Modelling long-term corrosion of mild steel in unpolluted fresh and sterile seawaters

The progression of corrosion loss and maximum pit depth, once initiated, is of much interest for predicting the future or remaining long-term life of infrastructure systems. Their failure can have substantial economic and environmental consequences. The model most widely used for the prediction of long-term corrosion is a strictly concave function such as the power law c = A t B where c is corrosion loss or maximum it depth, t is time and A and B are constants obtained from fitting the function to data. However, as now shown repeatedly and for immersion, tidal, coastal atmospheric and for different steel, including chromium steels and also for freshwater and brackish water immersion, this model is inconsistent with longer-term empirical observations. These mostly exhibit a bi-modal trend, even for sterile seawaters and triply distilled freshwaters. Herein it is proposed that the bi-modal behaviour arises from changes in the nature of the corrosion reactions as the rust layers deposited on the exterior surface of the steel become increasingly less permeable, particularly to oxygen. The mechanisms and chemistry involved are described. It is shown that the bi-modal characteristic primarily is the result of a change from corrosion rate controlled by the rate of oxygen diffusion with oxygen as the electron acceptor in the corrosion process, to one involving hydrogen reduction and associated diffusion for a short period followed by other possible rate-controlling mechanisms.