Speciation and Solubility Calculations for Waste Relevant Radionuclides in Boom Clay

EXTERNAL REPORT SCK•CEN-ER-198 14/Ssa/P-16

Speciation and solubility calculations for waste relevant radionuclides in Boom Clay

First Full Draft

Sonia Salah and Lian Wang

SCK•CEN Contract: CO-90-08-2214-00, RP.W&D.0064 NIRAS/ONDRAF contract: CCHO- 2009-0940000, LTBC02-GEO-01 Radionuclide migration and retention processes in Boom Clay

April, 2014

SCK•CEN RDD Boeretang 200 BE-2400 Mol Belgium

EXTERNAL REPORT OF THE BELGIAN NUCLEAR RESEARCH CENTRE SCK•CEN-ER-198 14/Ssa/P-16

Speciation and solubility calculations for waste relevant radionuclides in Boom Clay

First Full Draft

Sonia Salah and Lian Wang

SCK•CEN Contract: CO-90-08-2214-00, RP.W&D.0064 NIRAS/ONDRAF contract: CCHO- 2009-0940000, LTBC02-GEO-01 Radionuclide migration and retention processes in Boom Clay

April, 2014 Status: Unclassified ISSN 1782-2335

SCK•CEN Boeretang 200 BE-2400 Mol Belgium

Phone +32 14 33 21 11 Fax +32 14 31 50 21 http://www.sckcen.be

Contact: Knowledge Centre [email protected]

All property rights and copyright are reserved to SCK•CEN. In case of a contractual arrangement with SCK•CEN, the use of this information by a Third Party, or for any purpose other than for which it is intended on the basis of the contract, is not authorized. With respect to any unauthorized use, SCK•CEN makes no representation or warranty, expressed or implied, and assumes no liability as to the completeness, accuracy or usefulness of the information contained in this document, or that its use may not infringe privately owned rights. SCK•CEN, Studiecentrum voor Kernenergie/Centre d'Etude de l'Energie Nucléaire Stichting van Openbaar Nut – Fondation d'Utilité Publique ‐ Foundation of Public Utility Registered Office: Avenue Herrmann Debroux 40 – BE‐1160 BRUSSEL Operational Office: Boeretang 200 – BE‐2400 MOL

Abstract

The Boom Clay formation represents the reference host rock for the geological disposal of high- level and/or long-lived radioactive waste (HLW-LL) in Belgium. The Belgian authority responsible for the long-term management of radioactive waste is ONDRAF/NIRAS (Organisme National des Déchets Radioactifs et Matières Fissiles Enrichies/Nationale Instelling voor Radioactief Afval en Verrijkte Splijstoffen). The current focus of the research programme of ONDRAF/NIRAS (O/N) is the so-called Safety and Feasibility Case 1 (SFC-1), which represents a body of sound arguments and evidences describing, quantifying and substantiating that geological disposal of high-level and/or long-lived radioactive waste is a safe and feasible long- term solution.

In order to calculate the release and transport of the waste relevant radionuclides (RNs) through the near-field of the repository and the host clay layer (far-field), different parameters, such as concentration/solubility limits, retardation factors, diffusion accessible porosity values and pore diffusion coefficients are needed.

In this report, the solubility assessment methodology, i.e. results of the speciation calculations and derivation of the RN concentration/solubility limits applicable for "undisturbed" Boom Clay conditions (far-field) will be presented.

The solubility and speciation modelling was performed with the geochemical computer code The Geochemist's Workbench (versions 8.08, 8.10 and 8.12). The reference thermodynamic database that was used for the calculations and also developed at SCK•CEN is named MOLDATA thermodynamic database (2010_MOLDATA_nov_b.dat; MOLDATA TDB, version 1).

Table of Contents ABSTRACT ...... 5 TABLE OF CONTENTS ...... 6 LIST OF FIGURES ...... 8 LIST OF TABLES ...... 9 STRUCTURE OF THE DOCUMENT ...... 10 1 INTRODUCTION ...... 11 2 SYSTEM DEFINITION ...... 12

2.1 THE DISPOSAL SYSTEM ...... 12 2.2 BOOM CLAY REFERENCE PORE WATER ...... 12 2.3 COMPUTER CODE AND DATABASES ...... 13 2.4 MOLDATA ...... 13 2.5 GDP ...... 14 3 SPECIATION CALCULATIONS AND POURBAIX DIAGRAMS ...... 16

3.1 SPECIES ACTIVITIES AND ACTIVITY COEFFICIENTS ...... 17 4 SOLUBILITY ‐ DEFINITION AND THEORETICAL BACKGROUND ...... 19

4.1 SOLUBILITY CONSTANTS...... 19 4.2 SOLUBILITY AND SATURATION ...... 20 4.2.1 Calculation procedure ...... 21 4.2.2 Reasoning and approach of solid phase selection ...... 21 4.2.3 Solubility Source and Expert ranges ...... 23 4.2.4 Uncertainties ...... 23 4.3 PARAMETERS INFLUENCING SOLUBILITY ...... 25 4.3.1 Influence of particle size ...... 25 4.3.2 Ostwald rule and ripening ...... 25 4.3.3 Influence of temperature and pressure ...... 26 4.3.4 Influence of ionic strength ...... 26 4.3.5 Influence of inorganic complexation ...... 26 4.3.6 Dissolved Organic Carbon in BC porewater ...... 27 4.3.7 Influence of organic complexation ...... 28 4.3.8 Influence of colloids ...... 28 4.3.9 Influence of Eh, pH and pCO2 ...... 30 4.3.10 Influence of radiation ...... 30 4.4 THE ROLE OF SOLID SOLUTIONS ...... 31 4.4.1 Drawbacks for the application of solid solution models ...... 32 5 RESULTS ...... 33

5.1 ACTINIUM (AC) ...... 35 5.2 AMERICIUM (AM) ...... 36 5.3 BERYLLIUM (BE) ...... 41 5.4 CALCIUM (CA) ...... 43 5.5 CAESIUM (CS) ...... 45 5.6 CARBON (C) ...... 46 5.7 CHORINE (CL) ...... 48 5.8 CURIUM (CM) ...... 49 5.9 IODINE (I) ...... 50 5.10 MOLYBDENUM (MO) ...... 53 5.11 NEPTUNIUM (NP) ...... 57 5.12 NICKEL (NI) ...... 61 5.13 NIOBIUM (NB) ...... 66 5.14 PALLADIUM (PD) ...... 69

5.15 PLUTONIUM (PU) ...... 73 5.15.1 The redox behaviour of tri‐ and tetravalent (Np and) Pu ...... 78 5.16 PROTACTINIUM (PA) ...... 80 5.17 RADIUM (RA) ...... 83 5.18 RUBIDIUM (RB) ...... 87 5.19 SAMARIUM (SM) ...... 88 5.20 SELENIUM (SE) ...... 92 5.21 SILVER (AG) ...... 97 5.22 STRONTIUM (SR) ...... 102 5.23 TECHNETIUM (TC) ...... 104 5.24 THORIUM (TH) ...... 108 5.25 TIN (SN) ...... 113 5.26 URANIUM (U) ...... 117 5.27 ZIRCONIUM (ZR) ...... 122 6 ANNEX I: SUMMARY OF SOLUBILITY CALCULATIONS FOR PHASES COMPRISED IN THE SR & ER RANGES .... 126 7 ANNEX II: SUMMARY OF SOURCE AND EXPERT RANGES* ...... 128 8CALCULATION ANNEX III: OF THERMODYNAMIC UNCERTAINTIES...... 130 9 ANNEX IV: TECHNICAL NOTE ON TH‐DATA, I.E. BINARY TH‐CARBONATE AND TERNARY TH‐HYDROXO‐ CARBONATE COMPLEXES ...... 140 10 REFERENCES ...... 146

