Abstract
Owing to microstructural effects, the corrosion rate of stainless steels undergoing intergranular corrosion in nitric acid (HNO3) solutions can be strongly anisotropic and hence corrosion testing of specimens containing exposed side and end grain can lead to significant overestimation of the corrosion rate of plate surfaces. In addition, prolonged testing (typically in excess of a thousand hours) is normally required to reveal the steady-state corrosion rate, which is attained only after the onset of stable grain dropping. It has been found that by use of a pretreatment that induces grain dropping, the long-term rate eventually shown by untreated (i.e., initially ground) specimens can be revealed within a relatively short time. A method is also described that allows deconvolution of the corrosion rate with respect to each exposed face, thus allowing reliable prediction of the long-term penetration rate of plate surfaces. The deconvoluted plate corrosion rate is shown to be in good agreement with the metal's true penetration rate obtained by sectioning, provided that the former refers to the stable grain dropping condition. Significantly, the penetration rate of the metal is shown to be time-independent, displaying no induction period, and is not altered by the onset of gross grain dropping. The apparent increase in corrosion rate with time found by weight-loss measurements made prior to the onset of stable grain dropping is found to be an artefact, reflecting uncompensated increase in the area undergoing active corrosion. The present investigations form the basis of a rapid test procedure with the potential to become a standard method for measuring long-term intergranular corrosion rates.
