Defining the passive state of Alloy 22 using electrochemical impedance spectroscopy and deterministic modeling

Storage of High Level Nuclear Waste (HLNW) requires prediction of corrosion behavior with unprecedented accuracy over thousands of years. To support this need, high-accuracy electrochemical data are required. A laboratory with special features to support long-term electrochemical tests has been designed, built, and proven. High-accuracy electrochemical impedance spectroscopy (EIS) data and passive current densities have been measured for Alloy 22 in deaerated pH 6, 4 M NaCl and pH 3, 0.1 M Cl- electrolytes at temperatures of 30°, 60°, 75°, and 90°C. Accuracy of the EIS data has been verified by checking for hysteresis of back to back scans and in data collected hundreds of hours apart, and through application of the Kramers-Kronig transforms. The data are interpreted in terms of the Point Defect Model (PDM) of growth and breakdown of passive films. Mott-Schottky analysis is applied to evaluate the semiconductor properties of the passive film. At low potential the passive film is found to be an n-type semiconductor. The film transitions to a p-type semiconductor as potential is increased towards the transpassive potential. Impedance data collected into the transpassive range show a large drop in capacitance at low frequency. This can be interpreted as thinning of the film and weakening of the lattice structure. The observed behavior can be explained in terms of the PDM reaction mechanism as cation vacancy generation by oxidative ejection of cations at the film/solution interface. Comparison of low frequency impedance data, particularly phase shift, between 4 M NaCl and 0.1 M Cl- solutions indicates that passive film properties are relatively insensitive to chloride concentration. Mott-Schottky analysis indicates carrier density is relatively insensitive to chloride concentration. The equations derived from the PDM have been programmed into a commercial curve fitting software package, which optimizes estimated parameters on the PDM to derive a "best fit" parameter set using nonlinear curve fitting. It may also be used to calculate impedance from a given set of model parameters derived from another method such as phase space analysis. A "best guess" parameter set from phase space analysis was found to provide a good match between calculated and measured impedance, current density, and film thickness. However, the curve fitting method is unconstrained and was found to generate multiple solutions for a single impedance spectrum. When multiple parameter sets that provide good matches between calculated and measured impedance data were used to calculate barrier layer thickness and passive current density, many solutions resulted in unrealistic values. This demonstrates that the curve fitting method must be constrained by including thickness and current density equations in the model. Though unconstrained, the modeled parameters provide useful insights into the PDM reaction mechanism indicating cation interstitials are the dominant defect.