List of Figures Figure 1: Sulphur and carbon speciation as function of Eh and pH ...... 17 Figure 2: Eh-pH diagram of americium (Am-C-S-O-H) for the BC reference porewater system...... 37 Figure 3: Eh-pH diagram of beryllium (Be-C-S-O-H) for the BC reference porewater system...... 41 Figure 4: Eh-pH diagram of calcium (Ca-C-S-O-H) for the BC reference porewater system...... 43 Figure 5: Eh-pH diagram of caesium (Cs-C-S-O-H) for the BC reference porewater system...... 45 Figure 6: Eh-pH diagram of carbon (C-S-O-H) for the BC reference porewater system...... 46 Figure 7: Eh-pH diagram of chlorine (Cl-C-S-O-H) for the BC reference porewater system...... 48 Figure 8: Eh-pH diagram of iodine (I-C-S-O-H) for the BC reference porewater system...... 51 Figure 9: Eh-pH diagram of molybdenum (Mo-C-S-O-H) for the BC reference porewater system...... 54 Figure 10: Eh-pH diagram of neptunium (Np-C-S-O-H) for the BC reference porewater system...... 58 Figure 11: Eh-pH diagram of nickel (Ni-C-S-O-H) for the BC reference porewater system...... 62 Figure 12: Eh-pH diagram of niobium (Nb-C-S-O-H) for the BC reference porewater system...... 66 Figure 13: Eh-pH diagram of palladium (Pd-C-S-O-H) for the BC reference porewater system...... 70 Figure 14: Eh-pH diagram of plutonium (Pu-C-S-O-H) for the BC reference porewater system...... 74 Figure 15: Comparison of Np(III)/Np(IV) and Pu(III)/Pu(IV) speciation...... 78 Figure 16: Eh-pH diagram of protactinium (Pa-C-S-O-H) for the BC reference porewater system...... 80 Figure 17: Eh-pH diagram of radium (Ra-C-S-O-H) for the BC reference porewater system...... 83 Figure 18: Eh-pH diagram of rubidium (Rb-C-S-O-H) for the BC reference porewater system...... 87 Figure 19: Eh-pH diagram of samarium (Sm-C-S-O-H) for the BC reference porewater system...... 89 Figure 20: Eh-pH diagram of selenium (Se-C-S-O-H) for the BC reference porewater system...... 93 Figure 21: Eh-pH diagram of silver (Ag-C-S-O-H) for the BC reference porewater system...... 98 Figure 22: Silver speciation as function of total dissolved Ag concentration ...... 100 Figure 23: Eh-pH diagram of strontium (Sr-C-S-O-H) for the BC reference porewater system...... 102 Figure 24: Eh-pH diagram of technetium (Tc-C-S-O-H) for the BC reference porewater system...... 105 Figure 25: Eh-pH diagram of thorium (Th-C-S-O-H) for the BC reference porewater system...... 109 Figure 26: Eh-pH diagram of tin (Sn-C-S-O-H) for the BC reference porewater system...... 114 Figure 27: Eh-pH diagram of uranium (U-C-S-O-H) for the BC reference porewater system...... 119 Figure 28: Eh-pH diagram of zirconium (Zr-C-S-O-H) for the BC reference porewater system...... 123 Figure 29: Eh-pH diagram of thorium (Th-C-S-O-H) for the BC reference porewater system. Assumed activity of dissolved [Th] = 10-8. Database: MOLDATA. Code: The Geochemist's Workbench - 8.10 ...... 140

List of Tables Table 1: The (recalculated) Boom Clay water composition at 25°C ...... 13 Table 2: Solubility of Am in the BC reference porewater system at 25°C, pH 8.355 and Eh -281 mV...... 38 Table 3: Species distribution of Am in equilibrium with AmCO3OH(am,hyd)...... 38 Table 4: Solubility of Be in the BC reference porewater system at 25 °C, pH 8.355 and Eh -281 mV...... 42 Table 5: Solubility of Ca in the BC reference porewater system at 25 °C, pH 8.355 and Eh -281 mV...... 43 Table 6: Solubility of Mo in the BC reference porewater system at pH 8.355 and Eh -281 mV...... 54 Table 7: Solubility of Np in the BC reference porewater system at 25°C, pH 8.355 and Eh -281 mV...... 59 Table 8: Solubility of Ni in the BC reference porewater system at 25 °C, pH 8.355 and Eh -281 mV...... 63 Table 9: Species distribution in equilibrium with Ni(OH)2(beta)...... 64 Table 10: Solubility of Nb in the BC reference porewater system at 25 °C, pH 8.355 and Eh -281 mV...... 67 Table 11: Species distribution of Nb in equilibrium with Nb2O5(cr)...... 67 Table 12: Solubility of Pd in the BC reference porewater system at 25 °C, pH 8.355 and Eh -281 mV...... 71 Table 13: Solubility of Pu in the BC reference porewater system at 25°C, pH 8.355 and Eh -281 mV...... 75 Table 14: Species distribution of Pu in equilibrium with PuO2(am,hyd)...... 75 Table 15: Solubility of Pa in the BC reference porewater system at 25°C, pH 8.355 and Eh -281 mV...... 80 Table 16: Solubility of Ra in the BC reference porewater system at 25 °C, pH 8.355 and Eh -281 mV...... 84 Table 17: Species distribution of Ra in equilibrium with RaSO4(s)...... 84 Table 18: Solubility of Sm in the BC reference porewater system at 25 °C, pH 8.355 and Eh -281 mV...... 89 Table 19: Solubility of Se in the BC reference porewater system at 25 °C, pH 8.355 and Eh -281 mV...... 94 Table 20: Solubility of Ag in the BC reference porewater system at 25 °C, pH 8.355 and Eh -281 mV...... 99 Table 21: Species distribution of Ag in equilibrium with AgCl(cr)...... 99 Table 22: Solubility of Sr in the BC reference porewater system at 25 °C, pH 8.355 and Eh -281 mV...... 103 Table 23: Solubility of Tc in the BC reference porewater system at 25 °C, pH 8.355 and Eh -281 mV...... 106 Table 24: Solubility of Th in the BC reference porewater system at 25 °C, pH 8.355 and Eh -281 mV...... 110 Table 25: Solubility of Sn in the BC reference porewater system at 25 °C, pH 8.355 and Eh -281 mV...... 115 Table 26: Species distribution of Sn in equilibrium with SnO2(am)...... 115 Table 27: Solubility of U in the BC reference porewater system at 25 °C, pH 8.355 and Eh -281 mV...... 119 Table 28: Species distribution of U in equilibrium with UO2(am,hyd)...... 120 Table 29: Solubility of Zr in the BC reference porewater system at 25 °C, pH 8.355 and Eh -281 mV...... 123 Table 30: Mixed Th-hydroxo-carbonate complexes comprised in MOLDATA ...... 140 Table 31: Solubilty constants of Th(OH)4(am) in carbonate solution (log10K°s,1yz) and derived formation constants (logβ°1yz) ...... 142 Table 32: Recalculated formation constants of ternary Th(IV) complexes ...... 143 Table 33: Formation constants of ternary Th(IV) complexes recalculated by NEA ...... 144 Table 34: Recalculated formation constants logβ*1yz of Th(IV) complexes using the bicarbonate ...... 144

Structure of the document The present report comprises 5 main chapters and 4 appendices, which contents are briefly summarized in the following:

Chapter 1: Introduction

Chapter 2: In this chapter, first the Belgian disposal concept and Boom Clay porewater chemistry are briefly described. Afterwards, an overview about the geochemical code and the different thermodynamic databases that were used for the solubility and speciation calculations is given. Addionally, the in- house developed Geochemical Database Processor (GDP), that was used to compile the SCK•CEN reference TDB MOLDATA is presented.

Chapter 3: In this chapter, "speciation" is defined and the theoretical background for the calculation and interpretation of Pourbaix diagrams is given.

Chapter 4: In this chapter, first the solubility term as it is used in this report is defined, and then the thermodynamic background of solubility constants and saturation states is given. In the following, the solubility concept, the calculation procedure, the identification and selection of the solubility limiting solids are described. Afterwards, Source and Expert Ranges (SR and ER), and Best Estimate (BE) values are defined, and the uncertainty calculation approach that was used is illustrated. At the end of this chapter, the effects of different parameters (e.g. Eh, pH, pCO2, particle size) on solubility are discussed.

Chapter 5: This chapter comprises a description of the properties of each safety relevant radionuclide and the results of the speciation (Pourbaix diagrams calculated with the different TDB’s) and solubility calculations. The concentration limits for all radioelements and potential solubility limiting solids that were calculated with MOLDATA are presented. The reasoning that was taken as basis for the selection procedure is given and the determination of the Source and Expert ranges illustrated. Possible effects of changing/evolving conditions (pH, Eh, pCO2, ionic strength) on the stability of the selected solubility limiting solids are given, if known from literature and/or experimental results. In case experimental data were used to delimit, i.e. enlarge or restrict the SR, they are described. At the end of the element descriptions, general as well as specific remarks concerning the different thermodynamic source data(bases) and/or MOLDATA can be found.

Appendices: Annex I comprises a summarizing table of the concentration limits calculated with MOLDATA for each radionuclide and considered solid, in Annex II the Source and Expert Ranges for each radionuclide are tabulated, in Annex III details on the uncertainty calculations and thermodynamic data are given, and Annex IV represents a Technical Note, in which thorium data are presented and discussed that were published more recently than the ones comprised in the respective NEA review (Vol. 11; Rand et al., 2009) and in MOLDATA.

1 Introduction This report was prepared for Research Plan RP. W&D 0064 (WP 3: Radionuclide Speciation and Solubility) and can be considered as an extended update of a previously published report by Wang et al. (2000). It represents also a supporting document of the Level 4 (L4) integration report "Radionuclide migration in the far-field" (Bruggeman et al., in prep.), which corresponds to a deliverable of WP 1 and has to be prepared within the frame of the Safety and Feasibility Case 1 (SFC-1). In the latter report, the processes and mechanisms influencing the radionuclide migration/retention behaviour in the potential host formation and the selection strategy of migration parameters developed by SCK•CEN together with ONDRAF/NIRAS (O/N) are described in detail. Furthermore, a phenomenological model, which has been developed based on the research performed at SCK•CEN during the last 30 years, and which reflects the current state of knowledge with respect to the main BC relevant retention and migration processes, is put forward in this L4 report.

