The concurrent goals of cement hydration are to percolate (bridge) the original cement particles into a load-bearing network and to depercolate (dam) the original water-filled capillary porosity. The initial volume, particle size distribution, and flocculation/dispersion state of the cement particles have a large influence on both hydration rates and microstructure development. Likewise, the capillary porosity as characterized by its pore size distribution, percolation state, and saturation state also influences both hydration kinetics and microstructure. In this paper, experimental techniques and computer modeling are applied to further understanding several of the critical connections between these physical parameters and performance properties. First, the setting or bridging process is explored via a combination of needle penetration and rheological measurements, in concert with three-dimensional microstructural modeling. Second, low temperature calorimetry is shown to be a valuable indicator of the percolation state or damming of the water-filled pores with various size entryways in the three-dimensional microstructure. Porosity percolation (or depercolation) is shown to be strongly influenced by both curing conditions and the alkali content of the cement pastes. Finally, it is proposed that future efforts in this field be directed towards a greater understanding of the (nano)structures of cement hydration products, particularly the calcium silicate hydrate gel, and their influence on performance properties.
Air void spacing equations have been proposed in the literature by a number of authors: Powers; Philleo; Attiogbe; and Pleau and Pigeon. Each proposed spacing equation attempts to characterize the true “spacing” of entrained air voids in concrete. While efforts have been made to correlate these spacing equation calculations to freeze-thaw performance, no test has been performed to assess the geometrical accuracy of these spacing equations. Herein is a computerized accuracy test of these proposed spacing equations. A computer model of air void systems is used, and various “spacings” are measured in the model system. The results of these measurements are then compared to the appropriate spacing equation prediction, along with equations developed by Lu and Torquato.
X-ray microanalyses have been made on a suite of nine oilwell cement clinkers. Elemental data were obtained on alite, belite, and ferrite phases and, for two clinkers only, also on aluminate. The alite and belite compositions are broadly similar to those previously reported for other portland cements. Guest ion concentrations suggest that several charge-balancing substitutions occur. Bulk MgO and SO3 levels determine the Mg and S contents of both phases, but for Fe and Al the substitution levels are not strongly correlated with bulk composition. For ferrites, the data are generally consistent with the results of Bergstrom et al. (Adv Cem Res 1992,4, 141–147): the Mg content varies widely, is controlled by the bulk MgO, and is coupled with Si in a charge-balancing substitution for Fe. Estimates of the total phase assembly using the directly determined mineral compositions are compared with the predictions of a modified Bogue calculation (similar to that of Taylor, Adv Cem Res 1989, 2, 73–77). The agreement is variable and some refinements to the method of calculation are indicated.
The ionic diffusivity of a concrete is a function of its microstructure at many length scales, ranging from nanometers to millimeters. The microstructure is largely controlled by the initial concrete mixture proportions and the ultimate curing conditions. Linking a property like ionic diffusivity to the microstructure then requires a multiscale approach. A multiscale microstructural computer model for ionic diffusivity has been previously developed. This model has been developed specifically to compute the chloride diffusivity of concretes with various mixture proportions and projected degrees of hydration. The three key parts of this model were dependent on large-scale supercomputer-magnitude simulations to: (1) determine the total volume of interfacial zones for a given aggregate distribution; (2) simulate the hydrate cement paste microstructure around a typical aggregate; and (3) compute the effect of the aggregates and interfacial zones on the overall diffusivity of the concrete. The key feature of this model is that one can approximately take into account the redistribution of cement paste between interfacial transition zone regions and bulk paste regions, and its important effect on overall concrete diffusivity. In the present article, we review the previously developed model and show how analytical equations can accurately replace the large scale computer simulations of parts (1) and (3). This accomplishment will make the model more usable by those who do not have access to supercomputer computing power. Published by Elsevier Science Ltd.
