The basic fuel unit of the High Temperature Reactor (HTR) of the Pebble Bed Modular Reactor (PBMR) is a uranium dioxide kernel coated with a buffer layer, an inner pyrolytic carbon (IPyC) layer, a silicon carbide (SiC) layer and an outer pyrolytic carbon (OPyC) layer and is commonly referred to as a TRISO particle. Thousands of these micro-spheres are embedded in a graphite matrix and pressed to form a fuel sphere. During the manufacture of the TRISO particles and the fuel spheres there is a production of TRISO particles with cracked/broken layers, especially the SiC layer. Before the irradiation of the fresh fuel in the nuclear reactor it is of the utmost importance to quantify the failed fractions in fresh fuel as this information is very useful in the general understanding of fuel behaviour, calculation of risk and safety margins, and prediction of long term fuel behaviour. For this reason the burn-leach method has been applied for the quality control of the fresh fuel. In this work, several aspects of the burn-leach method that affect the efficiency of the method were studied. Aspects that were investigated are: qualitative aspects, layer properties, quantitative aspects, variants of the burn-leach method and lastly statistical information from the burn-leach data.
The results obtained were as follows: Studies in this dissertation suggest that partial leaching of uranium in TRISO particles with a defective SiC layer was a phenomenon that exists. Although UO2 kernel equivalents were successfully determined by burn-leach method for particles with fully broken SiC layers, certain particles leached uranium amounts that did not correspond to single UO2 kernel equivalents; Evidence of occurrences of ‘slow leaching’ in an acidic medium were evident for certain particles. There were remnants of uranium dioxide kernels that had been partially leached after the full 16 hours. This behaviour led to inconclusive results on the absolute number of defective particles in a given population; Investigations suggest that there is at least circumstantial evidence that the BL method combined with X-ray tomography provides information about the integrity of the SiC layer, and why one particle leaches and the other does not. Neither the burn-leach nor the leach-burn-leach analysis is sufficient to be used as a stand-alone method to quantify the number of particles with defective SiC layers in a given TRISO particle population. The two tests need to be coupled to other techniques such as high resolution tomography for an extensive quantification of the layer defects; Burn-leach has to be designed to test for the layer integrity on a microscopic level as opposed to testing for the broken shells only, as was done by the normal burn-leach based on the German program. The leach time was not sufficient in its present form; Burn-leach results indicated that oxidation times of 96 hours at 750 °C under atmospheric pressure did not negatively affect the mechanical strength of the silicon carbide layer of freshly-manufactured TRISO particles, as these particles did not have a high failure fraction.