IntroductionWorldwide, there are three high-level nuclear waste repositories under construction. Both the Forsmark granite project in Sweden and the Olkiluoto granite project in Finland are in an advanced stage of licensing. These two projects (SKB 2006; POSIVA 2006) and the former Columbia River Basalt (CRB) project in the USA (Lutton et al. 1986) envisage the storage of waste in copper-shielded steel containers surrounded by a bentonite buffer. The Opalinus Clay project in Switzerland has the option of both copper-shielded containers and unshielded containers (HSK 2005). The Bure clay project in France is in an early stage of licensing. The design of this project is exceptional in that it exclusively considers unshielded steel containers implying very high corrosion rates (ANDRA 2016). However, the final design may change in the course of the licensing procedure.Copper-shielded waste containers are the optimal solution, and there is hardly any room for improvement in the general container design and emplacement method (KBS-3V disposal concept; SKB 2006; POSIVA 2006). However, the search for the optimal mineralogical composition of the bentonite buffer surrounding the container (NDA 2014; POSIVA 2017) has not been completed yet. This is shown by the Forsmark project, which is one of the two projects in the advanced stage of licensing. In a nutshell, the applicant (SKB) claims that sulfur was not the main corrosion agent for copper in a 20-year corrosion experiment with copper embedded in MX-80 bentonite, which has a relatively high calcium sulfate content (1.3%; Carlson 2004). The licensing authority SSM has considerable doubts in this respect and has been looking for advice from consultants even up to early 2021 (MKG 2021).Prior to 2006, MX-80 bentonite had been the only reference buffer material in the KBS-3V disposal concept (NDA 2014). In 2006, SKB initiated the Alternative Buffer Material project with 10 additional materials (Eng et al. 2007; Svensson et al. 2011); these materials have a calcium sulfate content above and below that of the MX-80 bentonite.The mineralogy exerts a nonnegligible influence on the corrosion rate of the copper shield, especially with respect to sulfate mineral concentration. The objective of this paper is the quantification of the buffer-dependent corrosion rates of the copper shield.MethodThe structured orthogonal model mesh contains 36 elements and has dimensions of 1 × 1 × 0.36 m (Table 2). The mesh represents the standard design for nuclear waste disposal (SKB 2006; POSIVA 2006) with the host rock (first node) and the 35-cm-thick bentonite buffer (intermediate nodes) placed adjacently to the 5-cm-thick copper shield (last node). Except for the first node, the cell volume of the individual elements is 0.01 m3. The first element has a volume of 1052 m3, which imposes Dirichlet conditions, that is, the thermodynamic conditions of this element do not change at all.Table 2. Model setupParameterDimensionValueSimulation perioda106Model length/width/heightcm36/100/100Temperature°C25PressureMPa0.1Porosity—0.4Permeabilitym210−20Diffusion coefficient (all species)m2/s7 × 10−12The transport conditions are simulated with the thermodynamic data of the EQ3/6 V8b.2 database (TherAkin10.dat; Wolery 1992), which were derived using SUPCRT92 (Johnson et al. 1992), supplemented with the thermodynamic data for the solid-phase lepidocrocite (Majzlan et al. 2003a, b) and the liquid species CuCl° and CuCl− (Xiao et al. 1998). All liquid species in the database are allowed to take part in the reactions, the total comprising 128 liquid species. The solid phases react under kinetic constraints, with precipitation rates equal to dissolution rates. The rate constants are taken from Palandri and Kharaka (2004), except for native copper and chalcocite (Table 3). The rate constant for native copper (4.195 × 10−10 mol m−2 s−1) is calculated from a general corrosion rate of 9.4 × 10−7 m/a (King 2009). The rate constant for chalcocite is set equal to the rate constant for native copper.Table 3. TOUGHREACT flow-transport model setup: initial volume fractions, reactive surface areas, and kinetic propertiesTable 3. TOUGHREACT flow-transport model setup: initial volume fractions, reactive surface areas, and kinetic propertiesMineralChemical composition
