De-alloying of Stainless Steel and Monel induced cleavage model of SCC, this is an essential point.

Benjamin Lynch, Jane Deakin, The second study used Monel 400 in various Steven M. Hodges, Zehua Dong and Roger C. Newman conditions of cold work, exposed to copper (II) sulfate solutions. The de-alloying that occurs in this system is UMIST, Corrosion and Protection Centre very similar to that occurring in Cu-Zn or Cu-Al systems, Sackville Street, PO Box 88 with the lower anodic overpotential of the Ni being Manchester, M60 1QD, UK compensated by its larger concentration in the alloy (though some limited de-alloying and SCC was also The relationship between de-alloying and stress corrosion demonstrated in 70Cu-30Ni). The de-alloying in this cracking of solid-solution alloys is well established, system is strikingly intergranular but does not appear to especially for model gold-base systems such as Au-Cu or be related to any specific grain-boundary phase; indeed Au-Ag. In engineering alloys such as stainless steels, such the predominant mode of SCC in slow strain rate tests correlations are more difficult to make, partly because de- was transgranular. Intergranular “SCC” has not been alloying always occurs under conditions of simultaneous demonstrated conclusively in our work, though obviously oxidation of the alloy components. Even in copper-base the rate of penetration of the de-alloying along boundaries alloys, where in principle de-alloying might occur under can be enhanced by opening up the crack. The conditions of thermodynamic immunity of copper, in phenomenon is striking evidence that equilibration of the practice it is found that the copper must be dissolving, or more noble alloy component with its dissolved ions can at least equilibrated with its dissolved cations. For dramatically enhance de-alloying, just as in Cu-Zn or Cu- stainless steels, where nickel is the most noble alloy Al. In those materials the effect of the dissolved Cu was component, interpretation of chloride-induced SCC in to permit de-alloying at low Zn or Al concentrations, near terms of de-alloying is persuasive but analytically the minimum possible which is the 3D site percolation difficult, not least because Ni-rich de-alloyed layers less threshold for the fcc lattice. In the case of Cu-Ni the effect than 50 nm thick are unstable in air and their pores are is to compensate for the rather small difference in filled with oxidized Cr, which complicates depth profiling standard electrode potentials of Cu and Ni, in other words analysis. The present work is part of a study of caustic to allow enough surface mobility of Cu that Ni can SCC of stainless steel, which provided an opportunity to emerge continually at a modest anodic overpotential (less explore the de-alloying behavior of stainless steel under than 600 mV). conditions of continuous layer growth, with little or no Ni dissolution. In parallel with this work, we developed an In the Cu-Ni system the layer composition is interest in a peculiar de-alloying mechanism that can lead close to pure Cu and shows no sign of intermediate to SCC of Monel 400 (70Ni-30Cu). compositions as seen for stainless steel. The 50-50 layer composition appears to be characteristic of very small The stainless steel study used 316L material anodic overpotentials (for iron in 50% NaOH at the open- which was exposed to aqueous NaOH concentrations up circuit potential of stainless steel, less than 200 mV) and to 50% at temperatures near the boiling point and in the perhaps also assisted by high temperatures. At ambient absence of dissolved oxygen. Previous authors had temperature the corresponding overpotential might be reported “nickel” layers under similar conditions, but the closer to 500 mV, in which case de-alloying would not be reality is more interesting. The stainless steel surface seen because we would be in a region of Ni passivation develops a uniform de-alloyed layer whose composition is which would freeze its surface mobility. very close to 50Fe-50Ni (atom %), with smaller amounts of O and Cr that probably reflect the presence of oxidized Electrochemical impedance spectroscopy proved Cr in the pores of the layer. X-ray diffraction shows peak to be a very useful way of monitoring de-alloyed layer broadening and shifting consistent with the presence of a growth in the stainless steel system. If the capacitance solid solution rather than a mixture of pure Ni and results are taken at face value, the ligament size in the de- unoxidized stainless steel. alloyed layer is at least as fine as that seen in gold systems (a few nm). High-resolution microscopy is in progress to The composition of the de-alloyed layer is confirm this point. interesting on two counts. First, it appears to be the first observation of a de-alloying process that arrests at a layer composition corresponding to the typical 50-50 ‘parting limit’ seen in bulk solid solutions such as Ag-Au. That this has not been observed before may be attributed to a coincidental value of anodic overpotential for Fe that is naturally established in the Fe-Ni system at its open- circuit potential. It should be possible to obtain an analogous layer composition by choosing a particular potential in a system such as 90Ag-10Au. At higher potentials all the Ag (or Fe) will be stripped out of the layer. The second interesting feature is that the de-alloyed layer solves a space-filling problem that was identified many years ago by Newman and Sieradzki: a Ni content of only ca. 10% is not enough to form a 3D connected network upon removal of all the other alloy components, but if half the iron is left behind, the de-alloyed layer is approximately 20% space-filling which is enough to form a strong random porous solid. In the context of the film-