Pourbaix diagrams, solubility predictions, and corrosion product deposition modelling for the supercritical water reactor

Thumbnail Image
Journal Title
Journal ISSN
Volume Title
University of New Brunswick
The supercritical water-cooled reactor (SCWR) is one of the Generation IV nuclear reactor concepts. The Canadian version, the CANDU® -SCWR, promises to improve on existing reactor technology with a thermodynamic efficiency of ⁓50% along with a reduction in capital cost realized mostly by eliminating the need for steam generators. All SCWRs will operate with water above its thermodynamic critical point as the coolant. Water undergoes drastic changes in its physical properties as the critical point is approached and traversed. As a result, material selection promises to be challenging and predictive models that can provide insight into corrosion at these extreme conditions is seen as crucial to SCWR development. Semi-empirical models have been developed to allow for thermodynamic extrapolations of chemical species from standard states to elevated temperatures and pressures. Of these, the most widely used is the revised Helgeson-Kirkham-Flowers (RHKF) model. By using this model, after carefully optimizing its parameters to several published solubility studies, predictions of metal-oxide stability, solubility and transport under supercritical conditions can be made. Pourbaix diagrams were drawn for iron, nickel, chromium, aluminum, and titanium at conditions slightly below and slightly above the critical point for several dissolved corrosion product concentrations. The diagrams emphasized the need for strict water chemistry control to avoid unwanted transitioning from passive oxides to active corrosion regions. The high temperature solubilities of magnetite, bunsenite, eskolaite, alumina, and rutile over a range of temperatures including the supercritical regime have been estimated. For each oxide, it was found that the solubility dropped dramatically at the critical point, mirroring the drop in water density. Additionally, it was determined that of the potential chemicals that can be added to the coolant for pH control, lithium hydroxide was able to marginally affect pH above the critical point while ammonia could not. Building upon the calculations of solubility and metal-oxide stability, the onset and extent of corrosion product deposition along a hypothetical fuel channel was examined through a comprehensive corrosion product transport and deposition model. It was demonstrated that the dissolved corrosion product concentration in the coolant plays a large role in the deposition behavior along the fuel channel and highlights the argument that feedtrain corrosion must be controlled as best possible to minimize in-core deposition. It was also found that pH control agents were somewhat effective in reducing crud deposit thickness - mainly due to their lowering the solubility of the corrosion products along the feedtrain leading to lower concentrations entering the reactor core.