Corrosion Behaviour of AGR Simulated Fuels (SIMFUELs)

Rauff-Nisthar, Nadya and Boxall, Colin (2022) Corrosion Behaviour of AGR Simulated Fuels (SIMFUELs). PhD thesis, Lancaster University.

Text (2021Rauff-NistharPhD)
2021Rauff_NistharPhD.pdf - Published Version
Download (17MB)

Abstract

The UK radioactive waste inventory includes spent fuels from various reactor types, many of which have compositions and irradiation histories that are unique to the UK. These include uranium oxide fuels from existing Advanced Gas-cooled Reactor (AGR) power stations, of which approximately 4,500 tonnes is forecast not to be reprocessed and thus will require disposal in a Geological Disposal Facility (GDF)(1). The direct disposal of spent nuclear fuel into a geological repository is part of the nuclear waste policy of several mature nuclear states. The vast majority of this fuel is from light-water moderated reactors (LWR) and a significant amount of research has been carried out to support the direct disposal concept in terms of the physical and aqueous durability of the container, the cladding and the irradiated UO2-based fuel. In the UK, a large proportion of SNF is from indigenous Advanced Gas-cooled Reactors. AGRs, whilst also using UO2-based fuel, employ CO2 as coolant and are graphite moderated. Further, the fuel assembly cladding is comprised of 20/25/Nb stainless steel (20% Cr, 20% Ni) rather than zircalloy as is the case in Pressurised Water Reactors (PWRs). Consequently, AGR fuel has unique characteristics that need to be evaluated in order to satisfy safety case requirements before it can be disposed of in a geological repository. It is expected that, at some point after repository closure, the canisters within which the SNF is sealed will fail, allowing the ingress of groundwater. This groundwater may come into contact with the 20/25/Nb stainless steel cladding and subsequently the UO2 fuel pellets, initiating corrosion on either one or both elements of the SNF. At this failure point, most β/γ radiation will have decayed away, with just α-radiation remaining, initiating the process of the α-radiolysis of water. Alpha-radiolysis of the invading groundwater contacting the UO2 fuel pins will generate H2O2, which through a conversion mechanism can lead to the formation of studtites. Furthermore, the 20/25/Nb cladding may also show susceptibility to near field corrosion. The objectives of this work are therefore to examine the mechanisms that occur during the corrosion of annular AGR simulant fuel. This will lead to the establishment of the baseline corrosion behaviour and give a better understanding into the formation of secondary uranyl phases. iii | P a g e A range of AGR SIMFUELs of simulated burn-ups have been fabricated. Using combined Raman and electrochemical studies, the effects of the added dopants on the UO2 crystal structure have been investigated. The effect of deliberately added H2O2 on the surface of simulated AGR fuels has also been studied. Additionally, the electrochemical baseline behaviour of the minor actinide americium in nitric acid media representative of reprocessing within the nuclear fuel cycle and the possible subsequent uptake of americium onto stainless steels representative throughout the nuclear industry have also been investigated. It was found that SIMFUEL samples prepared in the absence of dopants (pure UO2 pellets) exhibit the cubic fluorite structure expected of UO2. This is observed for both 25 and 43 GWd/tU burn-up SIMFUEL, reflecting the introduction of dopant-associated lattice defects into the H2O2 matrix. As might be expected, the peak intensity is greater for the 43 GWd/tU sample. Voltammetry indicates the presence of three oxidation waves in neutral media typical of UK groundwater concepts: the oxidation of UO2+x sites at grain boundaries at -0.6 V vs SCE; the oxidation of UO2 to UO2+x within the grains at -0.1 V; and the oxidation of UO2+x to U(VI) species at +0.3 V. The currents associated with all three waves increase with simulated burn-up indicating a similarly increasing susceptibility to anodic corrosion reactions with burn-up. The size of the feature at -0.6 V suggests a high concentration of UO2+x sites at grain boundaries. As these may act as preferential and localized corrosion sites, it may be that fission products residing in grain boundaries will be instantly released on contact with groundwater. A literature review exploring the effects of hyperstoichiometric phases on UO2 dissolution suggest that spent fuel, where grain boundaries can contain non-stoichiometric sites, is able to corrode more quickly in comparison with stoichiometric grains. Therefore the assumption that radionuclides on grain boundaries will be released instantly upon contact with groundwater is a realistic assumption (2)). An increase in the lattice damage peak in the Raman spectrum is observed when the SIMFUEL is exposed to H2O2 representative of a repository near-field, suggesting that the structure at the surface is becoming more distressed. Supported by concurrent EOC measurements, this is consistent with additional point defects being established as the concentration of interstitial oxygen is increased in the lattice via H2O2 induced surface oxidation of the SIMFUEL – so increasing the iv | P a g e concentration of UO2+x sites both on and within the grains and thus increasing the susceptibility of the SIMFUEL to corrosion. Open circuit potential measurements on 43 GWd/tU burn-up AGR SIMFUEL in modified simplified groundwater indicate that the SIMFUEL may corrode in the modified simplified groundwater even in the absence of a peroxide generated oxidative stress. Similar studies of the same SIMFUEL held in contact with samples of the 20/25/Nb steel cladding show that this coupled system exhibits a mixed potential of ~-0.12 V vs SCE. This potential is (i) negative of that adopted by the isolated SIMFUEL and (ii) positive of that adopted by the isolated cladding in electrolytes of the same composition implying that, whilst corrosion at SIMFUEL’s grain boundaries may still obtain, the UO2 matrix is at least in part protected against on-grain corrosion at the expense of the steel cladding. In the case of americium we found that, by use of cyclic voltammetry, it is possible to indirectly observe the electrochemical oxidation of Am(III) to Am(VI) on glassy carbon electrodes. This oxidation occurs stoichiometrically, yielding a 100% conversion to Am(VI) in the local diffusion layer of the electrode. The electrochemically generated higher oxidation states of Am are stable on the timescale of a typical voltammetric experiment, opening the possibility of an electrochemical route to the reagentless generation of Am(VI) in advanced Am separation schemes that employ the higher oxidation states of americium. It was also confirmed that, under conventional solvent extraction conditions, Am(III) cannot act as a process steel corrosion accelerator via the generation of Am(VI).