Experimental Considerations for Modeling Galvanic Corrosion in Aluminum and Its Alloys

Abstract

In this paper, experimental data are compared for an aluminum/copper galvanic system with model predictions from a commercially available software package. The experimental data consisted of optical profilometry measurements of damage accumulation in a bimetallic Al-Cu strip as a function of immersion time in a chloride environment. Model predictions of corrosion rate were accomplished with software that used a solution of the Laplace equation to determine the potential distribution for a mesh of the system. Current density and, ultimately, corrosion rates were determined with this model by correlating the potential at a mesh point with a current density from the experimental potentiodynamic polarization curves for the materials. Although this is a proven method for systems where the anode material is under activation control, this methodology presents a problem for the case of pitting corrosion, such as Al and its alloys in chloride solution. In the case of pitting corrosion, the surface average current density from the polarization curve (i.e., the couple current density) is not equal to the current density at the base of the pit resulting in a model prediction that underestimates the corrosion rate. To overcome this shortcoming, an alternate method for determining the polarization curve for Al is offered from artificial pit electrodes (pencil electrodes) and model predictions using that method are presented.