Initial volume fractionReactive surface area (m2/m3)k25 (mol/m2/s)bEa (kJ/mol)bn(H+)bk25 (mol/m2/s)Ea (kJ/mol)k25 (mol/m2/s)Ea (kJ/mol)n(H+)CalciteCaCO30.03,2005.01 × 10−114.41.001.55 × 10−623.53.31 × 10−435.41.00QuartzSiO20.02,800———1.03 × 10−1487.7———GypsumCaSO4 · 2H2O0.0–0.0073,200———1.62 × 10−303.89 × 10−1548.0−0.13AnhydriteCaSO40.0–0.0073,200———6.46 × 10−414.3———HematiteFe2O30.0–0.0074,6004.08 × 10−1062.21.002.52 × 10−1566.2———GoethiteFeOOH0.0–0.0074,600———1.50 × 10−886.5———LepidocrociteFeOOH0.0–0.0074,600———1.50 × 10−886.5———PyriteFeS20.05,4003.02 × 10−856.9−0.52.82 × 10−1056.9———Native copperCu0.0–0.9540———4.195 × 10−10————ChalcociteCu2S0.0540———4.195 × 10−10————Na-smectiteNa0.29Al1.77Mg0.26Si4O10(OH)20.0–0.924,0001.05 × 10−1123.60.341.66 × 10−1335.03.02 × 10−1758.9−0.40K-smectiteK0.29Al1.77Mg0.26Si4O10(OH)20.024,0001.05 × 10−1123.60.341.66 × 10−1335.03.02 × 10−1758.9−0.40Ca-smectiteCa0.145Al1.77Mg0.26Si4O10(OH)20.024,0001.05 × 10−1123.60.341.66 × 10−1335.03.02 × 10−1758.9−0.40The reactive surface area of smectite is that of Friedland bentonite (Karnland 2010). The remaining values are taken from a TOUHGHREACT test case (Xu et al. 2014), whereby the value for anhydrite is used for calcite and gypsum, the value for goethite is used for hematite and lepidocrocite, and native copper and chalcocite use the surface area of pyrite multiplied by 0.1. All minerals are allowed to precipitate and dissolve.In the first cell, which is subjected to Dirichlet conditions, the initial volume fraction is zero for all minerals. The intermediate cells (No. 2-35) representing the bentonite buffer initially have 90% Na-smectite or 90% Na-smectite and accessory minerals, depending on the model setup according to the composition of common candidate buffer materials (Table 4):
•0.7% hematite (Fe2O3),•0.7% goethite (α-FeOOH),•0.7% lepidocrocite (γ-FeOOH),•0.7% gypsum (CaSO4 · 2H2O),•0.7% anhydrite (CaSO4), and•the combination of an iron oxide and calcium sulfate, for example, 0.7% hematite + 0.7% gypsum.Table 4. Mineralogical and chemical characterization of bentonites, which are candidate buffer materials for nuclear waste disposalTable 4. Mineralogical and chemical characterization of bentonites, which are candidate buffer materials for nuclear waste disposalBentoniteCountryChemical analysis S (%)ReferenceCalcium sulfate (gypsum) (anhydrite) (%)Hematite (%)Fe oxyhydroxide (goethite) (lepidocrocite) fcc (%)StranceCzech Republic——10–15<0.1Carlson (2004)MX-80USA1.3——0.31Carlson (2004)BH-200USA2.9——0.69Carlson (2004)FriedlandGermany0.1–1.20–0.70.1–1.40.5Karnland et al. (2006)KutchIndia0.3–2.20–0.81.1–1.60–0.1Karnland et al. (2006)AvonsealCanada2——0.4Quigley (1984)AvongelCanada3——0.6Quigley (1984)FiltaclayCanada3——0.6Quigley (1984)DeponitGreece1.8——SKB (2006)The remaining minerals are native copper, chalcocite, pyrite, Ca-smectite, K-smectite, calcite, and quartz, which all have an initial volume of zero.The last cell (No. 36) representing the copper shield initially has 90% native copper and 3% Na-smectite with additional components equal to those of the intermediate cells (No. 2-35). The initial fluid composition in all cells is that of each individual reference groundwater. The liquid flow is zero between all cells. The transport of liquid species occurs by diffusion only. The TOUGHREACT code calculates reactive transport with a single global diffusion coefficient. Because HS− is the most important corrosion rate-determining species, the global diffusion coefficient is set equal to that of HS− (7 × 10−12 m2 s−1; King et al. 2002). The