Main objective of the current report was to deliver a reference document, comprising information on the radionuclide speciation and solubility representative for "undisturbed" (far-field) BC conditions and to provide transparent and traceable input parameters, i.e. concentration limits for performance assessment (PA) calculations.

It should be explicitly mentioned that for the speciation and solubility calculations only inorganic ligands and complexation were considered. However, as revealed by long-term migration, batch and solubility experiments, transition metals, actinides and lanthanides are forming strong complexes with the humic substances of BC natural organic matter (NOM). The possible solubility enhancing effect related to the radionuclide association/interaction with the organic matter, which is thought to be of colloidal nature, is thus captured in transport/performance assessment (PA) calculations through the recently developed phenomenological model by Maes et al. (2011). According to this model, two species per radionuclide are allowed to migrate, i.e. the "free radionuclide" and the OM associated radionuclide (Rn-OM complex), which are both represented by their own parameter set.

Special attention has been paid to the description of the methodology put in place by SCK•CEN together with O/N to derive so-called solubility ranges (i.e. Source and Expert ranges) and Best Estimate (BE) values, as well as to the documentation of the selection procedure of the solubility controlling solids.

As such, the present work represents the basis for the numerical solubility values needed as input parameters in the ongoing safety and feasibility study and scoping calculations within the frame of SFC-1.

In total 25 elements are reviewed within this report, 23 safety-relevant radionuclides as well as calcium and carbon.

The radionuclide inventory for both spent fuel and vitrified high-level waste is provided by ONDRAF/NIRAS (https://www.nirond-km.be/gm/folder-1.11.315498, file "DCTs Version 2.2.xls") for a reference time of 10 years after unloading the spent fuel of the reactor. Based on these data, the respective inventory data are calculated taking into account a total cooling time of 60 years for category C waste (Weetjens et al., 2012).

2 System definition

2.1 The disposal system The recommended option for the long-term management of high-level and/or long-lived radioactive waste (HLW-LL)1 in Belgium is geological disposal in poorly indurated clay formations (i.e. Ypresian Clays, Boom Clay). Through its Waste Plan, ONDRAF/NIRAS (O/N) intends to satisfy its legal obligations, while providing the Government with all the elements needed to make a "decision in principle", in other words a general policy decision, about the long- term management of HLW-LL (ONDRAF/NIRAS, 2011). Within this framework, Boom Clay (BC) is investigated as one of the potential host formations. The Belgian concept (i.e. Supercontainer concept) for HLW-LL considers the installation of a multi-barrier repository system, which typically comprises the natural barrier, provided by the host rock and its surroundings (aquifers, biosphere), as well as the engineered barrier system (EBS). The former is also referenced as the geosphere or "far-field". The "engineered barrier system" in contrast, represents the man-made or artificial parts placed wihin a repository. The supercontainer design (Wickham et al., 2005) comprises different components, such as the vitrified HLW canisters or SF assemblies, surrounded by a carbon steel overpack, a concrete buffer and a stainless steel liner. The supercontainers are foreseen to be emplaced in galleries excavated in the host rock. Due to the plasticity of poorly indurated clays, the disposal galleries will be stabilized by a concrete lining. The space between the supercontainers and the concrete liner will be backfilled before the galleries will be sealed. The "near-field" includes the EBS and those parts of the host rock in contact with or near the EBS, whose properties have been affected by the presence of the repository.

2.2 Boom Clay reference pore water The solubility (and sorption) of the radionuclides depend on their speciation in the aqueous phase and thus on the chemistry of the porewater. In Table 1, the Boom Clay porewater composition for T=25°C is summarized. This composition was derived from the "Reference Boom Clay composition" given at T=16°C by De Craen et al. (2004). The major ion concentrations as well as the pH and Eh were calculated from/calibrated against an average (44 measurements) MORPHEUS water composition. This approach involved assuming equilibrium with the minerals calcite, pyrite, siderite, chalcedony and kaolinite, taking into account an ion exchange complex of 0.925 eq/5 kg of clay (i.e. 18.5 meq/100 g clay) and a pCO2 of 10-2.62. The same approach was applied to derive the Boom Clay porewater composition at 25°C. Since the Boom Clay reference porewater is highly supersaturated with respect to various clay minerals, mica, quartz, K-feldspar etc., these minerals were not allowed to precipitate during the simulations, in order to keep the water composition unchanged. This approach allowed radionuclide solubilities to be calculated under the reference Boom Clay conditions. It should be mentioned that for the derivation of the BC porewater composition at 25°C, the LLNL TDB (thermo.com.v.8.r6+.dat) was used.

1 Category A (LILW-SL): low- and intermediate-level short-lived waste. This category comprises radio-elements with half-lives of less than 30 years, emitting generally alpha radiation. Category B (LILW-LL): low- and intermediate-level long-lived waste. This category comprises radio-elements with half-lives of more than 30 years, emitting generally alpha radiation. Category C (HLW-SL and HLW-LL):high-level short lived and long-lived waste. This category comprises large amounts of beta and alpha emitting radio-elements having short or long half-lives. They are highly heat-generating. This waste arises from the reprocessing of irradiated nuclear fuel. Spent fuel that is not reprocessed also belongs to this category.

Table 1: The (recalculated) Boom Clay water composition at 25°C Species [mol/L] Al 4.63 ×10-8 -4 SiO2(aq) 1.95 ×10 Mg 6.56 × 10-5 Ca 5.04 × 10-5 -6 Fe 3.68 × 10 K 1.85 × 10-4 Na 1.55 × 10-2 Cl- 7.34 × 10-4 2- -5 SO4 2.41 × 10 -2.44 pCO2(g), atm 10

Eh, mV -281 pH 8.36

2.3 Computer code and databases The derivation of the pore water composition, as well as the speciation and solubility calculations presented within this report were performed using the recent releases of The Geochemist's Workbench code, versions 8.08, 8.10 and 8.12 (Bethke, 2010). Geochemist's Workbench represents a set of interactive software tools (ACT2, REACT, TACT and RXN) for solving problems in aqueous geochemistry, including those encountered in environmental protection and remediation, petroleum industry, and economic geology. Two of the four software tools, ACT2 and REACT, were used for the geochemical calculations. The former was applied for the construction of the so-called Pourbaix diagrams, while the latter program was used to derive the Boom Clay reference porewater composition, and to determine the prevalent speciation under Boom Clay conditions, as well as to calculate the radionuclide solubilities in this water.

With respect to the databases used for the different purposes, the following can be stated. The Pourbaix diagrams were calculated for each element with the four individual databases, i.e. the NAGRA/PSI, ANDRA, LLNL and NEA databases, as well with MOLDATA. Aim of this comparative study was to identify differences/discrepancies between the individual databases and MOLDATA, but also to "discover" possible inconsistencies and/or errors within MOLDATA.

The solubility calculations were also performed with all databases, however only the results of the calculations performed with MOLDATA will be presented in more detail, since the latter is considered to be the reference database for any SFC-1 related calculations.

2.4 MOLDATA The so-called MOLDATA project was launched in 2004 as part of the Research Plan "Geochemical and reactive transport modelling" (RP.WD.039) and has been followed-up within the frame of RP.WD.0064 "Radionuclide migration and retention processes in Boom Clay". The main objective has been the compilation of a high-quality and internally consistent thermochemical database that is to be used as reference database for geochemical and reactive transport calculations, which are increasingly applied to interpret phenomenological processes and safety assessment studies. In that sense, MOLDATA represents one of the building blocks on

which the PA of the radioactive waste disposal system is based. Therefore, reliability of the modeling results is essential and the quality of the thermodynamic data comprising the database has to be ensured. In 2006, ONDRAF/NIRAS defined quality measures (ONDRAF/NIRAS, 2006) in order to meet the high quality requirements of the Safety Case 1 (SFC-1) in general, and also to ensure the high-quality implementation and use of databases. Database implementation includes a verification process, meaning that the data need to be verified for their scientific correctness and need also to be well documented. Besides this, a high quality database is expected to be state-of- the-art, internally consistent and complete. The compilation of MOLDATA has been performed based on these quality measures. In order to strive for completeness, thermodynamic data from different so-called "source databases" were incorporated into MOLDATA. Thermodynamic data of the following databases have been evaluated within the frame of the MOLDATA compilation/selection procedure:

1) Lawrence Livermore National Laboratories (LLNL) database: thermo.com.v8.r6+.dat and thermo.com.r7beta.dat (Johnson et al., 1991; Delany and Lundeen, 1990; Helgeson et al., 1978); 2) NAGRA/PSI TDB: Chemical Thermodynamic Database 01/01, (Hummel et al., 2002); 3) ANDRA TDB: ThermoChimie v.5, June 2005; 4) NEA TDB’s:  Guillaumont et al. (2003): Update on the Chemical Thermodynamics of Uranium, Neptunium, Plutonium, Americium and Technecium (Vol. 5);  Gamsjäger et al. (2005): Chemical Thermodynamics of Nickel (Vol. 6);  Olin et al. (2005): Chemical Thermodynamics of Selenium (Vol. 7);  Brown et al. (2005): Chemical Thermodynamics of Zirconium (Vol. 8); and  Rand et al. (2008): Chemical Thermodynamics of Thorium (Vol. 11).

