Determining the provenance of nuclear materials using advanced analytical techniques

: A nuclear forensics investigation of bulk versus single particle analysis

Abstract

The main research aim of this thesis was to establish routine and robust methods to sequentially characterise nuclear materials which are out of regulatory control, from the bulk to the micrometre scale, similar to standards set by nuclear forensics. The tools and techniques used were high precision gamma spectrometry, XRT, XRF, XRD, SEM, EDS, FIB, synchrotron μXRT, μXRF, μXANES and MC-ICP-MS. The nuclear materials surveyed was collected from both ends of the radioactive release spectrum; from the FDNPP accident with a massive fallout over a relatively short period versus the Winfrith ALES which had a slow chronic regulated discharge.

The physical characteristics of material from the Winfrith ALES pipe matched those of iron pipe scale, closely resembling material found from drinking water systems. There were many variations in the external surfaces, internal morphology, elemental composition and chemical phases of the material at different sampling locations from the pipe. Most notable were samples collected from the Break Pressure Tank, the highest from sea level portion of the pipe. These samples had a heterogeneous density structure made up of partially cemented Fe oxides, Fe rich spheres and discrete ultra dense micrometre scale U rich particles distributed throughout the material. These samples also had environmentally significant concentrations of radionuclides (Co-60, Cs-137, Am-241), these uniquely radiaoctive pipe scale were found to be thermodynamically stable and would pose little environmental risk.

Generally, the largest levels of radioactivity found in coastal sediments around the UK are associated with ‘hot’ particles, some of which are actinide (U, Pu) based. Since earlier bulk analysis of the Winfrith pipe scale indicated the presence of discrete U rich particles, this warranted further study beyond lab based x-ray systems. Therefore, these U particles were isolated and extracted from pipe scale and analysed using micrometre scale x-ray systems at the Diamond Light Source, Harwell, UK. The results indicated that the U particles were conclusively anthropogenic and ascribed to nuclear fuel processing. The U particles were primarily UO2, mostly co-associated with the magnetite phases of the Fe pipe scale structure. Fortunately, this implied that the overall mobility of the U particles was extremely low and thus, posing little radiological risk.

The material collected from the Winfrith ALES pipeline was also analysed using MC-ICP- MS. This work exploited the well known U and Pu isotope pair radiotracers in order to make some qualitative assessments about the general movement of this nuclear material to the wider environment. U233/U236, 236U/238U, 240Pu/239Pu and 242Pu/239Pu atom ratios possessed characteristics that were clearly distinct from isotopic signatures present in Dounreay, Sellafield or La Hague. However, there was a large degree of internal variability which suggested that the nuclear material was derived from multiple sources of varying isotopic composition. This was not entirely unexpected as the Winfrith ALES pipeline serviced many different nuclear test reactors, testing many different types of fuel.

In the context of Fukushima, Cs isotope activity ratios are useful for estimating source terms, reactor conditions and tracing. In this portion of the work, we focused on the Cs- 134/Cs-137 activity ratio. Instead of using the traditional calibration methods employed by the scientific community, such as using in house reference standards or a simple ratio-ing of the Cs-134/Cs-137 gamma peaks, this work has evaluated an intrinsic calibration method. This is particularly helpful when analysing a high volume of variable environmental samples which is usual for FDNPP sampling missions. We found that this intrinsic calibration method had excellent agreement with literature results with slightly better precision which was necessary to better discriminate between different Fukushima nuclear reactor units. This work then employed the intrinsic calibration method in order to exploit the Cs- 134/Cs-137 as a radiotracer in the environmental. Tracking this activity ratio, we were able to illustrate basic spatial variability of the Cs micro-particles from the different FDNPP reactor units. We have also established some important correlations between these ratios and temporal, terrain and material variations. Notably, we have found that material from Unit 1 has been preferentially mobilising much more than those from Units 2-3.

This research project improved and validated new analytical methods predicted to have future utility for the nuclear forensics community. All the tools and techniques developed in this work were part of the usual complement of the nuclear forensics toolkit. Therefore, all developments studied in this work can apply or easily adapted to any nuclear forensics setting although the majority of work done in this project did not have a forensics element.

Date of Award	6 Dec 2022
Original language	English
Awarding Institution	University of Bristol
Supervisor	Tom A Scott (Supervisor) & David A Richards (Supervisor)