Durability of cemented waste in repository and under simulated conditions

The Romanian Radioactive Waste National Repository for low level and intermediate level radioactive waste was built in Baita - Bihor county, in an extinct uranium exploitation. The site is at 840 m above sea level and the host rock is crystalline with a low porosity, a good chemical homogeneity and impermeability, keeping these qualities over a considerable horizontal and vertical spans. To obtain the experimental data necessary for the waste form and package characterization together with the back-filling material behaviour in the repository environment, a medium term research programme (1996 - 2010) was implemented. The purpose of this experimental programme is to obtain a part of the data base necessary for the approach of medium and long term assessment of the safety and performance of Baita - Bihor Repository. The programme will provide: a deeper knowledge of the chemical species and reaction mechanisms, the structure, properties and performances of the final products. For safety reasons the behaviour of waste package, which is a main barrier, must be properly known in terms of long term durability in real repository conditions. Characterization of the behaviour includes many interactions between the waste package itself and the surrounding near field conditions such as mineralogy, hydrogeology and groundwater chemistry. To obtain a more deeper knowledge of the species and physical-chemical reactions participating in the matrix formation, as well as their future behaviour during the disposal period, a thorough XRD study started in 1998. For Romanian Radioactive Waste National Repository (DNDR) Baita - Bihor the following steps are planned for the conditioned waste matrix characterization in simulated and real conditions: - preparation and characterization of normal reference matrices based on different cement formulations; - preparation of reference simulated sludge cemented matrices containing iron hydroxide and iron phosphate; - selection of real and simulated disposal conditions for reference samples; - emplacement of reference samples in Baita Bihor Repository for real disposal characterization. All samples were positioned in five locations of Baita-Bihor repository where they will be kept for 1, 3, 5 and 10 years before mechanical and structural characterization. These locations have different humidity values, decreasing from location 1 to 5. The results will be compared with those obtained for the same type of samples kept in laboratory conditions. The mechanical performances of the matrix with cement-mineral additives are not influenced if bentonite or volcanic tuff concentration is is lower than 10%. The very good mechanical behaviour of the samples, drives at the conclusion that the additive cement-mineral mixture could be used in the disposal procedure, with good results, if the future structure characterization would prove the role as engineered barrier. Complexing reagents reduce drastically the mechanical performances. The chemical nature of the complexing agent plays an important role, the damaging effect being more pregnant for tartaric and citric acid than for oxalic acid. In any case there is a maximum limit of concentration which is acceptable from mechanical point of view. The mechanical resistance observed in all samples, after one year disposal, are within the accepted limits for the matrix durability. Thus, disposal conditions has not influence the mechanical properties of the cement matrix after one year disposal. The presented results, showed that bentonite seems to have a positive influence on the quality of the examined matrices . These results are preliminary. The data obtained so far will be completed with those to be obtained after three year disposal. XRD initial investigation of waste samples showed complex changes in the structure and components. To understand the systems, series of different compositions need to be characterized. In the next years the durability of samples kept for times longer than one year in normal and simulated environment will be determined. (authors)