For further details concerning the source databases and the MOLDATA compilation strategy, readers are referred to the report of Wang et al. (2011).

It should be clearly mentioned, that the MOLDATA compilation and selection procedure did not involve a full review of the thermodynamic data. The strategy of the selection procedure has been mainly based on incorporating the most accurate, state-of-the-art and accepted data by the scientific community. For example, with respect to radionuclides this approach means that priority was given to the NEA reviewed data. Other used criteria were the completeness and consistency within the database. When inconsistencies were encountered, the underlying thermodynamic data were reviewed in order to trace the problem and to make adaptations.

MOLDATA currently comprises in total 3855 species, with 84 basis species, 2140 aqueous species, 1436 minerals and 195 gases.

2.5 GDP The in-house developed Geochemical Data Processor (GDP) represents the tool, i.e. computer software (De Soete, 2010) that made the compilation of MOLDATA possible. As mentioned above, four source databases (NAPSI/NAGRA, ANDRA, NEA – Vol. 5-11, and LLNL), comprising a total of ~7500 species were used to compile MOLDATA. In order to process this huge dataset, a tool was needed enabling to upload and store them in a standardized format. The storage of the source data occurs in the so-called "Central Database" which can be seen as the

"core" of the GDP. The import of the data is generally done via a wizard like tool, by which the user - in an interactive manner - is guided through the "data upload process". As the source databases exist in different formats, i.e. GWB, PhreeqC or tabdelimited format, the data import into the Central Database requires a standardization process. This conversion of a datasets to a standardized format is done by so-called "import filters". In the same way, data can be exported/downloaded in different formats, for which so-called "export filters" were programmed. Each different format requires a specific import and export filter to be developed. For further details concerning the software, it is referred to the GDP manual (De Soete, 2010).

3 Speciation calculations and Pourbaix diagrams

Definition: Speciation is the distribution of aqueous species among free ions, ion pairs, and complexes (Nordstrom and Munoz, 1994).

Based on the water/rock equilibrium model, the speciation of elements (i.e. major elements, transition metals, actinides, lanthanides) can be evaluated by the use of so-called Pourbaix diagrams. A Pourbaix diagram generally illustrates the equilibrium aqueous species distribution of an element as a function of Eh and pH. Additionally, it is possible to map out fields of Eh and pH over which different solid phases are possibly stable. The vertical axis represents the Eh, which is the voltage potential with respect to the standard hydrogen electrode (SHE), as calculated by the Nernst equation:

0 0.059 aOx Eh  E  log n aRed where: E0 is the standard potential; n is the number of electrons involved in the reaction;

aOx is the activity of oxidized species; and

aRed is the activity of reduced species.

The horizontal axis represents the pH, with

pH = -log aH+

and aH+: activity of hydrogen ions.

The boundaries of fields within which a particular aqueous species dominates the aqueous speciation of an element, and stability fields of minerals, are represented/delimited by lines. The diagrams that are presented within this report have been calculated using the reference porewater composition for the Boom Clay (see paragraph 2.2) and radionuclide activities of 10-8. The boundary between a field representing the predominance of a particular aqueous species and a solid phase represents equilibrium for this radionuclide concentration.

It has to be mentioned, that the activity of the dissolved ions, the presence of binding agents (ligands), as well as temperature may modify a Pourbaix diagram and shift the lines/boundaries in accordance with the Nernst equation. The in-situ Boom Clay conditions, i.e. Eh -281 mV and pH ~8.4 are represented by a cross within the Pourbaix diagrams and enable the immediate recognition of the dominant species or stable solid phase.

In calculating Pourbaix diagrams, the effect of complexing anions on radionuclide speciation has to be treated with care, since some anions have varying speciation over the Eh-pH range considered. The speciation of dissolved sulphur and inorganic carbon is redox- and pH-dependent (see Figure 1). Consequently, erroneous results would be obtained by using constant 2- - concentrations for the complexing ligands SO4 and HCO3 (for example) over the entire Eh-pH

region. In the present report, the speciation of sulphur and inorganic carbon are taken into account by using so-called mosaic diagrams (Bethke, 2010). The chloride (Cl-) speciation, in contrast, is constant over the whole Eh-pH range.

HSO - 2- 4 SO4

.5 CO2(aq)

- HCO3

Eh (volts) Eh 0 2- H2S(aq) CO3

Methane µ

–.5 HS- 25 C 0 2 4 6 8 10 12 14 pH

Figure 1: Sulphur and carbon speciation as function of Eh and pH

Complexation of the radionuclides by organic ligands was not taken into account in the current calculations.

3.1 Species activities and activity coefficients In order to understand the qualitative meaning of activity coefficients, it is important to consider how solution concentration affects ion interaction. Generally, ions in solution interact with each other as well as with the water molecules and neutrally charged solute species. At low concentrations of a species i (ci), ionic interactions can be ignored. The properties of such a solution approach those of an ideal solution, and would become ideal in an infinitely diluted solution, so that:

[i] = ci

Determining equilibrium constants experimentally is commonly done by using media of high ionic strength (e.g. solutions of NaClO4, KNO3, Na2SO4). These electrolytes represent so-called "non- ideal solutions". In order to account for their deviation from ideality, the concept of activity coefficients, γ, was introduced into thermodynamics, according to which:

[i] = i × ci

Under dilute conditions, γ is expected to be close to 1 (i.e. approaching an ideal solution).

According to Coulomb's law, the electrostatic force between two point electric charges is proportional to the product of the charges and inversely proportional to the square of the distance between the two charges. Therefore, activity coefficients in dilute solutions decrease with increasing concentration, due to the fact that the coulombic forces become stronger as ions pack more closely together.

Different methods exist for the calculation of activity coefficient in electrolyte solutions. The most commonly used existing activity models are all based on the Debye-Hückel theory. This theory was proposed by Peter Debye and Erich Hückel in 1923 to calculate mean activity coefficients for ions in dilute solutions, and was further elaborated by Robinson and Stokes (1968), who derived the well-known Debye-Hückel equation:

 A z2  I log  i i 0 1 ai  B  I where 0 ai (given in Å): is the activity coefficient; zi is the electrical charge; A, B are constants (T-dependent; at 25°C A = 0.5092 and B = 0.3283); and I is the ionic strength of the solution (given in molal units).

The ionic strength I is defined as half the sum of the product of each species (molality mi) and the square of its charge: 1 I  m z 2  i i

A disadvantage/limitation of the Debye-Hückel equation is that it becomes inaccurate at moderate ionic strength, i.e. above about 0.1 molal (Stumm and Morgan, 1996).

Another model, that can be carried to somewhat higher ionic strengths, i.e. 0.3-0.5, is the so-called Davies model. The Davies equation can be considered as a variant of the Debye-Hückel equation and is represented by:

2  I  log i  A zi   0.3 I  1 I 

0 The equation was simplified from the original Debye-Hückel equation by noting that ai  B is about 1 at 25°C.

An alternative correction term similar was also added to the Debye-Hückel equation by Helgeson (1969), to produce the so-called ‘Bdot equation’:

2  Azi I log i   0  B I 1 ai B I where:  B is a coefficient that depends on the charges of the dissolved species and temperature.

This ‘Bdot’ equation can yield accurate estimates of i for ionic strengths of up to about 1 molal in NaCl-dominated solutions.

4 Solubility - Definition and theoretical background Definition for solubility: The solubility of a given element is the sum of the stoichiometric concentrations of all dissolved species containing the element (Garrels and Christ, 1965).

4.1 Solubility constants Dissolution and precipitation reactions of crystals/minerals represent heterogeneous equilibrium reactions, meaning they are involving solid and liquid/aqueous species, i.e. different phases. Conventionally, equilibrium constants involving a solid compound are typically denoted as "solubility constants", rather than as formation constants of the solid (Brown et al., 2005). Solubility constants can be defined as follows:

↔

, where: , is the solubility product, the index "s" to the equilibrium constant indicates that the constant refers to a solubility process and the subscript "0" indicates that the equilibrium reaction involves only uncomplexed aqueous species.

In case that the solubility constant includes the formation of aqueous complexes, the following notation is used:

↔

MaLb is a solid phase having components M (typically a metal) and L (typically a non-metal) with stoichiometric coefficients a and b; [M] represents the activity of aqueous species M; and [L] represents the activity of aqueous species L with the solubility constant

,, where:

MmLq is an aqueous complex consisting of components M and L with stoichiometric coefficients m and q.

Example: + - UO2F2(cr) ↔ UO2F + F

+ - Ks1,1 = Ks1 = [UO2F ] [F ]

Similarly, an asterix is added to the solubility constant, if it simultaneously involves a protonation equilibrium:

↔

∗ ,,

Example: + 2+ - U(HPO4)2 × 4H2O(cr) + H ↔ UHPO4 + H2PO4 + 4H2O

∗ ∗ ,, ,

The equations above are consistent with notations used by the NEA and mainly extracted from Brown et al. (2005).

4.2 Solubility and saturation Various types of precipitates may be formed depending on the degree of solution saturation, i.e. oversaturated or saturated solutions, respectively.