The durability of type K and H cement based grouts under conditions potentially found to be in high level nuclear waste repositories was studied. Tests have been carried out to determine the effects of temperature on the hydraulic conductivity and the leaching resistance of the grouts. Measurements of mercury intrusion porosimetry and scanning electron microscopy with energy dispersive X-ray analysis have been used to investigate the changes in the pore structures of both grouts as function of leaching and permeating time. Type K and type II cement based grouts, made with low water to cementitious materials ratio, silica fume; and superplasticizer, were exposed to high temperature and pressure. Preliminary results indicate that both hydraulic properties and leaching behavior of the grouts investigated are affected by the increase in temperature. However, data shaw that leaching rate mid hydraulic conductivity of the grouts decrease with time. The results showed clearly that chemical reactions, presumably accelerated by the elevated temperatures (100 degrees C), led to the formation of a precipitate in the microcracks and on the surface of the leached specimens. This precipitate is likely the cause of the observed decrease in the hydraulic conductivity and leaching rate. (C) 1998 Elsevier Science Ltd.
The effect of the geometry of woven fabrics on the bond between monofilament polyethylene yarns and cement matrix was studied in the present work. The fabrics were all plain weave, with varied fills density: 5, 7, or 10 fills per cm; the warps’ density was kept constant at 22 warps per cm. The interfacial bond was evaluated by pullout tests. To characterize the influence of the fabric’s geometry on bond performance, the influence of different parameters of the fabric’s geometry that may affect bond were separated: (1) pullout of a single crimped yarn untied from the fabric to characterize the influence of the shape of the individual crimped yarn; (2) pullout of a single yarn from free fabric (not embedded in the cement matrix); and (3) pullout of a yarn from a fabric embedded in the cement matrix. Straight yarns were also tested for comparison. It was found that the woven fabric provided a considerably better bond to the cementitious matrix than the bond of a single straight yarn. The crimped geometry of the yarn in the fabric was found to have a significant influence on increasing the bond between the woven fabric and the cementitious matrix.
The paper industry in Western Europe generates around 6 million tons/year of sludges, which contain about 60% dry matter mainly composed of cellulose fibers, kaolinite, and calcite. The present study deals with an original way of utilizing such wastes: the production of metakaolin by calcining paper sludge in the temperature range of 700°C to 800°C. After calcination, pastes containing 50% calcium hydroxide and 50% burnt sludge were hydrated and the lime consumption investigated by differential thermal analysis. The results show that a very reactive pozzolan is produced by calcining paper sludge at 700°C or 750°C for 2 or 5 hours. Despite a smaller kaolinite content, the burnt paper sludge exhibits more pozzolanic activity than commercially available metakaolins, especially at early ages. Thermodesorption analyses show that this higher activity is due to the presence of superficial defects that occur during the sludge calcination.
Deicing chemicals are mired with corrosion inhibitors to reduce rebar corrosion in bridge decks. The corrosion inhibitors are thought to penetrate into concrete to the depth of the rebars and form the passive film on the rebar surface. In a previous study, it was found that corrosion-inhibitor-added deicing salts interacted with 3% NaCl-added concrete and produced precipitates through chemical reactions. The amounts of precipitates produced was dependent on the type and concentration of corrosion-inhibitor-added deicing salts. In this investigation, the precipitates formed by chemical reactions between concrete and corrosion-inhibitor-added deicing salts were identified by using chemical analysis and X-ray diffraction methods. The distributions of Cl-, SO42-, and PO43- in concrete slabs ponded with corrosion-inhibitor-added deicing salts were determined by chemical analyses of powder samples obtained from the slabs. The major precipitates were calcium and/or magnesium phosphates as major chemical compounds and gypsum as a minor component. High concentrations of phosphate were observed at the top portion of concrete; slabs when the deicing chemicals contained phosphate inhibitors. Voids were observed at the interface of aggregate and mortar in the concrete slabs tested with the deicing salts solutions containing corrosion inhibitors. (C) 2998 Elsevier Science Ltd.
27Al MAS NMR spectra of synthetic calcium aluminoferrites, Ca2AlxFe2−xO5 with x = 0.93, 1, 1.33, reveal only a few percent of the expected intensity for the 27Al central transition, indicating that the calcium aluminoferrite phase in Portland cements can barely be observed by 27Al MAS NMR. This result supports the use of 27Al MAS NMR for quantitative analysis of the tricalcium aluminate phase in Portland cements.
