An overview of the different problems to be solved for waste form disposal in a shallow-land burial site is presented. Examples of cementbased materials used to immobilize concentrates are given in relation to cement chemistry. Experimental studies, undertaken to assess concrete durability, obviously indicate that pH is the starting parameter of chemical degradation reactions. It is found that blended cement (CLC) has a better cesium retention ability and a stronger resistance to chloride and sulfate attack than Portland cement.

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 (I2) is potentially a serious problem.

Current plans for disposing various low-level radioactive and/or hazardous liquid wastes include solidification of the waste using cementitious materials. One process, known as grouting, involves mixing liquid wastes with a blend of cementitious materials and pumping the resultant slurry to lined, underground concrete vaults. As the grout slurry begins to solidify and harden, the temperature within the grout increases due to exothermic hydration reactions. Depending on the the particular grout composition and on the disposal conditions, the grout may be exposed to temperatures of around 90°C for extended time periods. Studies are being conducted to determine the effects of high-temperature curing on selected properties of grouts prepared with a simulated low-level liquid waste. Grout samples cured at temperatures up to 950C in the laboratory absorbed water during curing. The resultant leach resistance and compressive strength of these grouts decreased with increases in curing temperature and curing time.

Ground granulated blast-furnace slag (ggbfs)/OPC blends are among the materials being considered as possible encapsulating matrices for low and intermediate level radioactive wastes. It is therefore desirable to identify any microstructural modifications which may result in these systems frmn exposure to ionizing radiations. In order to investigate this, ggbfs/COC blends were γ-irradiated to >80M3y at 50°C over a period of 2 years. These were examined by electron optical, X-ray and thermal analysis techniques and the results ccmpared with controls subjected to a similar hydration regime. The morphological and ccmpositional similarities and differences revealed between the irradiated and control samples are presented and discussed.

The foundation necessary for a constitutive model to simulate the response of concrete at elevated temperature has been presented. A model including temperature dependent creep, material property degradation, and cracking is needed to evaluate the long term functional requirements of concrete as a barrier material for nuclear waste disposal. Since the stress due to thermal load is proportional to the modulus, the degradation of the modulus with time even at constant elevated temperatures requires continual redistribution of load. Furthermore, since this degradation is not recoverable, the response of the material at elevated temperatures exhibits a complex dependence not only on the temperature distribution, but on the prior thermal history of the structure.

This constitutive model for the response of concrete at elevated temperatures has been implemented into an implicit, finite element program called ANACAP [8]. Because of the direct coupling with temperature, both through thermal loads and material property dependency, ANACAP also contains a heat transfer module that includes thermal effects due to fluid flow and moving material boundaries. This program has been applied to problems involving underground waste storage tanks and grout vaults at the DOE Hanford site in Richland, Washington.

The Structures Laboratory of the U.S. Army Engineer Waterways Experiment Station has conducted research on cement-based composites for the Waste Isolation Pilot Project (WIPP) since 1977, in cooperation with Sandia National Laboratories. Field testing requirements guided initial development of grouts. Concurrent and later laboratory studies explored the chemical stability and probable durability of these mixtures. Beginning in 1985, a series of small-scale seal performance tests at the WIPP prompted development of an expansive salt-saturated concrete. Important lessons learned from this ongoing work include: (1) carefully tailored mixtures can tolerate phase changes involving Ca, Al, and S04, without loss of structural integrity; (2) handling and placement properties are probably more crucial to the mixtures than is exact phase composition; and (3) for the environment of a geologic repository, demonstrated chemical durability will be the best indicator of long-term performance.

Cementitious waste forms are being considered for immobilizing nuclear waste before disposal. In earlier work, it was found that irradiation of a mortar formulation consisting of slag, portland cement, fly ash, water, and up to 10 wt % KCI-LiCl salt resulted in the generation of hydrogen. Yields were relatively high and the rates of generation were constant for the irradiation period investigated. The addition of small amounts of oxygenrich electron scavengers (NaN03, NaI04, KMn04, or Ag20) to the mortar was investigated as a means for reducing hydrogen yields. The addition of NaN03 reduced the hydrogen yield; changed the radiolytic products from hydrogen to a mixture of hydrogen, nitrogen, and N20; and reduced the pressurization rate after exposure to 400 Mrads. The addition of NaI04 and KMn04 reduced hydrogen yields slightly while the addition of Ag20 increased the yield. Moreover, the addition of FeS to a non-slag mortar changed the radiolysis mechanism but the addition of FeO did not. The results of these experiments provided an insight into the nature of the radiolytic reactions occurring in the mortar formulations and indicated that the radiolytic generation of gases might be controlled with the proper choice of additive.

Ground granulated blast-furnace slag (ggbfs) /Ordinary Portland cement (OPC) blends are possible materials for use in intermediate and low-level radioactive waste repositories. The microstructural development in neat OPC is described. The effect of increasing the loading of ggbfs on the composition and microstructure of the hardened paste has been examined by a number of techniques, including transmission electron microscopy. The implications for performance are discussed. A ggbfs/OPC 9:1 blend which had been exposed, after normal hydration to aqueous leaching was also examined. Marked changes in the microstructure and composition were observed.