In order to determine whether a solution or natural water is over- or undersaturated or at equilibrium with respect to a solid phase, so-called saturation indices can be calculated. The saturation index (SI) is defined as follows: Q SI = log( ) Ks where Q is the ion activity product and Ks is the solubility constant. Depending on SI, a system can have three saturation states, i.e. SI > 0, supersaturation SI = 0, in equilibrium SI < 0, undersaturation The most stable solid phase is characterized by the highest positive SI, while the less stable phases have lower saturation indices. In other words, the most stable phase is the least soluble with the lowest solubility product.

Although the phase characterized by the highest SI represents the thermodynamically most stable phase under the prevailing chemical conditions, it might not be the relevant solubility-limiting phase for a given geochemical system. There can be a variety of different reasons for such a situation, such as kinetic constraints, different degrees in crystallinity, and/or the particle size, to mention only the main ones. The influence of these factors on solubility is discussed in more detail in section 4.3.

4.2.1 Calculation procedure As mentioned above, solubility limitation represents a significant retardation mechanism for the majority of waste-relevant radionuclides. As it is almost impossible to measure solubility limits for each radionuclide under in-situ conditions, the common approach is to perform geochemical equilibrium calculations to estimate solubilities. These calculations use the chemical composition of the porewater that is representative of the environment under consideration (here: the far-field). The most reliable and complete thermodynamic dataset is used and expert judgement decides which solid phases should be considered for each radionuclide.

The solubility assessment described in this report is mainly based on geochemical calculations, which comprised two main steps. In the first step, the potential solubility controlling phases present in the thermodynamic database were identified. During the second step the concentration limits for all identified phases (one-by-one) were calculated. This was done by equilibrating one gram of solid phase (SI = 0) with the reference BC porewater. It should be explicitly mentioned, that the calculations were performed without considering an initial concentration constraint. In other words, the radionuclide inventory of the waste (activity concentrations) was not used to calculate or estimate an "inventory-limited porewater concentration" for the far-field. Due to this approach, the calculations were done for a large number of phases (i.e. with no pre-selection of phases likely to form), resulting in solubilities ranging between < 1.0 × 10-3 M and 8 × 10-35 M. Since the Boom Clay porewater is highly supersaturated with respect to clays, mica, quartz and other phases, the simulation of precipitation of these minerals was "disabled" to keep the water composition unchanged during the calculations. All calculations were performed at 25 °C instead of 16 °C, the in situ temperature of Boom Clay, as thermodynamic data are generally reported for 25°C. The solubilities for most radionuclides at 25 °C are in general slightly higher than the values at 16 °C, with some exceptions such as the carbonate minerals. The solubilities calculated with MOLDATA and the other TDB’s are reported and discussed in chapter 0.

In order to underpin the solubility calculations, a second series of calculations was undertaken to recalculate the Pourbaix diagrams by including also the identified waste-relevant phases. As for the speciation calculations, initially trace radionuclide activities (i.e. 10-8) were added to the Boom Clay porewater composition (i.e. major anions) to illustrate the stability fields of the solids. If considered useful, diagrams at higher activities were also calculated. Depending on the specified radionuclide activity, the most stable phase(s) appear(s) in the diagram. In the following, a potential precipitation/formation sequence of minerals can be anticipated by suppressing this stable phase due to which successively the following ones (less stable ones) appear in the diagrams. The least stable phase represents the last one shown in the diagrams. Taking into account kinetic considerations, the latter solid would represent the most favourable and probable one to precipitate and to control the radionuclide solubility, respectively.

4.2.2 Reasoning and approach of solid phase selection The solubility assessment and solid phase selection procedure was mainly based on thermodynamic principles and calculations, high-quality and site-specific experimental data (if available), as well as on expert judgement. The latter was especially required, when selecting between all thermodynamically possible phases and the kinetically most likely/favourable phases to form. An "analogy approach" was applied for radionuclides for which neither thermodynamic calculations have been possible (due to the lack of data), nor experimental data were available, e.g. Ac and Cm. The latter two elements belong to the group of actinides and are known to show a

similar chemical behaviour as americium, due to which parameter values derived for Am can be transferred to and applied for Ac and Cm. With respect to the experimental values, only results of solubility experiments using so-called Synthetic Boom Clay Water (SBCW) as medium were considered. Characteristically, SBCW represents an undisturbed BC porewater of 15 mM NaHCO3 containing no organic matter. Therefore, taking into account only solubility data derived in this water is consistent with our thermodynamic calculations, in which organic complexation reactions have been neglected. This approach is in agreement with the strategy that radionuclide/organic matter interaction processes are captured afterwards in PA calculations within the NOM facilitated transport model (Maes et al., 2011).

From the viewpoint of thermodynamics, the most stable solid would be selected as the controlling phase because thermodynamically less stable phases would ultimately be replaced by the most stable phase. However, it cannot be demonstrated that the thermodynamically most stable solid appears during the regulatory period under the expected repository conditions. This fact makes identification of the controlling solid purely from thermodynamic calculations unreliable. The Ostwald Step Rule provides a useful guide for this situation "of conflict" (see section 4.3.2 for further details). Precipitation kinetics is the governing factor for the Ostwald Step Rule. In other words, more stable phases are prevented from precipitation because less stable phases are kinetically favoured. The latter are often "amorphous", and may with time, i.e. during ageing transform to more crystalline solids. It should be mentioned that Ostwald’s theory is not accepted by the entire scientific community and that there are observations contradicting his theory, which have cast some doubt about the universal application of Ostwald’s rule (Deelman, 1999).

Despite the fact, that in the last decade solid solution thermodynamics has become more and more relevant (also) in the nuclear field, the implementation of solid solution formation into geochemical models is not as evolved as the theory and remains a challenge. Therefore, the formation of solid solutions was not considered in the present solubility assessment and is only discussed qualitatively (see section 4.4).

The formation of "real", "intrinsic" or "eigencolloids" was also not integrated into the solubility assessment, although it is well known that this process may drastically increase the equilibrium concentration, i.e. lead to so-called "apparent" solubilities that may exceed the thermodynamic one by several orders of magnitude.

Even though also Natural Analogue observations are often considered in solubility assessments and represent commonly an integral part in PA studies of nuclear waste repositories (Bruno et al., 2007), they were not taken into account in our current evaluation approach.

Based on the strategy described above, a consistent set of so-called Source and Expert Ranges (SR and ER) of concentration limits, as well as "Best-Estimate" (BE) values and associated uncertainties was derived.

A detailed sensitivity analyses, i.e. evaluation of effects on the solubility limits due to parameter variation (e.g. pH, pCO2, Eh, ionic strength) was not performed, but tendencies and possible effects are mentioned.

4.2.3 Solubility Source and Expert ranges Within the frame of SFC-1, ONDRAF/NIRAS has developed a new methodology for the uncertainty treatment of safety parameters (ONDRAF/NIRAS, 2011). As a detailed description of this methodology would go beyond the scope of this report, only the definitions will be given here and it is referred to the reference given above for further details.

Expert range (ER): the range within which experts expect the parameter value to lie. This range is also referred to as the realistic or likely range.

Source range (SR): the range outside of which the experts do not expect the parameter value to lie. This range is also referred to as the support range.

As solubility is a function of both, the type of solid/mineral and aqueous speciation, the Source Range comprises all reasonable phases predicted by the thermodynamic calculations to be able to control the solubility of the radionuclide under consideration. Consequently, the source range may comprise solubility values varying over several orders of magnitude. Expert judgement was used to eliminate unreasonable and non-relevant phases, such as high-temperature phases and/or phases that are unstable under the reference Boom Clay conditions. Very soluble phases, characterized by solubilities > 10-3 M were discarded from the selection as it seems unrealistic that such high concentrations may ever be reached in the far-field.

The Expert Range (ER) is represented by the "Best Estimate" (BE) solubility value, i.e. the solubility of the phase considered to be the most relevant one, including its (thermodynamic) uncertainty. The evaluation of the most relevant and probable solubility limiting phase has been mainly based on the calculations of Pourbaix diagrams and expert judgement. The calculations enabled to anticipate, although in an indirect manner, the formation/precipitation sequence and the stability of the potentially possible solubility limiting phases, which was then evaluated by expert judgement. As an underlying rule, the Ostwald principle was applied, by which the least stable (most soluble) phase represents the kinetically favoured phase to form. In applying this approach, reliable and soundly based values are transferred to PA.

In cases where experimental solubility values were available (e.g. for U, Tc and Se), they have been also included in the parameter ranges.

4.2.4 Uncertainties It should be clearly mentioned that the uncertainties as referred to in section 4.2.3, represent only the uncertainties associated to the solubility products calculated for the reference Boom Clay pore water composition. This composition is considered to be representative for the "undisturbed" Boom Clay geochemical conditions in the Mol-Dessel area. Other possible uncertainties (e.g. pore water composition) are considered to be negligible compared to the "thermodynamic uncertainties".

