SCIENCE Thin-Layer Mixed-Potential Model for the Corrosion of High-Level Nuclear Waste Canisters D.D. Macdonald and M. Urquidi-Macdonald* ABSTRACT A thin-layer mixed-potential model (TLMPM) is described for cal- culating redox and corrosion potentials for high-level nuclear waste (HLNV) canisters in a tuff repository. The model assumes that the canister surface covered by a thin liquid (water) layer, which is irradiated by y-photons from the decay of radionuclides in the waste. The dose rate and the temperature at the surface are assumed to decay exponentially with time. Redox and corro- sion potentials are calculated by equating the sum of partial cur- rents for charge-transfer reactions involving various electroactive radiolysis products and the meta/ substrate to zero. Estimated polarization data for type 304L (UNS' ) S30403) and type 316L (UNS S31603) stainless steels, for alloy 825 (UNS N08825), and for copper alloys CDA 102 (UNS C10200), CDA 613 (UNS C61300), and CDA 715 (UNS C71500) in high-temperature aque- ous solutions were combined with the TLMPM to calculate redox and corrosion potentials over a 1000-year time period. The model predicts that the corrosion potential wilt increanse with time due to the dominant effect of decreasing temperature. KEY WORDS: corrosion, high-level nuclear waste, mixed poten- tial model, water radiolysis. INTRODUCTION The disposal of high level nuclear waste (HLNW) represents a ma- jor technological challenge for all nations that operate nuclear re- actors. The principal goal of this technology is to isolate the waste forms containing various radio nuclides from the biosphere for a period in excess of 10 ° years and possibly longer. This must be done with exceptional reliability, since any break of the canister is unacceptable ecologically. The most likely cause of failure of the engineered barrier is corrosion with various forms of localized at- tack (e.g., pitting corrosion) being considered the most probable. The problem of designing an engineered barrier to contain the waste form therefore involves careful consideration of the corro- sion behavior of candidate metals and alloys in the expected envi- ronment over a millennium or more in time. However, our experi- "Submitted for publication August 1989; in revised form, December 1989. *SRI International, 333 Ravenswood Ave., Menlo Park, CA 94025. 'UNS numbers are listed in Metals and Alloys in the Unified Numbering System, published by the Society of Automotive Engineers (SAE) and cosponsored by the American Society for Testing and Materials (ASTM). ence with the corrosion of materials under the appropriate environ- mental conditions extends back no more than 100 years, so that method must be developed either to extrapolate our meager corro- sion data base over a period of 100 to 1000 times our current ex- perience or to predict the corrosion behavior using suitable physi- cal/mathematical models. The latter approach is attractive, only if the models are sufficiently sophisticated to yield reliable predic- tions. Of the various repository systems under consideration,' the proposed disposal in a tuff environment at Yucca Mountain, Ne- vada, is particularly attractive because the repository will be re- mote from any known extended groundwater system.' However, the canister will be in contact with moist air during the initial isola- tion period (up to 100 years) when the temperature of the canister surface is above the boiling temperature of water.' At longer times, the canister may also be in contact with liquid water once the temperature has decayed below the boiling point of water at the elevation of the repository (-95°C). In this paper, we describe a thin-layer mixed potential model for calculating the redox and corrosion potentials at the surface of a high-level nuclear waste canister in a tuff environment. Briefly, this model seeks to calculate the spontaneous electrical potential of the canister surface in contact with a thin liquid (water) layer that contains nonquilibrium concentrations of oxidizing (e.g., O 2 , H20 2 , OH) and reducing (H2 , e ; q ) species that are produced by the radiolysis of water. The problem is further complicated because the temperature (Figure 1) and the -y-ray dose rate at the canister surface decay with time in a nearly exponential manner, from initial values of —250°C and —6.1 x 10 5 rad/ht 2) (8.47 x 10 16 eV•cm - 3 •s - '), respectively. Because of the nonequilibrium nature of the system, a kinetic rather than a thermodynamic model is required to account for reactions between radiolytically gener- ated species in the bulk of the liquid film and between various electroactive species and the canister surface. These latter reac- tions are potential dependent, and the redox or corrosion potential is that voltage at which the total current for the oxidation and re- duction of all electroactive species at the surface (including the canister material) is equal to zero. Calculation of the corrosion po- tential has been carried out for type 304L stainless steel (SS), type 316L SS, alloy 825, copper (CDA 102), and copper alloys 121This value was kindly supplied to us by Dr. R. Glass, Livermore National Laboratory, Livermore, CA. 0010-9312/90/000115/$3.00/0 © 1990 , National Association of Corrosion Engineers CORROSION —May 1990