Pitting Corrosion Probability of Copper in Solutions Containing Chloride and Sulfate Using Coupled Multielectrode Arrays

Abstract

The preferred disposal method for used nuclear fuel is to seal it in containers emplaced in a deep geologic repository (DGR) in a suitable rock formation. In Canada, the container is a robust steel vessel coated with 3 mm of Cu as a corrosion barrier. Eventually, the DGR environment will become anoxic, and in the absence of oxygen copper is thermodynamically stable in water. However, since exposure conditions will initially be oxidizing due to the presence of trapped O2 upon sealing the DGR, there is a possibility of localized corrosion to occur. Mass balance calculations show that uniform corrosion during this period will be within the designed corrosion allowance, however, since the Cu coating is relatively thin and the required lifetime is long, the risk of pitting corrosion during this period must be carefully evaluated.Pitting corrosion susceptibility is determined by the corrosion potential of the material (Ecorr, the potential at which the rates of all anodic and cathodic reactions are equal), and the breakdown potential (Eb, the potential at which the copper oxidation rate on the surface increases abruptly due to the breakdown of the protective film), both of which are distributed parameters. This is due to the stochastic nature of passive film rupture and any uncontrollable variations in the structure and local environment at the metal surface. It is important to note that the analysis is based on the concept that pitting is only possible if the Ecorr is equal to, or more positive than the Eb.Our previous studies showed a passive behavior of copper surface when exposed to high pH solutions. In this work, a multielectrode array was used to obtain a statistically meaningful set of potential measurements through the simultaneous monitoring of 30 electrodes to determine their corrosion and breakdown potentials. From the extensive database that acquired, the Ecorr and Eb distributions were defined for the copper electrodes exposed to solutions containing various chloride concentrations and a more thorough evaluation of the relative values of Ecorr and Eb has undertaken to determine the susceptibility to pitting.Experiments were conducted using a multielectrode array in chloride and sulfate containing solutions at 25 °C and pH 8 and 11. In alkaline solutions, the copper oxide is more stable and can potentially form a passive film, whereas chloride and sulfate promote the breakdown of the passive film. As the concentration of aggressive anions increased, the values of Ecorr and Eb decreased, suggesting that the solubility of copper increased with in the presence of both chloride and sulfate. However, there is a critical sulfate concentration (~5 × 10-3 M) that above which the pitting probability of copper decreased.In general, the possibility of pitting increased with growing overlap between the distribution curves of Ecorr and Eb. To evaluate the possibility of pitting, histograms and distribution curves of Ecorr and Eb in the solutions that contain different chloride and sulfate concentrations were compared. In lower chloride concentrations, there was a very small overlap between distributions of Ecorr and Eb values; therefore, the possibility of pitting under these conditions is very low. On the other hand, a high chloride concentration at high pH indicated a higher probability of pitting, as the overlap between Ecorr and Eb distributions increased. Also, the extent of the passive zone decreased with increasing chloride concentration. In solutions containing sulfate content, the overlap between Ecorr and Eb increased with increasing sulfate concentration up to a certain content (~5 × 10-3 M), after this point the overlap decreased as the sulfate concentration increased. The results from potentiodynamic experiments indicated that increasing the chloride concentration resulted in decreased the extent of the passive zone; however, the sulfate did not seem to play a significant role in changing the extent of the passive zone. Figure 1