simulation period is 1 million years.ResultsAll simulations showed a satisfactory convergence behavior, except for two sets of simulations with accessory anhydrite (Forsmark and Olkiluoto groundwater). In the remaining 47 models, the general corrosion depth of the 5-cm-thick native copper (Cu°) shield at the end of the 1-million-year simulation period is in the range of 0.002–0.08 cm (Fig. 1).The groundwater composition has a relatively strong influence on the corrosion depth, whereby a groundwater composition corresponding to a relatively short distance from the Cu°–chalcocite (Cu2S) boundary in the Eh–pH space of the Cu–S–Cl–O–H system [Fig. 2(a)] allows relatively little corrosion. The minimum corrosion depth (0.002 cm) is recorded for the CRB test case, which represents a groundwater composition with a relatively small distance from the Cu°–Cu2S boundary. The maximum corrosion depth (0.08 cm) is recorded for the Forsmark test case, which represents a relatively large distance from the Cu°–Cu2S boundary. The remaining three test cases occupy an intermediate position both in terms of the Cu° corrosion depth and in terms of distance from the Cu°–Cu2S boundary.The influence of bentonite composition is weaker than that of groundwater composition but, nevertheless, significant. The calcium sulfate-bearing buffer has a corrosion depth that is a maximum eight times greater than that recorded for a buffer without calcium sulfates.DiscussionThe general corrosion depth of native copper (Cu°) that is determined with the reactive-transport model in this paper only considers the low-Eh side of the Cu° stability field (Fig. 2). Nevertheless, the high-Eh side may be relevant for groundwater with an especially low Cu/Cl activity ratio. The five reference groundwaters used in this study are shallow groundwaters with a relatively high Cu/Cl activity ratio. However, deeper groundwaters that have equivalent Cu concentrations but inevitably have a higher salinity (thus, lower Cu/Cl activity ratios) may be under consideration in other disposal projects. Fortunately, there is a simple technical solution that alleviates the problem of low Cu/Cl groundwaters. Wang and Hadgu (2020) propose to embed 1% copper wires or meshes in the bentonite buffer albeit for a different reason (heat dissipation). The side benefit of such a design will be an enhanced Cu/Cl activity ratio in the interstitial water of the bentonite buffer, because some copper of the wires or mesh dissolves; thus, the Cu/Cl activity ratio of the fluid reaching the copper shell of the waste container is relatively high. Note that low Cu/Cl groundwater, which will justify the modified design for purely chemical reasons, is not modeled in this paper, because there are no real-world examples available that can show that the high-Eh side of the Cu° stability field is critical.On the low-Eh side, the Cu° corrosion depth that is determined with the reactive-transport model in this paper tends to be lower than that of previous nonreactive transport models. Previous calculations predict a general corrosion depth of 0.04–0.1 cm at the end of a 1-million-year simulation period for Scandinavian and North American sites (Schwartz 2008, 2018, 2019), whereas the new values range from 0.002 to 0.08 cm. The general corrosion depth has to be multiplied by a correction factor to obtain the actual value for pit corrosion. King et al. (2002) estimate that the correction factor is in the range of 5 (realistic value) to 25 (conservative value). Accordingly, the minimum pit corrosion depth should be in the range of 0.01–2 cm. Other types of localized corrosion are stress corrosion, hydrogen embrittlement, and radiation-induced corrosion (Nacka District Court 2018; KTH 2020). These other types of localized corrosion will partly add to pit corrosion.References ANDRA (Agence nationale pour la gestion des déchets radioactives). 