Corrosion Behaviour of AGR SIMFUELS

In the UK, the majority of spent nuclear fuel (SNF) is from indigenous Advanced Gas-cooled Reactors (AGRs). AGRs, whilst 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 rather than zircalloy as is the case in Pressurised Water Reactors (PWRs). Consequently, AGR fuel has unique characteristics that need to be evaluated before disposal in an expected anoxic UK geological repository.

Pure UO₂ pellets and, for the first time, AGR SIMFUEL pellets simulating 25 (pellet interior) and 43 GWd/tU burn-up (rim structure) have been fabricated at the UK National Nuclear Laboratory. Using combined electrochemical and Raman spectroscopic studies, we have investigated the effect of the SIMFUEL dopants on the UO₂ crystal matrix. We have also studied the effect of H₂O₂ on the SIMFUEL surface as a simulant for the effect of a radiolysis on intruding groundwater in the near field.

Raman microscopy reveals an increase in the lattice damage (500 – 700 cm^-1) with increasing burn-up, as the UO₂ cubic fluorite lattice structure becomes more distressed and moves towards a tetragonal structure. Cyclic voltammetric studies of the AGR SIMFUELs reveal them to more susceptible to electrochemical oxidation then PWR SIMFUELs; in particular the peak associated with the in-grain UO₂ to UO_2+x oxidation (onset -0.2 V vs SCE,  E_p = +0.14 V) is ~3X larger for AGR SIMFUELs than PWR SIMFUELs of similar simulated burnup. Importantly, analogous studies on PWR SIMFUELs have shown that UO_2+x locations act as preferential corrosion sites.

The open circuit potential (E_OC) for UO₂ and AGR SIMFUELs show a logarithmic dependence on [H₂O₂] in modified simplified groundwater(10 mmol dm^-3 NaCl, 2 mmol dm^-3 NaHCO₃, pH 6.8). In contrast, the E_OC of PWR SIMFUELs of ~0.1 V vs SCE in aqueous alkaline solution is near independent of [H₂O₂], an observation attributed to surface-catalysed H₂O₂ decomposition by non-stoichiometric U(IV)/U(V) sites. This suggests that, in contrast to the non-complexing alkaline conditions, carbonate complex-ation of U(VI) in groundwater simulant removes the H₂O₂ decomposition sites from the UO₂ surface so resulting in a larger net [H₂O₂] and thus oxidative stress being induced at the SIMFUEL surface.

At [H₂O₂] = 10 mmol dm^-3, close to that encountered in an a dominated near field, H₂O₂ generates an oxidative stress of 0.05 V vs SCE on 43 GWd/tU AGR SIMFUEL in simplified modified groundwater. This corresponds to the near top of the voltammetric wave for the in-grain UO₂ to UO_2+x process, suggesting that near field peroxide may be capable of inducing a substantial population of anodically active UO_2+x sites on AGR SNF and consequent localised site corrosion at the spent fuel surface.

Extended duration Raman studies on the oxidation of UO₂ on 25 GWd/tU SIMFUELs in 1 mmol dm^-3 H₂O₂ (E_OC  = +0.19 V vs SCE) in O₂-free deionised water reveal that, over a 1 week period, there is an increase over in the intensity of the peak at 540 cm^-1, associated with the formation of defect sites in the UO₂ fluorite lattice. Shorter term studies on undoped UO₂ samples in 0.01 to 1 mmol dm^-3 H₂O₂ solutions also reveal the in-growth of a peak at ~610 cm^-1, a feature that has been attributed to O sublattice distortions due to the generation of UO_2+x locations. This again indicates of the potential for near field concentration H₂O₂ to induce an increased susceptibility to corrosion for AGR SIMFUEL.

The ultimate product of H₂O₂ induced corrosion of either UO₂ or AGR SIMFUELs would be expected to be U(VI). The extended duration Raman studies show that a peak appears at 820 cm^-1 after 5 hours exposure to 1 mmol dm^-3 H₂O₂. This grows in intensity up to 50 hours exposure and has been assigned to metastudtite (UO₂(O₂) 2H₂O), a uranium peroxide secondary phase formed at the SIMFUEL surface as a result of the reaction between H₂O₂ and UO₂²⁺ – the latter necessarily generated as a result of above-expected H₂O₂–driven corrosion of the AGR SIMFUEL UO₂ matrix.

Peroxide aside, open circuit potential measurements on 43 GWd/tU simulated burnup AGR SIMFUELs samples in intimate contact with 20/25/Nb stainless steel and in modified simplified groundwater electrolyte find that they exhibit a mixed potential of ~-0.12 V vs SCE. This corresponds to nearly a quarter of the way up the oxidation wave for in-grain UO₂ to UO_2+x process in the associated voltammogram – implying that UO₂/steel coupling can induce UO₂ corrosion even in the absence of a peroxide generated oxidative stress.