Localized Corrosion of Carbon Steel in the Presence of Gamma Radiation

Abstract

Canada’s long-term plan for used nuclear fuel includes permanent disposal in a deep geological repository (DGR) using a multiple-barrier system with a key barrier being the used fuel container (UFC). The UFC design consists of an inner vessel of carbon steel (CS), which provides the structural strength to withstand repository loads, and an outer layer of copper, which functions as an external corrosion barrier. The inner vessel of the current UFC design is composed of a pressure grade CS hemispherical head and pressure grade CS piping as a body, the two parts to be laser welded on-site. One of potential issues with this design is the effect of g-radiation on corrosion of the crevice near the weld region. The environment inside the sealed UFC would be humid air. Gamma-irradiation of humid air produces NOx and HNO3, and these species could accelerate the formation and condensation of water droplets in crevices (which might also have surface roughness features that act as preferential condensation sites). The chemical environment in such acidic condensed water, coupled with the presence of oxidizing water radiolysis products, could be very aggressive. Corrosion within the crevice between the head and the body of the container assembly or in the stressed regions near the welds could lead to a localized build-up of corrosion products (oxide deposits and H2gas). The effect of g-radiolysis on crevice corrosion of CS is being investigated using a combination of corrosion exposure tests and electrochemical measurements. These measurements are augmented by post-test analysis of oxide surface using SEM, XPS and Raman spectroscopy. In the corrosion tests reported here, a prepared CS crevice was wetted by adding a couple of water droplets to the mouth of the crevice. The test sample was then exposed to a desired environment in a sealed container for 20 h. Optical images of the corroded surfaces show that the corrosion of the wet bold surface progresses more extensively in air than in Ar, more at 80 oC than at 25 oC and more in the presence rather than in the absence of radiation. The composition of the oxide formed on the wet bold surface (as determined by Raman spectroscopy) shows the oxidation leads to formation of a mixed FeII/FeIII oxide/hydroxide in Ar, and to the formation of FeIII oxyhydroxide and FeIII oxide in addition to the FeII/FeIII oxide in air. Increasing temperature increases the ratio of oxide to hydroxide or oxyhydroxide but does not affect the metal oxidation states. More corrosion products are formed in the presence than absence of radiation. In contrast to the bold surface the interior surface of crevice remained smooth. The XPS analysis of the crevice surface indicates that the oxides formed there are thin (< ~ 9 nm) and that a larger fraction of the oxide is FeIII oxide/hydroxide rather than FeII/FeIII oxide. The thickness of the oxide formed at 80 oC is slightly higher than that formed at 25 oC. Radiation slightly decreases the oxide thickness at 25 oC while it increases the oxide thickness at 80 oC. The different corrosion behaviours of the crevice and bold surfaces raise the possibility of galvanic coupling between the two surfaces. This galvanic effect was investigated by monitoring the coupling current between two electrodes, one electrode representing the bold surface and the other representing the crevice surface, in the electrochemical coupling tests. To aid in interpretation of the galvanic coupling experiments, the corrosion potential and linear polarization measurements were made on each electrode separately. The post-test surfaces from these tests were examined by SEM and XPS. The observed effects of temperature and g-radiation on the corrosion potentials are consistent with the different oxide formation and growth observed in the coupon exposure study. The corrosion potential on the crevice electrode is always slightly higher than that of a bold electrode with the difference depending on temperature and radiation environment. A galvanic coupling current observed between the two electrodes shows that the crevice electrode is the net cathode. The measured coupling currents are consistent with the results from the dynamic polarization tests performed on each electrode. A CS crevice corrosion mechanism that can explain both the crevice corrosion tests and the electrochemical results is discussed.