Rim formation and fission gas behaviour: some structure remarks

In high burn-up LWR nuclear fuel an increase of the Xe-mobility is observed in the rim region according to EPMA. This often coincides with an increase of the local porosity and the grain subdivision of the material in regions around the pores. The restructuring does not always imply disappearance of the prior grain boundaries. This seems to occur in a final step. Micro-XRD studies also show a contraction of the fuel lattice in the rim zone, reflecting mainly the release of accumulated stresses during irradiation, via reordering of defects and defect complexes, including sub-grain formation and displacement of Xe traps. The lattice contraction is not measurable when the fraction of restructured areas is low and the prior grain structure still remains. Nevertheless, in such a case, even the Xe signal by EPMA is observed to decrease, anticipating the displacement of Xe inside the grains, probably towards cavities. However, the quantitative proportion of Xe in matrix and pores can not be given by EPMA. This is confirmed by TEM examinations, showing still plenty of gas bubbles inside restructured grains, in spite of the low Xe signal detected by EPMA. An alternative determination therefore appears necessary. The fission gas release (FGR) behaviour of the rim zone seems then to depend basically on the efficiency of gas retention in its porosity. The closed character of these pores and the low percolation probability derived from the high pore to grain size ratio anticipate a low incidence of open porosity. Also, mechanical tests suggest a low pore interconnection probability by microcracking. However, at very high local burn-ups (>150 GWd/tM), too high porosity values are determined compared to the values derived from immersion density and solid swelling, suggesting the potential existence of open channels. Also, abnormally high porosity values by quantitative metallography might arise from grain pullout during sample preparation. Here, a rough estimation of the release fraction from the rim zone is made, considering the persistence of all prior grain boundaries and the interconnection of all pores near these boundaries in a range of one pore radius. In this mostly improbable case, a maximum release of about 15% is estimated. For a precise FGR characterisation of the rim zone, quantitative determinations of the pore connectivity and of the real ratio: Xe in matrix/Xe in pores still seem to be lacking. (authors)