The leaching mechanisms of simulated low-level radioactive waste forms are being determined as support for development of an accelerated leach test. Two approaches are being used: (1) comparison of leaching data with results of a model that describes diffusion from a finite cylinder, and (2) observation of the leaching process at temperatures between 20°C and 65°C. To provide results that can be used for modeling, leaching at elevated temperatures must change neither the leaching mechanism nor the structural controls of leaching such as the porosity. Releases of 137Cs, 85Sr, calcium, sodium and potassium from portland cement containing sodium sulfate, as a simulated evaporator sludge, have been determined under a variety of experimental conditions. Data from the leach tests were compared to model results for diffusion from the finite cylinder. While most leaching appears to be diffusion controlled, notable exceptions occur. For all samples, activation energies ranging between 6 and 11 Kcal/mole have been calculated from the relationship of the effective diffusion coefficient to increasing temperature, close to the expected value of 5 Kcal/mole for diffusion.

The chemical behavior of cement in the natural environment is of interest because cement has been suggested as a means of containing low-level radioactive waste and as a barrier between radioactive-waste containers and the surrounding soil or rock. In the past, either experiments or computer modeling have been used independently to predict how cement will leach in the natural environment. In this study, a geochemical model for cement leaching was developed, and predictions by the model were compared to results from a static leach test. Calcium concentrations and pH were adequately predicted by the model; however, other major leachate constituents, such as Al, Si, and S04, were poorly predicted. Additional experimental data and refinement of the model are both required to better predict the observed leachate concentrations. Accurate prediction of experimental results by a geochemical model would increase confidence in the model for use in long-term performance assessment.

At Hanford, low-level liquid nuclear waste is being mixed with cementitious materials (grout) to form leach-resistant solid waste. Prior to grouting each liquid waste, an assessment must be performed to evaluate the long-term environmental impact. These predictions rely upon a diffusioncontrolled release model and short-term laboratory leach data on small grout samples. This paper describes size scale-up and inventory scale-up experiments that evaluate whether diffusion does in fact control the release of contaminants. The results of the volume scale-up test suggest that tests on grout cylinders between the sizes 3 cm dia. by 3 cm length and 30 cm dia. by 29 cm length yield comparable results. These data and other available literature suggest that extrapolation of leach results to large blocks of solidified waste is defensible.

The inventory scale-up tests for 125I and selenium show a constant effective diffusivity regardless of the original inventory. This suggests that diffusion processes do in fact control the release of iodide and selenium from grout. The 99Tc effective diffusivities may not remain constant with changes in inventory.

Numerical solution of Fick's second law of diffusion demonstrated that the error introduced in the ANSI/ANS-16.1-1986 procedure, by assuming a zero surface concentration, varied from about 30% for a leachability index of 5 to about 0.1% for an index of 10. Waste forms typically have indexes of 7 or greater, implying errors of less than 5% with this assumption. The estimated leachability index differs from the actual value by less than 0.2 when the numerical solution is analyzed as input experimental data, except for a leachability index as low as 5 (essentially the value for water).

Latex-modified Portland cement, which would be expected to have low permeability and ion diffusivity, was studied for possible application as a solidification agent for radioactive wastes generated from nuclear power plants. In order to predict the leaching ratio of radionuclides from the cementitious waste forms, the effect of water and latex content in the fresh cement paste on total porosity and ion diffusivity of hardened paste was quantitatively estimated.

Total porosity of hardened cement paste decreased with the reduction of water content in the fresh paste and it was also reduced by latex addition. This latter effect could be attributed to the latex emulsion forming a water-proof film and filling the capillary pores. Also Cs ions diffusivity, which is the ratio of the diffusion coefficient in pore water to that in bulk water, showed an exponential correlation with total porosity for both cement materials. An empirical equation, expressing ion diffusivity as a function of total porosity, was derived from the consideration that the water constrictivity in this porous medium could cause an increase of the apparent viscosity of pore water. These results suggested a possibility that the transport behavior of radionuclides through the cementitious matrix could be estimated from the mixing parameters of the original cement pastes.

The problem of predicting the future performance of cement-based systems is complicated by a poor understanding of the behaviour of cement systems at long ages, as well as of the complex interactions which can occur between cement and waste components - active as well as inactive - and with cement blending agents including fly ash, glassy slags and natural pozzolans. The progress achieved in developing a predictive capability is reviewed. Considerable success has been achieved in modelling the chemically-related features of cement based systems in terms of pH, Eh, and of element partition between solid and aqueous phases. The behaviour of model radwaste elements - iodine and uranium - has been studied in depth and indicates that both sorption and precipitation occur. U(VI), in particular, forms solubilitylimiting compounds, e.g. uranophane. But in general, presently-available data are inadequate to predict many cement-waste stream teractions; future progress in modelling is likely to rely heavily on additions to the data base.

The repository environment will also condition chemical exchanges in cement-based systems. Progress is being made in predicting the impact of CO2, a common ground water component, on the performance of cement systems.