Four different series of mortars were cast to study the influence of the microstructure on the transport coefficient of tritiated water. Test parameters included type of binder, water/binder ratio (0.25 and 0.45), and sand volume fraction (0% and 50%). Mercury intrusion porosimetry was used to characterize the pore structure of all mixtures. The effective tritiated water diffusion coefficients were determined using a simple diffusion test. The test results show that the reduction of the water/binder ratio and the use of silica fume significantly contribute to reduce the transport properties. The test results also indicate that aggregates modify both the microstructure and the transport properties of mortars. The diffusion coefficient of tritiated water was found to decrease with an increase of the sand volume fraction. The increased tortuosity of the matrix induced by the presence of aggregates thus appears to be more important than the influence of the interfacial transition zone. Results also clearly underline the influence of the preparation technique prior to a mercury intrusion experiment.
Self stress generated in polymer impregnated gypsum (referred as GPC) when it is composed is estimated, and its influence on flexural strength is discussed. The estimation of the self stress is based on measured values of shrinkage caused by polymerization of impregnated monomer and elastic modulus of dried gypsum base just before impregnation. The effect of this self stress on flexural strength of GPC is investigated. It was found that the following equation is valid to predict flexural strength of GPC (бb) in terms of the self stress as a variable: бb = бgb + Vp (бp − бsp), where бgb = flexural strength of gypsum base, Vp = specific volume of polymer, бp = tensile strength of polymer, and бsp = self stress generated in polymer phase. If extremely low water-gypsum ratio is adopted to prepare gypsum base, cracking is observed just after polymerization preceding any flexural loading. For somewhat higher water-gypsum ratio, specimens are not cracked, but their flexural strength is decreased after polymer impregnation. The self stress corresponding to this case turns out to be higher than the tensile strength of polymethyl methacryrate used for the impregnation. Since prediction of бsp in the equation is based on tri-axial compressive strain of gypsum base that is within its elastic region, бsp in polymer phase should positively exist. Even for this condition, the validity of the equation seems to be maintained, although the value in the parenthesis of the equation becomes negative. Based on this fact, an unstable physical state where one phase of a composite material is stressed beyond its macroscopic strength as an individual material owing to the crack arresting effect of the other phase (gypsum in this case) has been postulated. This state is designated as a “superstressed” state, taking its resemblance with supercooling or supersaturation into consideration.
This study presents a method for estimating the strength of lightweight aggregate. Cylindrical specimens with various aggregate volume ratios (volume of coarse aggregate/total aggregate volume) were cast and tested. Micromechanics method was applied by considering a perfect bond between mortar and aggregate. The approximate aggregate strengths determined from the concrete strength, component properties, and the volume ratio of aggregate are between 15 and 30 MPa. Both matrix strength and composite strength are much higher than the lightweight aggregate strength.
This article outlines an experimental and numerical study on quasi-instantaneous and long-term deformations of high performance concrete subjected to sealed curing. For this purpose more than 100 cylinders and 400 cubes were made of eight concretes and studied in relation to creep and shrinkage, hydration, internal relative humidity, and compressive strength. One heat-cured concrete was studied at temperatures other than 20°C varying between −20°C and 60°C. Analyses were carried out of quasi-instantaneous deformation, short- and long-term basic creep, and autogenous shrinkage. Relationships were obtained between elastic modulus and creep compliance, and hydration, internal relative humidity, and compressive strength. New and original results are presented on relationships between autogenous shrinkage and internal relative humidity. Other results confirm and validate earlier findings of normal strength concrete regarding relationships between creep compliance, porosity, compressive strength, and maturity for high performance concrete. The project was carried out at Lund Institute of Technology between 1992 and 1996.
Relationships among mixing conditions, extensional viscosities of the formed paste, and the morphology and flexural strength of hardened composites have been studied for a calcium aluminate cement-phenol resin composite. Mixing torque was monitored as a function of temperature and resin pH. The behavior of the torque time curves as well as C-13 nuclear magnetic resonance and differential scanning calorimetry measurements suggest that paste formation is characterized by an induction period of flat torque, corresponding to dissolution of cement ions,followed by a stiffening: period of rapid torque rise, corresponding to an ionic interaction between the resin and cement particles. The viscoelastic nature of the paste was studied using a biaxial squeeze flow device. The steady biaxial extensional viscosity trends with extension rate are consistent with the formation of an increasingly cross-linked bulk organic phase in the paste. Flexural strengths of hardened material processed tinder varying degrees of resin pH suggest that an optimal structure forms when the resin is allowed to simultaneously polymerize and ionically interact with cement particles. This conclusion is supported by evidence of scanning electron microscopy, which shows structure formation for a given set of mixing conditions. (C) 1998 Elsevier Science Ltd.
