in J-13 and in solutions at pH 2, 4, and 10 were low and similar to those of SS-316 and Alloy C-22. Corrosion rate data for MWF materials were also compared to those for copper and mild steel. These results are not surprising, considering that the solutions tested did not contain chloride ions that could have initiated localized corrosion.

The list of tests proposed to be performed after June 1999 had not been finalized completely. Electrochemical tests are to be performed at elevated temperatures in order to assess the effect of increased temperature on corrosion rates. Given that electrochemical processes proceed at higher rates at higher temperatures, it might be better to concentrate on a few key samples, expose them at higher temperatures, and obtain electrochemical and surface analysis data. Tests were also to be conducted in chloride solutions at concentrations up to 10,000 ppm. These are credible conditions that might be encountered in a repository. In this case it might be better to carry out pitting scans according to ASTM G-61, which allows a comparison of different alloys' relative susceptibility to pitting. Data on corrosion rate can also be obtained from these measurements. Finally, a study of corrosion mechanisms was proposed by ANL.

Metal Waste Form Release Rate

Work was presented to the committee at its meeting at ANL-E on the development of a radioisotope release rate model for the stainless steel MWF.7 Because the MWF behaves much like stainless steel, the modeling approach is based on the results of ANL's current uniform corrosion test program, the use of known stainless steel degradation mechanisms as a basis of MWF release modeling, and the use of MWF corrosion rate data to adjust the models empirically. Factors considered important for radioisotope release include MWF metallurgy, degradation mechanisms, and environmental conditions. A number of potential degradation mechanisms were discussed, including uniform corrosion, localized corrosion (crevice, pitting), microbially influenced corrosion (MIC), selective leaching, intergranular attack, and galvanic corrosion. 8 (Models will also be established for the chemistry of the repository water.)

Waste canister modeling is concerned with the corrosive attack of the carbon steel and C-22 shells. Corrosion testing has used concentrated J-13 well water. In discussions between representatives of ANL and the committee, it was suggested that it is highly unlikely that any corrosive attack of C-22 would be observed under the mild conditions used in the corrosion testing. Corrosion models for long-term waste canisters will be based on empirical multivariate regressions. Key variables considered in ANL's corrosion testing included temperature, pH, and chloride content. For stainless steel, additional variables such as H2O2, HCO3-, and NO3- concentrations will be taken into account. The experimental data to be used for the development of models for uniform corrosion were those discussed in ANL's presentation at the committee's meeting at ANL-E in October 1998. It was mentioned that additional tests were scheduled to reduce model uncertainties, although it was not clear what types of tests would be used. In


M. C. Petri, ANL-E, presentation to the committee, Argonne, IL, October 26, 1998.


In discussions between representatives of ANL and the committee at the committee's meeting at ANL-W in June 1998, a member of the committee suggested that it was unlikely that MIC would play an important role and that obtaining meaningful MIC data would probably be quite difficult.

The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement