MGAU ENRICHMENT MEASUREMENTS OF U/TH COMPOUNDS: CHALLENGES AND RESULTS

Determining uranium content is one of the most common measurements in nuclear safeguards. In most cases these types of measurements include determination of uranium enrichment, which is then either used in conjunction with other information in the derivation of total uranium mass, or simply to categorize the special nuclear material (SNM) present. The Multi-Group Analysis code for Uranium (MGAU) is one of the widely used computer codes for the determination of the relative abundances of the uranium isotopes, and it has proved to work well for many different types of samples. In some cases, however, the enrichment analysis results may be compromised by a non-optimal counting geometry, non-ideal detector availability, and adverse sample conditions. One particular example is the measurement of uranium bearing materials containing high quantities of thorium. This type of material can be generated at a uranium conversion plant, where uranium tetrafluoride is fed into a flame reactor with gaseous fluorine to be converted to UF6. During the conversion process non-volatile uranium compounds, called unburnts, with high thorium content are accumulated and stored at the conversion facility. Unburnt material exits at the NECSA Uranium Conversion Plant in South Africa. The uranium needs to be measured and verified. It has been long known that the presence of natural thorium in uranium bearing materials may lead to a positive bias in the MGAU analysis results. This bias can sometimes be quite significant for samples having high thorium content, and as a result the latest version of MGAU (v4.2) has been improved to detect the presence of thorium in the sample and minimize the effect. A set of test measurements has been performed on drums containing unburned material generated at the NECSA Uranium Conversion Plant. The unburnt material in the form of dry UF4 cake with high thorium content was measured using a high purity germanium detector. The measured spectra were analyzed using both the latest version of MGAU (v4.2) as well as the older version (v3.2), which is still commonly used in the safeguards community. The MGAU (v4.2) results showed a significant reduction in the reported enrichment bias for the measured samples of this type.