The uncertainties associated to the solubility constants/products were calculated according to the NEA guidelines (Silva et al., 1995), i.e. by using a simple error propagation method:

…

In case the uncertainty on the solubility constant under consideration and the formation constant of the dominant aqueous complex were available, the "total uncertainty" was directly calculated from the corresponding ("individual") uncertainties. For many solid phases however, the uncertainty on the solubility constant needed first to be derived from the standard molar Gibbs energies of formation fGm° of the reactants and products involved in the solubility equilibrium. The following example illustrates the taken approach:

Example: Americium

Am(OH)3(am) represents one of the potential solubility limiting phases comprised in the source range. The thermodynamic uncertainty related to this phase arises from the uncertainty on the - solubility product/constant and the formation constant of Am(CO3)2 , which corresponds to the dominant aqueous complex under BC conditions. Propagating the formal uncertainties of the relevant equilibria produces the following uncertainty:

+ 3+ ° Am(OH)3(am) + 3 H ↔ Am + 3 H2O log10K = 16.90 3+ - - + Am + 2 HCO3 ↔ Am(CO3)2 + 2 H log10β2= -7.75 + - - ° Am(OH)3(am) + H + 2 HCO3 ↔ Am(CO3)2 + 3 H2O log10K = 9.15

° ∆ of Am(OH)3(am) : no data available ° - -1 ∆ of HCO3 : ± 0.251 kJ mol ° - -1 ∆ Am(CO3)2 : ± 5.911 kJ mol ° -1 ∆ of H2O : ± 0.041 kJ mol

2 0.251 5.911 3 0.041  = ± 5.93 kJ mol-1 / 2.303 RT (5.708) = ± 1.04 where R is the gas constant, 8.314510 x10-3 kJ mol-1 K-1 and T is the temperature in [K], here 298.15 K.

- The equilibria and associated uncertainties for Am(OH)3(am) and Am(CO3)2 given by Guillaumont et al. (2003) are:

+ 3+ ° Am(OH)3(am) + 3 H ↔ Am + 3 H2O log10K = 16.9 ± 0.8 3+ 2- - Am + 2 CO3 ↔ Am(CO3)2 log10β2 = 12.9 ± 0.6

0.8 0.6  = ± 1.00

+ 2- - ° Am(OH)3(am) + 3 H + 2 CO3 = Am(CO3)2 + 3 H2O log10K = 29.8 ± 1.0

It can be seen that the uncertainty calculated from the standard molar Gibbs energies of formation (1st approach of calculation) is higher than the one calculated from the uncertainties on the solubility/formation constants given by NEA, despite the fact that even no uncertainty on the standard molar Gibbs energy of formation for Am(OH)3(am) is comprised in the first calculation. In case, uncertainties on reaction constants were available in the source databases, the uncertainties derived from these reaction constants were selected to determine/delimit the expert range as they are considered to be qualitatively better as they generally represent measured values.

It should be mentioned though, that not for all solids reaction constants (and uncertainties) are given in the databases, so that the second calculation method needed to be applied to derive the uncertainties.

4.3 Parameters influencing solubility

4.3.1 Influence of particle size The influence of the particle size or molar surface area on solubility was elaborated by Schindler (1967). Small solids or finely dispersed particles are generally more soluble than large crystals and solubility increases as the size of a crystal/particle decreases below a threshold size of about 1 μm. Consequently, small crystals/particles are thermodynamically less stable or metastable and with time should recrystallize and convert to larger ones. For particles that are smaller than the threshold it is important to take into account the particle size when evaluating the solubility of a solid phase. The following equation was derived to quantify this effect of molar surface/particle size on the solubility product: log ∗ log ∗ → 3.454 where:

*KA is the solubility constant for the solute particle with the molar surface area Am;

*KA→0 is the solubility constant for the substance with the molar surface area tending towards zero (i.e., when the particles are large);  is the surface tension of the solute particle in the solvent;

Am is the molar surface area of the solute; R is the universal gas constant; and T is the absolute temperature.

4.3.2 Ostwald rule and ripening According to the current state of knowledge the so-called Ostwald Step Rule or Law of Stages (Steefel and Van Cappellen, 1990) describes the phenomenon of precipitation of a more soluble and hence less stable phase and its replacement by a series of progressively more stable phases (see also 4.2.2).

This behaviour is explained by a combination of precipitation kinetics and energetic factors. The mineral with the fastest precipitation rate under the considered conditions will form preferentially. In other words, the nucleation of a more soluble phase is kinetically favoured over that of a less soluble analogue/polymorph, due to the lower mineral/solution interfacial energy of the more soluble phase. If the solution/porewater is sufficiently oversaturated, a metastable phase may

precipitate first due to its higher precipitation rate. Over time, as the solution composition evolves, the relative dissolution and precipitation rates will change and a more stable secondary mineral assemblage may form.

As the Ostwald step rule has been taken into account only in a qualitative way for the solubility assessment, the thermodynamic theory will not be elaborated in detail here and the interested reader is referred to relevant published literature (e.g. Stumm and Morgan, 1996; Helgeson et al., 1984; Oelkers and Schott, 2009; Zhang and Nancollas, 1990; Deelman, 1997).

4.3.3 Influence of temperature and pressure Solubility constants, like other types of equilibrium constants, are functions of temperature and pressure. The reference temperature used in the calculations has been fixed at 25°C, although the in-situ groundwater temperature is 16°C. This adaptation was done as the thermodynamic data comprised in MOLDATA (and also in the other TDB’s) refer to standard state conditions, i.e. 298.15 K (25°C) and 1 bar. It is judged, however, that this temperature adaptation has no important effect on the solubility calculations. It is known that the solubilities of many inorganic salts increase with temperature, while those of a number of compounds of interest in natural waters (e.g. CaCO3 and CaSO4) decrease with temperature. Thus, in order to give general statements on the influence of temperature on solubility, sensitivity calculations should be performed for each radioelement. This would however require that reaction enthalpy, entropy and heat capacity data are available for all aqueous species and solids, which is currently not the case for MOLDATA.

In contrast to temperature, pressure has only a slight effect on solubility, except if extreme pressure conditions are encountered, i.e. outside the range relevant for the assessment of geological disposal of HLW-LL in Boom Clay.

4.3.4 Influence of ionic strength The solubility equilibrium is influenced by the ionic strength of the solution (Stumm and Morgan, 1996). In general, solubility increases as the concentration of the electrolyte solution increases. This increase in solubility is due to the change in activity coefficients by changing ionic strength, and is accounted for by the Debye-Hückel theory. If a solubility constant, as defined under section 4.1 refers to standard state conditions of 25°C, 1 bar and infinite dilution (I = 0), then the ion activities are equal to the ion concentrations under these conditions. In this case, the activity coefficients and mean ion activity coefficients become equal to unity. It is known that these equilibrium constants (in terms of activities) are obtained by extrapolation of experimental solubility data to zero concentration. If these constants are used at ionic strength of I  0, activity corrections have to be performed, which is/has to be done by applying the appropriate activity models.

4.3.5 Influence of inorganic complexation It is known that the solubility of solid phases can be enhanced by the presence of inorganic ligands in solution, such as hydroxide and carbonate, to mention only the environmentally most relevant ones. The complexation behaviour of the light actinides (Th, Pa, U, Np, Pu, Am, Cm) has been studied and described in detail by many authors (Kim, 1991; Neck and Kim, 2001; Runde, 2000 a; Fanghänel and Kim, 2002). The complexation of a dissolved actinide ion with ligands in solution

generally increases the total actinide concentration. The stronger the complexation affinity and the larger the stability of the complex, the higher is the influence on solubility. Kim et al. (2008) determined the following (inorganic) complexation affinity/strength for actinides:

- 2- - - 2- 2- - - OH , CO3 (HCO3 ) > F , HPO4 and SO4 > Cl , NO3 .

Generally, the stabilities of actinide complexes at a given oxidation state increase with atomic number. This increase in stability follows the trend of the actinide contraction, wherein the ionic radii decrease with increasing atomic number. Another parameter affecting the stability of complexes, and thus solubility, is the effective charge. The higher the effective charge, the more stable complexes are formed in solution and the more stable solids (with lowest solubility) are precipitated. On the other hand, for the actinides forming the weakest complexes, the solids are characterized by the highest solubility (Runde, 2000 a).

Generally major cations do not affect directly the complexation behaviour of actinides, but they may have an indirect influence on the solubility of actinides through their effects on ionic strength and their competition with the actinides for complex-forming anions (Appelo and Postma, 2005).