2016. Dossier d’options de sûreté. Châtenay-Malabry, France: ANDRA. Butler, B. S., and W. S. Burbank. 1929. The copper deposits of Michigan, 1–234. US Geological Survey Prof. Paper No. 144. Reston, VA: USGS. Carlson, L. 2004. Bentonite mineralogy. POSIVA Working Rep. No. 2004-02. Olkiluoto, Finland: Posiva OY. Eng, A., U. Nilsson, and D. Svensson. 2007. Äspö hard rock laboratory alternative buffer material. SKB Rep. No. IPR-07-15. Stockholm, Sweden: SKB (Svensk Kärnbränslehantering). HSK. 2005. Gutachten zum Entsorgungsnachweis der Nagra für abgebrannte Brennelemente, verglaste hochaktive sowie langlebige mittelaktive Abfälle (Projekt Opalinuston). Würenlingen, Switzerland: HSK (Hauptabteiliung für die Sicherheit der Kernanlagen). Johnson, J. W., E. H. Oelkers, and H. C. Helgeson. 1992. “SUPCRT92: A software package for calculating the standard molal thermodynamic properties of minerals, gases, aqueous species, and reactions from 1 to 5000 bar and 0 to 1000°C.” Comput. Geosci. 18 (7): 899–947. https://doi.org/10.1016/0098-3004(92)90029-Q. Karnland, O. 2010. Chemical and mineralogical characterization of the bentonite buffer for the acceptance control procedure in a KBS-3 repository. SKB Technical Rep. No. TR-10-60. Stockholm, Sweden: SKB (Svensk Kärnbränslehantering). Karnland, O., S. Olsson, and U. Nilsson. 2006. Mineralogy and sealing properties of various bentonites and smectite-rich clay minerals. SKB Technical Rep. No. TR-06-30. Stockholm, Sweden: SKB (Svensk Kärnbränslehantering). King, F. 2009. Critical review of the literature on the corrosion of copper by water. Nanaimo, BC, Canada: Integrity Corrosion Consulting. King, F., L. Ahonen, C. Taxen, U. Vuorinen, and L. Werme. 2002. Copper corrosion under expected conditions in a deep geologic repository. POSIVA Rep. No. 2002-01. Olkiluoto, Finland: Posiva OY. KTH (Kungliga tekniska högskolan). 2020. “The most important comments to the SKB LOT-report TR-20-14.” Accessed April 23, 2021. http://www.mkg.se. Lasaga, A. C., J. M. Soler, J. Ganor, T. E. Burch, and K. L. Nagy. 1994. “Chemical weathering rate laws and global geochemical cycles.” Geochim. Cosmochim. Acta 58 (10): 2361–2386. https://doi.org/10.1016/0016-7037(94)90016-7. Lutton, J. M., W. F. Brehm, H. P. Maffei, C. L. Rivera, and R. P. Anantatmula. 1986. “General corrosion studies of candidate container materials for the Basalt Waste Isolation project.” In High-level nuclear waste disposal, edited by H. C. Burkholder, 563–572. Richland, WA: Batelle Press. Majzlan, J., K. Grevel, and A. Navrotzky. 2003a. “Thermodynamics of Fe oxides: Part II. Enthalpies of formation and relative stability of goethite (α-FeOOH), lepidocrocite (γ-FeOOH) and maghemite (γ-Fe2O3).” Am. Mineral. 88 (5–6): 855–859. https://doi.org/10.2138/am-2003-5-614. Majzlan, J., B. E. Lang, R. Stevens, A. Navrotsky, B. F. Woodfield, and J. Boerio-Goates. 2003b. “Thermodynamics of Fe oxides: Part I. Entropy at standard temperature and pressure and heat capacity of goethite (α-FeOOH), lepidocrocite (γ-FeOOH) and maghemite (γ-Fe2O3).” Am. Mineral. 88 (5–6): 846–854. https://doi.org/10.2138/am-2003-5-613. Metz, V., B. Kienzler, and W. Schüßler. 2003. “Geochemical evaluation of different groundwater-host rock systems for radioactive waste disposal.” J. Contam. Hydrol. 61 (1–4): 265–279. https://doi.org/10.1016/S0169-7722(02)00130-4. MKG (Miljöorganisationernas kärnavfallsgranskning). 2021. “The regulator SSM expected to report LOT results to the government in the beginning of March—MGK has made several inputs.” Accessed April 23, 2021. http://www.mkg.se. Nacka District Court. 2018. “Summary of opinion regarding the proposed final repository for spent nuclear fuel at Forsmark.” [ Unofficial translation by MKG] Accessed April 23, 2021. http://www.mkg.se. NDA (Nuclear Decommissioning Authority). 