Chemistry of corrosion products in lead-bismuth cooled nuclear systems: oxides and intermetallics of iron, nickel and manganese

MYRRHA is an accelerator-driven system, combining a linear particle accelerator and a subcritical nuclear reactor core. MYRRHA uses the liquid leadbismuth eutectic alloy (LBE) as spallation target material and primary coolant. For a reliable operation of such system, high purity of the LBE coolant should be guaranteed. A precise control of the dissolved oxygen concentration (DOC) in LBE can mitigate the two major sources of coolant contamination, namely the impurities continuously released by corrosion and the coolant oxidation itself (PbO). Although the DOC mitigates corrosion, the release rate of corrosion products is still sufficient to affect the DOC control by reacting with dissolved oxygen in LBE. In addition, the corrosion products can interact with LBE itself forming solid precipitates that can precipitate in cold regions of the reactor and cause blockages. Thus, the general aim of this research is to study the chemistry of corrosion products in LBE to support the design of appropriate oxygen control and LBE purification system. The methodology of this work relies heavily on the use of potentiometric oxygen sensors, which provide a highly accurate in-situ method to measure and control the DOC in LBE (Chapter 3). Combined with thermodynamics (Chapter 2), this technology allows studying oxides in LBE, their equilibrium and kinetics. Furthermore, the DOC measured at the oxide equilibrium can be used to characterize the behavior of metallic impurities, which can be used to model their interaction with each other and LBE. Three corrosion products might influence reactor operation: iron, nickel and manganese. Iron is proven to form magnetite (Fe3O4), which is a dominant oxide in LBE and, together with PbO, determines the DOC control limits. In Chapter 4, the Fe3O4-Fe2O3 reaction was identified in LBE in the MYRRHA operating range by comparing the response of the reaction DOC to the addition of metallic impurities. The measured data enabled selection of appropriate thermodynamic data to model its equilibrium. In Chapter 5, the reaction was shown to take place on the surface of suspended Fe3O4 particles and to follow the adsorption model with a negative apparent activation energy. A validation case confirmed the ability of the proposed model to reproduce the DOC trend during cooling. Nickel (Ni) could be problematic due to the formation of NiO and Bi3Ni. NiO is theoretically predicted to form in the MYRRHA operating range and therefore has the potential to affect the DOC, while the conditions for Bi3Ni formation in LBE are not well defined. To address this, the behavior of Ni in LBE was investigated (Chapter 6). Firstly, the activity coefficient of Ni was derived from the NiO equilibrium measurements in LBE. These data fill the gap in the literature data, and therefore clarify the conditions for Bi3Ni formation in LBE. Secondly, the conditions for NiO nucleation were identified in LBE with high nickel concentration. Based on these experiments, NiO can nucleate in the LBE bulk and on the surface of PbO particles, but it is unlikely to occur at the dissolved nickel concentrations expected in MYRRHA. Manganese (Mn) is one of the main activation products and, like Ni, it can form an intermetallic with bismuth (BiMn1.08), but the conditions for its formation are unknown. Therefore, the final part of this research focused on the behavior of Mn in LBE (Chapter 7). Similar to the NiO study, the MnO equilibrium was used to measure the activity coefficient of Mn in LBE. The activity coefficient of Mn shows the same characteristic deviation from the theoretically expected trend as the activity coefficient of Ni, which can be justified in terms of the associate solution model. Calculating equilibrium between dissolved manganese, its pure solid phase and BiMn1.08, two correlations describing the maximum concentration of dissolved manganese in LBE were derived.