4.3.6 Dissolved Organic Carbon in BC porewater Boom Clay porewater at Mol is rich in Dissolved Organic Matter (DOM) and characterized by a mean Dissolved Organic Carbon (DOC) content of ~115 mg C/L (De Craen et al., 2004). The "dissolved" fraction of organic carbon is generally an operational classification, referring to compounds < 0.45 µm. DOC is mainly composed of humic substances, which can be divided into three groups according to their solubility: humic acids, fulvic acids and humin (Stevenson, 1982). Humic acids represent the fraction of humic substances that is not soluble in acid solutions, but is soluble at higher pH values (alkaline solutions). Fulvic acids correspond to the fraction of humic substances that is soluble in solutions under all pH conditions. They remain in solution after removal of humic acid by acidification. Humin belongs to the third group of humic substances and is neither soluble in acid nor in alkaline solutions. Within the frame of a PhD thesis (Dierckx, 1995) and the TRANCOM-Clay project (Diercks et al., 1999), the Boom Clay organic matter and different types of humic substances were investigated in detail. Results revealed that concentrates of organic matter from Boom Clay pore water contain about 70% of humic acids and 30% of fulvic acids (Dierckx et al., 1999).With respect to the chemical/elemental composition of humic and fulvic acids C, H, N, S and O represent the major constituting elements. While humic acids have a higher contribution of C than fulvic acids, the latter have a higher O content. The hydrogen content is about the same in HA and FA (Dierckx, 1995). Due to the simultaneous presence of carboxylic, phenolic and hydroxyl groups, humic and fulvic acids are often referred to as polyfunctional structures. These functional groups account for the strong complexation capacity of humic and fulvic acids and are of great importance as centers of metal ion binding (Dierckx, 1995). The main difference between humic and fulvic acids, is their molecular weight (MW). While fulvic acid molecular weights range between 500 to 2000 dalton (Da or amu), humic acids molecular weights range between 1000 to more than 10,000 Da and can attain colloidal dimensions (Silva and Nitsche, 1995). Henrion et al. (1991) determined that the DOC in the Boom Clay was smaller than 104 Da (100 kDa) and about half of the DOC was smaller than 1000 Da (1 kDa, Bruggeman, 2010). Since the size of humics in sea water and natural waters approach or fall in the range of colloidal systems (Choppin, 2003), it is difficult to draw a clear boundary between truly dissolved humics and colloidal/particulate matter. Based on the definition put forward by Stumm and Morgan (1996), particles with a size between 1 nm and 1 µm belong to the group of

colloids. Size distribution studies of different BC NOM batches performed by Bruggeman et al. (2010), enabled to determine a bimodal distribution consisting of small (< 30 kDa) and larger (30- 300 kDa) organic matter molecules. The NOM fraction smaller than 30 kDa is considered to represent the "dissolved" organic matter pool, while the 30-300 kDa fraction is considered to represent the "colloidal" organic matter. Additionally, the presence of very large (> 300 kDa) organic matter colloids was revealed. The latter are assumed to be immobile under in-situ BC conditions, while particles smaller than 300 kDa are considered to be mobile in the BC porewater and able to act as complexing ligands and/or (carrier) colloids for the waste relevant elements, mainly actinides.

4.3.7 Influence of organic complexation Besides inorganic complexation, the formation of dissolved organic complexes may also have a significant impact on the solubility of radionuclides/actinides. According to Olofsson and Allard (1983), "actinides in all their oxidation states (III – VI) can form complexes with humic and fulvic acids comparable in strength to the hydroxide and carbonate complexes". The strong complexing capacity of humic substances (HS) is mainly based on the simultaneous presence of carboxylic, phenolic and hydroxyl groups.

It should be mentioned that within the NEA TDB project, a review of U, Np, Pu, Am, Tc, Se, Ni, Zr as well as H, Na, K, Mg and Ca compounds and complexes with oxalate, citrate, ethylenediaminetetra-acetate (EDTA) and iso-saccharinate (ISA) was performed in 2005. The complexation behaviour of the latter ligands was considered to be important due their potential importance for radioactive waste problems. Organic material comprised in intermediate and low- level waste repositories, such as bitumen and cellulose upon degradation may form products that exhibit complexing properties able to affect the solubility and mobility of radionuclides. Oxalate, due to it complexation strength represents the major product of radiolytic degradation of bitumen (Hummel et al., 2005). With respect to alkaline degradation of cellulose (C6H10O5) containing waste under cementitious conditions, isosaccharinic acid (ISA) represents the organic ligand of major concern in speciation calculations and performance assessment evaluations (Hummel et al., 2005; Valcke et al., 2003; Van Loon and Glaus, 1998). After all, it is due to this strong complexation behaviour, that these organic ligands are used in ion exchange resins and decontamination agents, respectively. As such, they can also become part of the radioactive waste inventory.

4.3.8 Influence of colloids When discussing colloids, first the terminology should be clearly defined. There exist a number of different classification approaches which may be based on the size, origin and/or composition of the colloid. If classifying colloids by composition, two types of colloids are differentiated (Van der Lee et al., 1994), i.e. type I colloids (also named "intrinsic colloids", "real colloids", "true colloids" or "eigencolloids") and type II colloids, which are also named "carrier colloids" or "pseudocolloids"). The former consist mainly of polymerized complexes of a particular element, while colloidal clay minerals, metal hydroxides, silica polymers, organics or others to which an actinide ion may bind are Type II carrier or (pseudo)colloids.

Intrinsic colloid formation is particularly important for actinides and has been described by many authors (Kim, 1991; Runde, 2000 a). The primary step to polymerization/polynucleation and thus colloid generation is hydrolysis. The tendency of an actinide to hydrolyse (or to form

complexes) is determined by its redox behaviour on the one hand and by its effective charge on the other hand. While in the III and IV oxidation states, the actinides form hydrated An3+ and An4+ cations, the highly charged penta- and hexavalent cations are unstable in solution and hydrolyse + 2+ instantaneously to form so-called trans-dioxo/actinyl cations AnO2 and AnO2 , respectively. The order in the tendency of actinides to form intrinsic colloids was given by Kim (1991) and corresponds to the order of decreasing charge of the central ion:

4+ 2+ 3+ + An > AnO2 > An > AnO2

With respect to tetravalent actinide ions, Pu(IV) is known to hydrolyse already under very acidic pH conditions (pH = 0-1) and at concentrations of > 10-6 M will polymerize and form very stable intrinsic colloids. Similar behaviour has been described for Th(IV), although the latter actinide is characterized by somewhat lower/weaker hydrolysis constants, i.e. the onset of hydrolysis starts only at pH 2.5-3 (Fanghänel and Kim, 2002). Hexavalent actinyl ions are characterized by the second strong hydrolysis tendency and are also susceptible to form colloids. Compared to Pu(IV) and Th(IV) colloids, Am(III) colloids are reported to be much less stable in aqueous solutions. Despite the fact that Am(III) intrinsic colloids seem less stable, they have been evidenced by different studies (Kim 1991; Olofsson and Allard, 1983; Buckau et al., 1996). With respect to + pentavalent actinyl cations, AnO2 are the most stable against hydrolysis among all actinide ions + in aqueous solutions (Kim, 1991). Lierse et al. (1985) could show that NpO2 remains stable as an unhydrolysed ion up to 10-2 M without undergoing polynucleation or colloid formation under near-neutral pH conditions.

Within the group of organic colloids, humic and fulvic acids play an important role, as they represent high molecular weight compounds and can be present in groundwaters as negatively charged macromolecules. Due to their high affinity for metal cations and radionuclides, pseudocolloids as described above may be formed.

According to Stumm (1992) and Stumm and Morgan (1996), the term "colloid" in the field of nuclear waste applies to any particle with a size ranging between 1nm – 1µm and which, due to Brownian motion and special means of stabilization, may be suspended or dispersed in solution/groundwater and maintained in suspension over a long period of time. Species smaller than 1nm are considered to behave similarly to dissolved species, while colloids larger than 1µm will settle out as they are too large for Brownian motion and to overcome gravitational forces.

When classifying colloids by their origin, typically natural groundwater colloids are distinguished from anthropogenic colloids (Triay et al., 1995). Groundwater colloids may be composed of organic and/or inorganic molecular constituents or microorganisms to which a particular element may be attached. Anthropogenic colloids are generated by physical, chemical and/or biological processes acting on materials created by humans. In the context of radioactive waste disposal, anthropogenic colloids may be derived from the waste itself and the repository construction and sealing materials.

Colloids are characterized by high specific surface areas and high specific surface energies. Although colloids are generally stable species, their stabilities depend on their surface charges. Due to the latter (i.e. repulsion of electrical charges of the same sign) aggregation of colloidal particles is inhibited. Any (environmental) process which neutralizes the surface charge can result in the aggregation of the colloidal particles and cause their precipitation. The most commonly

encountered changes which result in precipitation are those of pH and ionic strength. Owing to their surface charges, colloids may sorb radionuclides and/or charged complexes of radionuclides. The extent of this sorption will depend upon the charges and sizes of the colloids and the charges of the radionuclide species.

All described processes of colloid formation are of great importance and can lead to enhanced and even extremely high so-called "apparent solubilities" which have no relation to the thermodynamically defined solubilities (Dozol and Hagemann, 1993).

4.3.9 Influence of Eh, pH and pCO2

The oxidation state of the water/solid system, as commonly represented by the parameter Eh, probably has the biggest influence on the solubility of many elements of interest to safety assessment. The Eh affects most importantly the oxidation state of some radionuclides/actinides and consequently their dissolution and precipitation behaviour. The oxidation state of the element under consideration will, however, also determine its complexation behaviour, which as mentioned above affects the solubility, with more complexation generally being reflected in greater solubility. The pH, often in combination with the partial pressure of CO2 (pCO2), plays an equally important role in controlling the aqueous speciation of many radionuclides. General statements concerning the combined effects of Eh, pH and pCO2 on solubility are difficult. Only a few experimental/solubility data exist for the reference BC conditions. Thus, the influence of these parameters might currently only be assessed by the help of the Pourbaix diagrams.

