EN | FR

Studies of behaviour



Natural analogues and protective gels

Studies of behaviour
Laboratory studies of the long-term behaviour of waste packages carried out on the shielded process line at the ATALANTE facility in Marcoule.
CEA/P.Stroppa

The long-term containment of the radioactive atoms contained in the packages is mainly guaranteed by the resistance of the containment matrices to their main enemies: water and irradiation by the radiation inside them. Through research, models for the release of the radioelements have been developed for each containment matrix in various situations.

These models do not claim to predict precisely the behaviour of the packages over extremely long periods of up to a million years. They cannot reproduce the full complexity of the phenomena at work over these very long periods, but they do try to provide a reasonable estimate of the maximum risk based on current knowledge. They therefore maximise the quantities of radioelements released and minimise the life of the containment.

These estimates depend on the conditions in which alteration takes place, particularly with glass matrices. The extreme is what happens after the end of the disposal period, once the different containment barriers have degraded, the container has lost its containment capability and the package comes into contact with water, and through that with the surrounding environment.

Archaeological glass
Studying archaeological glass that has been in the sea for nearly 2000 years, since Roman times, tells us a lot about the effects of the fracturing of glass. Not only does the existence of this glass demonstrate that a block of nuclear glass could withstand the constant, aggressive renewal of seawater for millennia, but also that any cracks appearing in it may be filled in as time passes.
CEA/DCom

With vitrified waste, the release of heat and self-irradiation do not cause significant releases of radioelements in the long term provided that the disposal conditions do not deteriorate, or if they do, provided that there is no water present. The main phenomena that can significantly affect long-term behaviour are encountered only in 'open' systems saturated with water.

Research carried out on glass that is the natural analogue of nuclear glass is contributing to the validation of models of long-term change. Scientists now understand how the glass changes over very long periods (from several thousand years to several million years). They can now check whether their models agree with these changes and with knowledge acquired in the laboratory.

Damage due to the recoil of a nucleus
When an atom decays, the nucleus that recoils and the rays emitted inherit some of the energy released, which on an atomic scale is vast. They lose this energy in the form of heat, colliding with the atoms in their environment and displacing them by a few hundred microns. The figure shows a cascade of collisions started by a 70 keV recoil nucleus. The dots correspond to all the displaced oxygen atoms. The scales are expressed in angstroms (ten thousandths of microns or ten millionths of millimetres).
Clefs CEA N°53

When water is present, glass is eroded but its erosion leads to the formation of a thin layer of amorphous material: a gel. This phenomenon is observed with all types of nuclear glass. The gel, which consists mainly of silica, forms a barrier between the healthy glass and the solution, drastically slowing down its continued erosion. The evolution of the gel has been studied in the laboratory. It has been demonstrated that, in the containment conditions of a waste disposal facility, not only can the continued existence of the gel be guaranteed but the increase in the protection it provides over time can also be predicted. With R7T7 borosilicate glass, it would take a thousand years to erode the glass over a thickness of one micron.

In the last 40 years a large number of studies have been carried out on the long-term behaviour of nuclear glasses, which are a proven matrix for high-level waste packages. Overall there has been good progress with modelling the release of radioelements from industrial waste packages, but the parameters used in these models still need to be refined further.

Go to page in French