Published online by Cambridge University Press: 01 January 1992
Two methods were used to measure grain-boundary inventories of 137Cs, 90Sr and 99Tc in used CANDU fuel, to corroborate source term estimates based on a fission gas release code. Used fuels were partially oxidized at 200°C in air to overall compositions of UO2+x (0.15≤ × ≤0.25) to expose UO2 grain boundaries, followed by leaching in aqueous solution. Only a fraction (2 to 18%) of the calculated gap + grain-boundary inventories for 37Cs was released. This suggests that the calculations overestimate Cs release or that oxidation does not expose all grain boundaries, or that Cs release from grain boundaries is slow. Release of 90Sr (0.01 to 0.7%) agreed reasonably well with the source term estimates (0.001 to 0.3%). Release of 99Tc (0.3 to 1.5%) suggests that the source term estimate for the upper boundary of 99Tc release (25%) may be too high. A second technique involved leaching of crushed and size-fractionated used fuel in either a static or dynamic system. A direct one-to-one correlation between calculated and measured gap + grain-boundary inventories for 137Cs was found for low- and medium-power fuels.