Electrochemistry of High Temperature Corrosion of Alloys in Molten Salts Relevant to Future Nuclear Reactors

Abstract

Eutectic molten salts are coolant candidates for molten salt-cooled nuclear reactors; however, alloy corrosion is the key materials-compatibility issue [1]. Corrosion in molten salts may involve thermodynamic considerations, thermal gradient-driven corrosion, dissimilar material corrosion, dealloying, and impurity-driven corrosion [2, 3]. New alloys must be developed, and their corrosion behaviour merits special attention at a fundamental level. We aim to understand industrial alloy behaviour through study of model alloys, leading to insights relevant to materials performance in Molten Salt Reactors.In the field of molten salt corrosion, dissolution of alloying elements is mostly discussed in terms of one-dimensional diffusion of a more easily dissolved element from the bulk to the surface, which necessarily involves lattice diffusion of metals. We report on electrochemical study of corrosion mechanisms in Fe-(Cr)-Ni model and industrial alloys, and report on a study of critical alloy compositions and porosity formation upon dealloying of one or more electrochemically reactive components in molten salts.Dealloying is selective electrolytic dissolution of one or more active elements from a metallic solid solution or intermetallic compound [4, 5]. The operative mass transport process in aqueous dealloying is diffusion of the more-noble component at the solid-electrolyte interface – enhanced by poorly-understood electrolyte effects on the diffusivity. The parting limit in dealloying is the minimum content of less-noble element(s) for dealloying, below which the dealloying is prevented by a passive layer of more-noble elements on surface formed at initial stage of corrosion. The more usual case (i.e. AgAu, CuAu, etc.) is a threshold of ca. 55-60 at. % less-noble element (Ag and Cu, respectively) [6]. According to Artymowicz et al. [7] the underlying parting limit is very close to 60 at. %, but increasing kinetics of surface diffusion could drop the parting limit to ca. 55 at. % in systems studied to that date.Fe-(Cr)-Ni model alloys were prepared using a Cold Crucible Induction Levitation Melter. Electrochemical studies were done using a well-controlled electrochemical cell for corrosion study in molten chloride salts. We have developed Mg|Mg2+ reference electrode (RE) for our eutectic chloride salts and it is found to be a reliable RE, as far as we can see from our electrochemical measurements (open circuit potential, cyclic polarization, and polarization resistance). Impurity contents were monitored electrochemically, and final water removal was carried out using Mg. Characterization is being carried out by analytical electron microscopy, X-ray diffraction and secondary ion mass spectrometry.We found that up to a certain temperature, there is dealloying of the type observed in aqueous solutions, with porosity formation and parting limits. In brief, porosity and parting limits were observed in molten salts, exactly as predicted, except that the de-alloying threshold for electrolytic dissolution of the less-noble element(s) was dropped by several percent, compared with aqueous solutions. This is due to the very fast surface diffusion of more-noble metal in the molten chloride salt. Dealloyed layers were nearly pure nickel, but with residual Fe and/or Cr at the ligament cores in the porous structure. Correspondingly, the porosity is very coarse, and shows new features such as secondary corrosion through the ligament cores, as shown in Fig. 1.More Ni suppresses dealloying in both Fe-Ni and Fe-Cr-Ni model alloys, but more Fe and Cr promote oxide formation in binary and ternary alloys where dealloying propagates below an oxide layer of Fe and Cr in binary and ternary alloys. So, to some extent there is a balance, masking the underlying dependency of dealloying on Ni content.This type of dealloying is mediated by surface diffusion. At higher temperatures, there is a shift to different mechanisms involving lattice diffusion in the metal and porosity changes its appearance. Even then, there is not necessarily a planar dealloying front – more of a “negative dendrite” type of interface would appear [8]. At very high homologous temperature the dealloying feature will revert to that mostly described in the molten salt literature, with planar interfaces.