The maximum heat of hydration rate seemed to correlate with the product of the cement fineness and C3A content regardless of plasticizer type. The fly ash cement had to be left out of the correlation plots due to its low initial heat of hydration.
The second, third and fourth hydration peaks were measured on the cement pastes with 0-0.8% SNF, lignosulphonate and PA2 by weight of cement. Lignosulphonate was found to be the strongest retarder while SNF had the least effect on the setting time of the three plasticizers. No correlations could be found between the setting times and cement characteristics such as cement fineness, aluminate and alkali contents for un-plasticized pastes probably because the setting times might have been too close to each other to be able to obtain accurate values. Correlations between setting time and cement characteristics were however found for pastes with plasticizers. The setting times did not correlate with the cement fineness (Blaine) as a single parameter. The product of cement Blaine and C3A content, however, resulted in a correlation. Furthermore the setting time correlated with the cubic modification of C3A. It may seem that the setting times depend more on the cubic modification of C3A than the sum of orthorhombic and cubic aluminate. This finding indicates that the cubic aluminate modification is more reactive than the orthorhombic. The setting time decreased with increasing content of easily soluble K-ions in the cements probably due to the formation of syngenite, K2SO4·CaSO4·H2O, which removes some sulphate from solution that would otherwise retard C3A hydration. A similar correlation was not found between the setting time and the sodium equivalent.
Cement interactions with plasticizers
Three plasticizers were studied namely SNF, lignosulphonate and polyacrylate (PA2). PA2 was the most efficient plasticizer of the three tested even thought it was found to adsorb to a lesser extent on cement than SNF and lignosulphonate. SNF and lignosulphonate brought about comparable results.
PA2 was observed to induce flow gain within the 2 hours of rheological measurements which might be caused by the polymer expanding in the water phase and thus improve the dispersion of the paste. Furthermore the grafted side chains of the polymer are considered to be long enough to provide steric dispersion even thought the backbone might be embedded in the hydration products. Cement pastes with SNF and lignosulphonate exhibited flow loss as a function of time which indicates that the plasticizer molecules were consumed by the hydration products.
The concentrations of superplasticizer in the pore water were not found to change markedly in the time range 20-95 min after water addition, indicating that most of the plasticizer molecules were consumed (i.e. adsorbed or intercalated in surface hydration products) within the first 20 minutes after water addition.
The adsorption characteristics were found to depend on the plasticizer type. The adsorption curves of SNF and lignosulphonate reached a plateau at saturation characterizing high-affinity adsorption or increased continuously as a sign of low affinity adsorption. The adsorbed amounts of polyacrylate decreased, however, after saturation had been reached which might indicate that surplus molecules in the water phase compress the ionic double layer or that adsorbed molecules expand and hinder molecules in the water phase to attach at the surface (i.e. osmosis).
The plasticizer saturation dosages were found to depend on cement surface area (Blaine), amount of cubic C3A and easily soluble sulphates. The saturation dosage of lignosulphonate seemed to have a dependency on the amount of soluble alkali that was somewhat stronger than observed for pastes with SNF. This difference might be caused by lignosulphonate forming complexes with solvated ions in a higher degree than SNF. Moreover alkali sulphates are furthermore often added to commercial SNF based products as the one used in this work. The best correlation, overall, was found for the product of cubic C3A and Blaine which is logical since high surface and cubic aluminate contents accounts for high cement reactivity and since the plasticizers are known to coordinate with calcium sites. Correlations were also found between saturation dosage with the product of Naeqv and Blaine as well as the product of Naeqv and cubic C3A. The investigations seemed to indicate that the plasticizer saturation concentration increase with increasing alkali content. These findings, however, are rather unclear. According to literature an increased concentration of alkali sulphate in solution results in both an increased hydration rate (which would lead to a higher plasticizer intercalation) and a reduced plasticizer adsorption (due to SO42- - superplasticizer competition). The easily soluble sulphates might, of course, entail the opposing effects of Blaine and C3A in a way that smoothen the correlation plots of the plasticizer saturation dosage with the cement characteristics.
Effect of temperature on rheology and plasticizer adsorption
Flow resistance and adsorbed amounts of SNF, lignosulphonate and PA2 were measured at temperatures ranging from 11 to 40oC. Limestone was used as a nonreactive model material for cement. The adsorbed amounts of SNF and lignosulphonate on limestone were found to decrease after reaching a maximum which occurred at approximately 25oC. Decreased amounts of adsorbed plasticizer with increasing temperature might be explained by increased kinetic energy to the molecules or by an entropy effect. The adsorption of PA2 on limestone seemed to be independent of paste temperature in the range of 16-34oC which might be caused by low reduction of entropy at adsorption due to its short backbone and long, grafted side chains. The flow resistance of the limestone pastes generally increased with increasing temperature which may be caused by reduced amounts of adsorbed plasticizer and/or dehydration of the paste during the rheological measurements.
Two types of cements were used to study adsorption and flow resistance with increasing temperature namely CEM I 42.5 RR and CEM I 52.5 R-LA. Amounts of plasticizer adsorbed and intercalated (consumed) by cement reached a plateau or even decreased with increasing temperature in the case of SNF and lignosulphonate. This finding might be caused by two opposing effects namely: increased number of adsorption sites due to increased hydration rate with increasing temperature and reduced adsorption due to increased kinetic energy and/or reduced entropy of the plasticizer. Amounts of PA2 consumed by cement increased linearly with increasing temperature as might be explained by the experiments with limestone where the adsorbed amounts of PA2 seemed to be independent of temperature. Increased consumption of plasticizer by the cements with rising temperature is thus probably governed by the increased number of adsorption cites due to increased hydration rate. The flow resistance of CEM I 52.5 R-LA cement increased exponentially with increasing temperature as a function of temperature most likely because of the increased hydration rate. The pastes of CEM I 42.5 RR cement were generally highly viscous and probably agglomerated. The flow resistance reached a plateau value with increasing temperature in this case. | nb_NO |
