Dealloying of Magnesium Alloys

The corrosion behavior of cast Mg-Al alloys such as AZ91D and AM60B is strongly dependent on the alloy composition and microstructure.  Alloy grain size, area fraction of phases and spatial distribution of the phases are important factors governing both full immersion and atmospheric corrosion behavior. Mg-Al alloy systems can be broadly characterized as being composed of a primary a-Mg matrix (grains) containing Al in solid solution and b-phase intermetallic particles mainly distributed within and adjacent to grain boundaries.  Heat treatment of the as-cast alloys significantly modifies the microstructure altering grain size, phase amounts and distribution as well as the a-Mg phase composition.

The general scenario that we envision evolving during corrosion of cast Mg-Al alloys is the following. In near-neutral electrolytes containing chloride, second phase particles such as the Mg₁₇Al₁₂ b-phase serve as cathodes for water reduction resulting in increases in pH.  The corresponding dissolution reaction involves the dealloying of Mg from a-matrix grains resulting in enrichment and redistribution of solid solution Al as well as impurities (such as Fe) that may be present within the grains.  The redistributed Al will form Al clusters on the surface of the now dealloyed a-phase that can also serve as cathodes. If the pH rises to ~ 9-9.5, the a-phase will begin to passivate and the b-phase can in principle dealloy (selective dissolution of Al) since the corrosion potential (~1.5V vs. SCE) is high enough for this to occur.  At this elevated pH, aluminum may dissolve as the soluble aluminate and will tend to precipitate on to the dissolving alloy surface. Thus, owing to these pH changes local cathodes can become anodes and vice versa. The degree to which these processes affect corrosion behaviors depends on the spatial density (mean separation) of existing and evolving surface cathodes and how this length scale compares to the thickness of a electrolyte diffusion boundary layer or condensed liquid layer (atmospheric corrosion) and grain size.

Given this hypothesis for corrosion behavior, we have been examining dealloying of AZ91D and AM60B as well the constituent a- and b-phases. Instead of relying on a technique such as EDS to characterize the time-dependent change in Al concentration on a corroding surface we developed an electrochemical assay involving Li underpotential deposition (UPD) that measures the electrochemically active surface area (ECSA) of Al. We show results comparing the new assay protocol yielding the ECSA of Al to that obtained from EDS. Anodic dissolution behaviors of the alloys and constituent phases have been measured in an ionic liquid electrolyte (neat1:2 molar ratio of choline chloride:urea) which allows for accurate determination of the evolving surface composition as a function of well-defined anodic dealloying rates. We discuss these results together with some preliminary Kinetic Monte Carlo modeling of corrosion processes of the constituent a-phase.