An experimental and numerical study of the long-term interaction between silica fume and Portland cement in concrete subjected to air, water, or sealed curing is outlined. About 250 kg of eight qualities of each concrete were studied at four different ages each over a period of 7 years between 1989 and 1996. Parallel studies of strength, hydration, and internal relative humidity were performed. Half of the concretes contained silica fume. New and original results and analyses of the interaction between Portland cement and silica fume related to compressive strength, split tensile strength, hydration, and internal relative humidity are presented. The specimens are available for future measurements.
The effects on the microstructural development of adding silica fume to cements and concretes during cement hydration have been studied using small-angle neutron scattering and ultrasmall-angle X-ray scattering. A previously developed fractal based microstructural model has been applied to extract representative microstructural parameters from the small-angle scattering data. A link has been established between the existence of coarse or agglomerated particles in the silica fume particle size distribution and possible deleterious microstructural evolution during cement hydration.Published by Elsevier Science Ltd.
Concrete is a composite, and its properties depend on the properties of the component phases and the interaction between them. It is known that the interfaces are the weakest link in concrete, playing a very important role in the process of failure. This process is strongly related with the characteristics of the aggregates (especially coarse aggregates) and with the relative differences in strength between matrix and inclusions. This paper analyzes the mechanical behavior of high strength and conventional concretes prepared with coarse aggregates having significant differences in strength, shape and surface texture, porosity and absorption, and interface bond strength. Two different gravels and two different crushed stones were used. Concrete mixtures with water/cement ratios of 0.30 and 0.50 were designed. The effects of aggregate type and strength level on concrete failure mechanism, including tensile and compressive strength, stiffness, energy of fracture, and crack pattern, are discussed.
Small amounts of polyvinyl alcohol (PVA) were added to cement paste in an attempt to increase the aggregate-paste bond strength. Specimens consisted of ordinary Portland cement mixed with PVA/water solutions and cast against ground surfaces of limestone and granite. The aggregate-paste bond strength after curing was tested in a wet state by three-point bending. The morphology of the interfacial transition zone was observed with scanning electron microscopy, and the composition was analyzed with infrared spectroscopy. The addition of 1.4 wt% PVA based on the mass of cement increased the strength for both limestone-paste and granite-paste bonds. The strength increase was about five-fold for limestone and nearly two-fold for granite. The failure mode also changed, from pure adhesive failure without PVA to cohesive failure of the aggregate with limestone and to a mixed cohesive failure of the paste and adhesive failure with granite. The gain in bond strength with the addition of PVA seems to arise from suppression of the porous interfacial transition zone and an inhibition of calcium hydroxide nucleation on the aggregate surface.
The exchange of water by ethanol in two water-saturated cement pastes has been investigated by carbon NMR. The two cement pastes differed only in their thermal history. The diffusion of ethanol into the cement paste was shown to be described by Fickian diffusion, assuming one-dimensional diffusion under perfect sink boundary conditions. The diffusion coefficients were calculated to be (1.28 +/- 0.14) 10(-7) cm2/s for the virgin cement sample and (4.38 +/- 0.57) 10(-7) m2/s for the preheated cement sample (preheated at 105 degrees C for 12 h), respectively. The measurements indicate an extensive exchange between water and ethanol.
In a cement hydration product-electrolyte system the hydration products are micron-sized charged particles. The principles of colloid electrochemistry predict that in such a system each solid particle is surrounded by a layer of concentrated solution of its counter-ion. The actual concentration of this layer depends on the surface charge and the concentration of the solution away from the particle, i.e., bulk solution. Divalent counter-ions preferentially concentrate around the solid particles. The principles also predict that the intrinsic diffusivity of an ion is proportional to the square root of its concentration in a solution. The literature on the expressed pore solution, ionic diffusivity through cement based materials, the formation factor, alkali-silica reaction, etc., has been examined from the point of view of colloid electrochemistry. Many of the reported but unexplained phenomena could be explained from the principles of colloid electrochemistry.