2014. A review of the development of bentonite barriers in the KBS-3V disposal concept. NDA Technical Note No. 21665941. Oxford, UK: NDA. Newcomb, R. G. 1972. Quality of the ground water in basalt of the Columbia River Group, Washington, Oregon and Idaho. US Geological Survey Water-Supply Paper No. 1999-N. Reston, VA: USGS. Palandri, J. L., and Y. K. Kharaka. 2004. A compilation of rate parameters of water-mineral interaction kinetics for application to geochemical modeling. US Geological Survey Open File Rep. No. 2004-1068. Menlo Park, CA: USGS. Pearson, F. J. 2002. Benken reference water chemistry. NAGRA Rep., quoted in NAGRA, 2002. Project Opalinus Clay—Safety Rep. NAGRA Technical Rep. No. 02-05. Wettingen, Switzerland: NAGRA. Pearson, F. J. 2003. Mont Terri project—Geochemistry of water in the Opalinus Clay formation at the Mont Terri rock laboratory. Serie Geologie 5: 1–319. Eidgenössisches Dept. für Umwelt. Bern, Switzerland: Berichte des BWG. Pitkänen, P., S. Partamies, and A. Luukkonen. 2004. Hydrochemical interpretation of baseline groundwater conditions at the Olkiluoto site. POSIVA Rep. No. 2003-07. Olkiluoto, Finland: Posiva OY. POSIVA. 2006. Expected evolution of a spent nuclear fuel repository at Olkiluoto. POSIVA Rep. No. 2006-05. Olkiluoto, Finland: POSIVA. POSIVA. 2017. Safety functions, performance targets and technical design requirements for a KBS-3V repository. POSIVA SKB Rep. No. 01. Olkiluoto, Finland: Posiva OY. Pruess, K., C. Oldenburg, and G. Moridis. 1999. TOUGH2 user’s guide, version 2.0. Berkeley, CA: Earth Sciences Division, Lawrence Berkeley National Laboratory, Univ. of California. Quigley, R. M. 1984. Quantitative mineralogy and preliminary pore-water chemistry of candidate buffer and backfill materials for a nuclear fuel waste disposal vault, 1–41. Rep. No. AECL 7827. Chalk River, ON, Canada: Atomic Energy of Canada. Reidel, S. P., V. G. Johnson, and F. A. Spane. 2002. Natural gas storage in basalt aquifers of the Columbia Basin, Pacific Northwest USA: A guide to site characterization. PNNL Rep. No. 13962. Richland, WA: Pacific Northwest National Laboratory. Schwartz, M. O. 2018. “The new Wallula CO2 project may revive the old Columbia River Basalt (western USA) nuclear-waste repository project.” Hydrogeol. J. 26 (1): 3–6. https://doi.org/10.1007/s10040-017-1632-y. Schwartz, M. O. 2019. Nuclear waste disposal—Groundwater contamination in three dimensions. Beau Bassin, Mauritius: Scholar’s Press. SKB. 2006. Long-term safety for KBS-3 repositories at Forsmark and Laxemar—A first evaluation. SKB Technical Rep. No. TR-06-09. Stockholm, Sweden: SKB (Svensk Kärnbränslehantering). Smellie, J. A. T., M. Laaksorharju, and P. Wikberg. 1995. “Äspö, SE Sweden: A natural groundwater flow model derived from hydrogeochemical observations.” J. Hydrol. 172 (1–4): 147–169. https://doi.org/10.1016/0022-1694(95)02720-A. Svensson, D., A. Dueck, U. Nilsson, S. Olsson, T. Sandén, S. Lydmark, S. Jägerwall, K. Pedersen, and S. Hansen. 2011. Alternative buffer material. SKB Rep. No. TR-11-06. Stockholm, Sweden: SKB (Svensk Kärnbränslehantering). Wang, Y., and T. Hadgu. 2020. “Enhancement of thermal conductivity of bentonite buffer materials with copper wires/meshes for high-level radioactive waste disposal.” Nucl. Technol. 206 (10): 1584–1592. https://doi.org/10.1080/00295450.2019.1704577. Wolery, T. W. 1992. EQ3/6, a software package for geochemical modeling of aqueous systems: Package overview and installation guide (version 7.0). Livermore, CA: Lawrence Livermore National Laboratory, Univ. of California. Xiao, X., C. H. Gammons, and A. E. Williams-Jones. 1998. “Experimental study of copper(I) chloride complexing in hydrothermal solutions at 40 to 300°C and saturated vapor pressure.” Geochim. Cosmochim. Acta 62 (17): 2949–2964. https://doi.org/10.1016/S0016-7037(98)00228-2. Xu, T., E. Sonnenthal, N. Spycher, and L. Zheng. 2014. TOUGHREACT v3.0-OMP reference manual: A parallel simulation program for non-isothermal multiphase geochemical reactive transport. Berkeley, CA: Earth Sciences Division, Lawrence Berkeley National Laboratory.