A waste repository for the below ground disposal of low level radioactive waste is planned at Chalk River Nuclear Laboratories. It relies greatly on the durability of concrete for the required 500 year service life. A research program to design durable concrete and predict its service life is in progress.

The degradation of the concrete depends to a large extent on the rate of ingress of corrosive agents. Penetration of chloride and sulphate ions are particularly relevant. Twenty mix formulations were developed to create various types and qualities of concrete, and to study their behaviour in different site environmental conditions. A total of 1000 concrete specimens are being exposed at 20°C and 45°C to 25 different combinations of the corrosive agents including CO2. Procedures to measure the ionic profiles and to determine the factors controlling diffusion of the ions in the various concretes have been developed. Results of selected concrete systems exposed to chloride and sulphate solutions for 1 year are presented and discussed in terms of pore structure and permeability parameters of the concrete.

A model to describe the pore sizes in cement paste and mortar, as determined by high pressure mercury intrusion porosimetry, has been developed. The model describes porosity using a compound lognormal distribution. For given material under a given set of curing conditions, the weighing factors and shape parameters of two sub-distributions in the lognormal model may be considered as constants, while the location parameters may be related to curing time and the relationship can be quantified. Therefore, it is possible to predict both the pore size distribution in cement and mortar at any age as well as the evolution in pore size during curing.

The durability of concretes in radioactive waste repositories is likely to be limited by their reaction with sulphate-bearing groundwaters. In such a situation the concrete slowly degrades by sulphate penetrating into the concrete where it reacts with the cement to cause stress and eventual disruption. Experiments have been carried out to quantify these reactions and the resulting expansions. The data have been incorporated into an overall mechanistic model for sulphate degradation in which the criterion for disruption is the accumulation of a critical amount of stored elastic energy in the reaction zone. The results of this model are in good agreement with those of an earlier empirical model and enable extrapolations to be made to different concretes and groundwaters.

In the concept of radioactive waste disposal developed in the UK, OPC blended with pulverised fuel ash or ground granulated blastfurnace slag is being considered for encapsulation of waste forms, as a material for backfilling and sealing a repository, and for concrete in repository construction.

This paper describes a laboratory study of the long term durability of such cements in contact with sulphate-bearing ground water under accelerated exposure conditions. Mineralogical analysis of the cements over the exposure period, carried out with the aid of scanning electron microscope observations and x-ray diffraction studies, provides an indication of the stability of cementitious phases exposed to an aggressive environment.

It is shown that for intact cement blocks there is minimal interaction between cement and sulphate-bearing ground water. Sulphate minerals produced by the reaction are accommodated in voids in the cement with no adverse effect on the cement structure. However, crystallisation of C-S-H and sulphate minerals along cracks in hardened cement specimens causes expansion of fracture surfaces resulting in a more accessible route for ground water intrusion and radionucleide migration.

The reaction of cement with ground water is greatly accelerated by the use of powdered material. Ettringite formed in the reaction is found to be unstable under these conditions. The mineralogical assemblage after exposure for 1 year is calcite, hydrotalcite, C-S-H and quartz.

Portland cement has been identified as a likely candidate seal material by programs investigating the deep burial of nuclear waste as a disposal mechanism. The longevity of performance of cement grout is currently being investigated, along with bentonite, under the auspices of the Stripa Project. Coordinated laboratory, field, and modeling studies are underway to produce fundamental data, practical experience, and estimates of long-range performance, respectively. Long-term performance of cement grout is of particular concern. Since most of the solid phases of which grout is comprised are metastable, it is likely that grout performance will decrease with time. The question is whether performance will still be acceptable after this decrease. This issue is being addressed with the coupled use of geochemical and permeability modeling. For a simplified cement system, two mechanisms for chemical degradation have been considered: phase change and dissolution. For dissolution, both equilibrium (slow flow) and open (fast flow) systems have been analyzed as bounding scenarios. Granitic terrain groundwaters ranging from fresh to saline have been used in the analyses. To assess the consequences in terms of flow, an empirical relation between cement permeability and porosity has been developed. Performance changes with time have been predicted by making conservative estimates of local hydraulic head conditions for successive periods of repository history. For the granitic rock environments considered, preliminary results indicate that cement grout performance may be acceptable for tens of thousands to millions of years, providing its initial hydraulic conductivity is on the order of 10−12 m/s. Other conditions favoring long-term performance include minimizing the ettringite content of the grout, and emplacement at a site where the groundwater has an elevated TDS, and where the local hydraulic gradient is flat or repository resaturation times are short.

The current concept for disposal of low- or intermediate-level nuclear waste in the UK includes a largely cementitious backfill. An important factor in determining nuclide release rates from the repository is the pH in the near-field. In this paper, we describe development of a coupled chemistry/transport model of cement degradation and its application in determining the evolution of the chemistry within the repository. In particular, the effect of the cement formulation and interaction with groundwater species on the evolution of the pH is examined. The model is sufficiently flexible to consider a number of different repository scenarios.