It is a widely-held expectation of corrosion specialists that carbon steels are subject to crevice corrosion in seawater. However, there is little, if any, practical evidence of such corrosion (the long-term test provided by the wreck of the Titanic and many other ships with riveted hulls might be expected to reveal the problem if it exists). Previous work [1] that measured the pH and potential inside an artificial crevice on carbon steel suggested that, at the free corrosion potential, the crevice behaved as a nett cathode and became alkaline. This can be rationalized by appreciating that the crevice tends to develop an increased level of ferrous ions, which causes the iron to ferrous equilibrium potential to rise to about the free corrosion potential, thereby reducing the nett rate of iron dissolution. However, some workers remain unconvinced (for example see the discussion entitled "Crevice corrosion of steel in seawater" on the NACE Corrosion Network in 2004). Owing to the large number of possible reactions it is difficult with simple 'manual' models of crevice electrochemistry to obtain an accurate indication of crevice behaviour. Numerical models of crevice behaviour allow us to test ideas about what happens inside a crevice, and this paper will describe a model of crevice behaviour that has been developed using a commercial finite element package.

Reference

1. A. Alavi and R.A. Cottis, "The Measurement of pH and Chloride Concentration in a Simulated Crevice", "Embrittlement by the Localized Crack Environment", ed. R.P. Gangloff, pp75-88, TMS-AIME,  Philadelphia, 1983.