Before one can take a decision on the technical feasibility of the geological disposal of vitrified high-level long-lived radioactive waste (HLW), the post-closure performance of the entire repository system must be assessed to be able to demonstrate the safety of the disposal concept. The common approach to obtain information for predicting the long-term performance of HLW glass basically consists of laboratory tests and modelling work. The laboratory tests include the evaluation of the various interactions between the HLW glass and the engineered and natural barriers. The experimental results are used to derive parameter values for the conceptual models that aim to describe the observed glass dissolution processes. Source terms can then be defined that describe the expected performance of the disposed HLW glass.

Surface laboratory tests inherently are performed under conditions that are not always representative for the expected repository conditions, especially when a longer test duration is envisaged. Moreover, most laboratory experiments do not integrate all components and processes that influence the glass corrosion. These drawbacks contribute to the existing uncertainties and unreliability of the modelling results. Especially the general public experiences these long-term predictions on HLW glass performance as non-convincing, relying on the argument that they are based on experiments performed under unrealistic laboratory conditions.

In situ tests in the URF HADES enable to perform glass corrosion tests under very realistic conditions, especially for longer test durations, and to integrate all the relevant processes. Therefore, the results from such in situ tests can put in perspective the results from laboratory tests, and hence contribute to increase the reliability of the glass alteration and source term model estimations for vitrified HLW.

The objective of the overall CORALUS project is to confirm that the existing long-term predictions on the performance of the SON68 18 17 L1C2A2Z1 reference glass, which are essentially based on the results and insights from surface laboratory experiments, are reliable. This will be done by confronting these data with the results on glass corrosion and radionuclide leaching and migration that are obtained from an integrated in situ corrosion test, performed under very realistic conditions (radioactive glass samples, backfill materials exerting a high swelling pressure, g?irradiation, controlled temperature,...). The conclusions of this comparison will be reported in a final report that will allow both the waste management agencies, the waste glass producers, and the regulatory bodies to evaluate their actual opinion on the technical feasibility of the geological disposal of this waste form, and on the relevance of laboratory-scale glass corrosion tests for deriving parameters for long-term HLW glass performance.

Together with the SON68 glass samples, also samples of other materials are placed on the different CORALUS test tubes for study of the global dissolution (glass), and global corrosion and local pitting corrosion (container materials):

At present, the radioactive test tubes are still in operation. Hence, there are only the first results on the composition of the interstitial solution of the different backfill materials and for the different conditions (30 °C, no gamma irradiation, 90 °C with gamma irradiation). All solutions are characterized by fairly high salt contents, which are higher for the test tubes at 90 °C. pH values are in the average about 1 unit lower for the tubes at 90 °C. In analogy with the results of the CERBERUS in situ test, we expect that this will have a beneficial effect on the glass dissolution (i.e. a lower dissolution at lower pH). In contrast, redox potentials seem to be higher (less negative) for the test tubes at 90 °C. The differences in pH and redox potential might affect the solubility and the speciation of the leached radionuclides, and hence their migration rate.

The total alpha activity in the piezometer solutions is always below the detection limit for the modules with the bentonite-based backfill materials. For the module with the dried Boom clay, the total alpha activity fluctuates around the detection limit. Alpha spectrometry showed that an occasional small measurable alpha activity is due to the natural radioactivity of Boom clay (U-isotopes). These results demonstrate also that the (four) pre-compacted backfill blocks have strongly swelled, sealing the remaining voids in the modules. This was already demonstrated during the dismantling of the first “dummy” test tube (see bibliography, EUR 19143).