F.P. Glasser

Thermal stability and decomposition mechanisms of ettringite at <120°C

Ettringite has been synthesised in phase-pure form and its decomposition and reformation studied as a function of water vapour pressure and temperature. Isobars at PH2O=6, 35, 75 and 400 mm were determined and combined with data on decomposition at the self-generated water vapour pressure achieved at ∼120°C. These data follow a well-defined curve although considerable hysteresis occurs in the temperatures for reformation, which are typically 15–20°C less than the corresponding decomposition temperature at each isobar. The dehydrated product is amorphous or nearly so to X-ray powder diffraction, but still gives an electron diffraction pattern: it has a water content that is fixed or variable within narrow limits, 10–13 H2O per ettringite formula unit. This product is termed meta-ettringite. Recommendations are also made for the upper safe service limits for ettringite-based cements.

Immobilization of chromium in cement matrices

Portland cement and blended cements containing blast furnace slag afford both physical and chemical immobilization of chromium. To separate physical and chemical effects, the pore fluid contained in set, hydrated cements has been expressed and analyzed. In Portland cement spiked with 5,000 ppm Cr(III), pore fluid levels are 0.1--1 ppm, whereas in well-cured slag blends, they decrease to <0.01 ppm. Both cement types give chemical immobilization, but slag cements give the better performance. Slag-containing cements are the most effective at removing Cr(VI) from the pore fluid, probably by reducing Cr(VI) to Cr(III). Electron microscopy coupled with energy dispersive X-ray analysis shows that Cr(III) can be substituted for Al in most of the calcium aluminated hydrate phases. In synthetic preparations, substitution is complete resulting in Ca-Cr phases that are isostructural to calcium aluminate phases. Three Cr analogues of calcium aluminates were synthesized: Ca[sub 2]Cr(OH)[sub 7] [center dot] 3H[sub 2]O, Ca[sub 2]Cr[sub 2]O[sub 5] [center dot] 6H[sub 2]O and Ca[sub 2]Cr[sub 2]O[sub 5] [center dot] 8H[sub 2]O, as well as solid solutions, e.g., Cr substituted hydrogarnet 3CaO [center dot] (Al[sub 2]O[sub 3]/Cr[sub 2]O[sub 3]) [center dot] 6H[sub 2]O. There is no real evidence that Cr is taken up by C-S-H gel.

A thermodynamic model for blended cements

Abstract A model is described for predicting the solid and solution chemistry of blended cements for use in equilibrium modelling. A computer program, CEMCHEM, has been written to predict the stable phase assemblage from the composition of the initial cement blend. Solubility models developed for the cement hydrate phases are then used to predict the aqueous solution composition at equilibrium with the cement. Example calculations for mature cement pastes are presented and compared with real pore fluids extracted from aged cements.

A thermodynamic model for blended cements. II: Cement hydrate phases; thermodynamic values and modelling studies

Abstract Blended Portland cements are likely to form a substantial proportion of repository materials for the disposal of radioactive waste in the UK. A thermodynamic model has been developed therefore in order to predict the composition of the solid and aqueous phases in blended cements as a function of the bulk cement composition. The model is based on simplifying cement to the system CaO SiO 2 Al 2 O 3 SO 4 MgO H 2 O, which constitutes 95% of most cement formulations. Solubility data for hydrogarnet and ettringite suggest that they dissolve congruently and that conventional solubility products can be used to model their dissolution. A solubility model for the siliceous hydrogarnet series, based on ideal solid solution on either side of an immiscibility gap, closely matches experimental solubility data. Solubility data for hydrotalcite and gehlenite hydrate are less consistent and indicative of more complex dissolution processes. On the basis of earlier work, an accurate solubility model is described for hydrated calcium silicate gels in the CaO SiO 2 H 2 O system. Together, these solubility models form a relatively complete thermodynamic model for blended cements. Model predictions for fully matured cement blends are compared to the compositions of pore fluids extracted from aged cement blends. Departures from expected behaviour occur in alkali-bearing systems and are discussed.

Chemical environment in cement matrices

General progress in modelling the behavior of cement matrices and their interactions with wastes is reviewed in terms of appropriate modelling concepts and their formulation in terms of parameters which are susceptible to evaluation. The redox behavior, particularly of slag cement blends is shown to differ sharply from that of Portland cement matrices. Experimental methods of measuring E/sub h/ and poising capacities are described and the composition of pore fluids determined. Slag cements contain appreciable soluble S/sup n-/ species as well as S/sub 2/O/sub 3//sup 2 -/ and SO/sub 4//sup 2 -/.

Encapsulation of Radioiodine in Cementitious Waste Forms

Risk assessment models applied to radioactive waste repository design disclose that iodine is one of the nuclides causing most concern. Computer calculations for these scenario studies assume that iodine, in the form of I − , is poorly sorbed on most geological materials. Therefore it is important that iodine be retained at source, ie. within the vault, for as long as is practicable. In the UK context, cements are likely to form a major part of the waste package for low and medium active wastes, and of engineered vault structures. These cements are likely to be blends of one form or another, including Portland cement blended with blast furnace slag, or fly ash. It is therefore important to assess the effect of Portland cement and blending agent on iodine speciation and uptake by the constituent solid phases. Data are presented on the uptake of I − on specific phases: Ca(OH) 2 , calcium aluminate sulphate hydrates, hydrotalcite and calcium silicate hydrogel (C-S-H). Combining this with a model for predicting phase assemblages in well-aged slag cements, yields an optimum blend for immobilising radioiodide, on ageing. Precipitation of I as AgI during cementation and radiolytic effects on I are also discussed. The production of gaseous radioiodine (I 2 ) is potentially a serious problem.

The stability of ettringite

Ettringite has been known as a natural mineral for more than 100 years and is a normal constituent of hydrated Portland cement pastes. Its nominal composition in cement pastes, Ca6[Al2(OH)6]2(SO4)3·24-26H2O, may be modified by single or double substitution, of carbonate for sulfate and of silicon for aluminium. The stability of ettringite in the CaO-Al2O3-SO3-H2O system is relatively well established; it decomposes at 114 C. However, ‘stability’ is frequently used as shorthand for ‘compatibility with other phases’ and indeed, it is also appropriate to include within the scope of this assessment ettringite stability at water vapour pressures less than saturation. The relevant data are reviewed. Some aspects of ettringite persistence (as distinct from stability) are discussed, for example formation of ‘metaettringite’ a phase which preserves the columnar structure of ettringite but has lower water content than ettringite itself. A distinction is made between ettringite stability and amount of substance: within the stability range of ettringite, the presence of additional soluble components, e.g., salts, can alter the amounts of ettringite present but this is not conclusive proof that ettringite is intrinsically unstable.

The Long-Term Properties of Cement and Concretes

The state of knowledge concerning the internal constitution of cement systems is reviewed. It is concluded that many properties relevant to radwaste immobilization are strongly influenced by the presence of thermodynamically metastable phases. In order to ascertain cement properties at long ages, models have been developed which include compositional and time-dependant factors. These require a validated data base. Progress in developing models and creating the data base are reviewed. In the course of accumulating data, specific radwaste species -cement interactions can also be quantified.