4.3.10 Influence of radiation The change and modification of mineral properties (e.g. volume, solubility), caused by radiation damage has been described by many authors (Ewing et al., 2000; Eby et al., 1992; Woodhead et al., (1991). Minerals of geologic interest include zircon (ZrSiO4) and monazite [(REE)PO4], uraninite (UO2) and the thorium polymorphs thorite [(Th,U)SiO4] and huttonite (ThSiO4). Zircon and monazite represent commonly occurring accessory minerals containing several hundreds to several thousands of ppm U and Th.

During the last three decades in particular, there has been increased interest in radiation effects in minerals and "complex" ceramics, due to the importance of radiation damage on the long-term behaviour of solids used for the immobilization of high-level nuclear waste, such as pyrochlore and zirconolite (Ewing et al., 2000). Radiation induced mineral transformations/damage comprise (besides other effects) amorphization/metamictization, and/or annealing. Radiation damage may be the result of α-decay events, ionizing radiation fields or heavy ion irradiations. In natural minerals, the dominant damage mechanism is a result of α-decay of 238U, 235U and 232Th and radionuclides in the decay chains of these isotopes. In nuclear waste forms the α-decay damage is the result of the decay of actinides, mainly 239Pu. A lot of effort has been expended in an attempt to determine the so-called α-decay-event dose required for the amorphization of the above mentioned and other minerals, but the estimated dose for the crystalline-to-amorphous transition may be very variable due to the varied thermal histories of metamict minerals (Lumpkin et al., 1998). Widely used techniques to study the effects of α-decay damage are actinide doping and ion-beam irradiation experiments. Via these experiments, it has been shown that the ion flux and temperature are critical parameters for the onset of amorphization. Besides this, the composition of a mineral seems to have an effect on both the amorphization dose, and radiation resistance.

Differences in susceptibility to radiation were shown to be also related to structural and bonding parameters.

With respect to the far-field of the repository, amorphization of the clay constituting minerals and/or other crystalline solids is not expected to be significant due to the long elapsed time before radionuclides will be released. During this long time, radionuclides initially present will have decayed significantly and hence exposure of the far-field to radiation following radionuclide release will be decreased. Additionally, the thermal stage is considered to last only around 2500 years and therefore, prior to the release of any radionuclides, temperatures will have decreased to values at which amorphization, metamictization and or annealing may not occur anymore.

Besides this, in most safety assessments amorphous phases have been chosen to represent the solubility limiting phases for the waste-relevant radionuclides. This approach implies that evaluating the possibility of increased solubilities due to amorphization is not needed.

4.4 The role of solid solutions Although the formation of solid solutions was not taken into account in the calculations presented within this report, the role of solid solutions in nuclear waste chemistry, in particular their generation during the nuclear cycle and their presence in waste repository components has gained increasing recognition in the last decade (e.g. SKIN EURATOM project). Special attention has been paid to understanding solid solution concepts used in nuclear waste geochemistry, since many actinides and fission products can be incorporated in secondary phases formed in the near- and far- fields of nuclear waste repositories. The following radioactive waste-relevant examples of solid solutions illustrate this fact:

 mixed oxide (MOX) fuel, representing a PuO2-UO2 solid solution;  irradiated fuel, corresponding to a solid solution of uranium and fission products (e.g.: Sr, Zr, Nb, REE);  metallic alloy precipitates containing Mo, Tc, Ru, Rh, Pd, Ag, Cd, In, Sn, Sb and Te;  oxide precipitates (Rb, Cs, Ba, Zr, Nb, Mo, Te);  borosilicate glasses, considered as solid solutions of hydroxides and metasilicates;  alteration/gel layers of borosilicate glasses;  zircon based alloys (= solid solutions containing Fe and smaller amounts of Cr, Ni, and Mo) as both stainless and carbon steels;  corrosion products of these alloys which form iron oxides and oxyhydroxides (e.g. green rust), and which at higher temperatures may form spinel phases, most of these phases forming solid solutions;  alteration phases of cement consisting mainly of Calcium-Silicate-Hydrate (CSH) solid solution phases, that are able to co-precipitate radioelements such as U, Np, Am, Cm and Eu;  calcite as a trace metal scavenger (e.g. actinides, REE, Cd, Co, Fe, Zn and Mn); and  barite, which commonly contains minor amounts of Sr and Pb, and trace amounts of radium (Ra), as solid solutions.

In order to simulate the dissolution and formation of a mixed solid phase, Bruno et al. (2007) suggest that the so-called kinetic co-dissolution and/or equilibrium co-precipitation approach should be applied in PA calculations.

The kinetic co-dissolution approach derives from the application of the stoichiometric saturation concept introduced by Thorstenson and Plummer (1977). According to this approach it is assumed that an ideally mixed solid solution dissolves in a congruent manner and therefore the dissolution of the trace component is proportional to its molar activity ratio in the solid phase.

The equilibrium co-precipitation approach is based on the thermodynamics of ideal solid solutions. Basically, it assumes that the activity of the trace phase is equal to the molar fraction when the concentration of the trace solid phase is sufficiently low. That is, the solution is assumed to obey Raoult’s Law under these circumstances.

For further details concerning these models, readers are referred to Bruno et al. (2007).

It should be mentioned, that where incorporated into PA calculations, these solid solution models have been mainly applied to dissolution and precipitation processes in the near-field, i.e. spent fuel alteration processes.

4.4.1 Drawbacks for the application of solid solution models In order to characterize and model an aqueous solid solution (AqSS) system, a lot of information is needed, including: T, p, pH, I, mi (equilibrium molalities of mixing components in the aqueous solution) and for the solid xi (mole fractions) as well as nss (amount of solid solution formed). These data are, however, not easy to obtain. Up to now, systematic solid solution data of sufficiently high quality are available only for carbonates, notably calcite. For other low-T mineral groups, such as sulphates, solid solution data are still scarce or even lacking.

5 Results The results are presented element by element in alphabetic order.

5.1 Actinium (Ac) General Actinium is a radioactive metal with an atomic number of 89. It is the first element of the actinides and gave the group its name. In nature, it is often associated with uranium ores and minerals. Chemically, actinium shows a similar behaviour as the rare earth elements (REE), in particular as lanthanum (La). The isotope 227Ac is the only naturally occurring isotope and also is produced in nuclear reactors via neutron irradiation of 226Ra. It decays with a half-life of 21.772 years by emitting mainly beta-particles. Its principal decay products are 227Th (half-life 18.5 days), 223Ra (half-life 11.4 days), and a number of short-lived products including radon, bismuth, polonium, and lead isotopes (Handbook of Chemistry and Physics, 1992-1993). All other Ac radioisotopes have half lives less than ten days. In terms of its potential for radiation induced health effects, 227Ac is about as dangerous as plutonium.

Solubility data provided to PA Taken to be identical to Am.

Data retained in the DCF’s BE/SR/ER: Taken to be identical to Am.

Remarks In the present MOLDATA TDB actinium data are not available. We therefore keep the conclusion of the previous DCF (Marivoet, 1999), that actinium data should be similar to that of americium.

5.2 Americium (Am) General Americium is a transuranic element and belongs to the group of actinides. The most common and longest-lived isotopes are 241Am and 243Am with half lives of 432.2 and 7.37 × 103 years, respectively. The latter two isotopes are produced in nuclear reactors via uranium or plutonium bombardment with alpha particles. Americium most commonly occurs in the trivalent oxidation state and as such forms insoluble fluoride, oxalate, iodate, hydroxide, phosphate and other salts (Penneman and Keenan, 1960). Other oxidation states have been described, i.e. Am(IV), Am(V) and Am(VI). In the trivalent and tetravalent state, and in the absence of carbonates or other ligands, actinides form hydrated An3+ and An4+ ions in solution, respectively. The positively charged or neutral An(III) hydroxo-complexes can be represented by the general formula 3-m An(OH)m , where m = 1,2,3 and 4. Tetravalent americium is only stable in the presence of strongly complexing agents, such as carbonate or fluorite (Hummel et al., 2002). The higher charged Am-ions in the V and VI states are unstable in solution, except under strongly oxidizing + 2+ conditions, where they hydrolyse instantly to form linear trans-dioxo cations, AmO2 and AmO2 , respectively (Runde, 2000a; Hummel, 2002). In the presence of carbonates, Am forms carbonate (3-2n) complexes of the general formula Am(CO3)n . For further details on Am geochemistry and experimental works, it is referred to the Topical Report on Americium (Bruggeman et al., 2012).

a) b) 1 1

- .5 .5 +++ AmO2CO3 Am+++ Am

++ AmHCO3 + AmCO+ AmCO3 3 - - Am(CO3)2 Am(CO3)2 Eh (volts) Eh (volts) 0 Am(CO )--- 0 --- 3 3 Am(CO3)3

Am(OH)3(aq) Am(OH)3(aq) µ µ

–.5 –.5 25°C 25°C 0 2 4 6 8 10 12 14 0 2 4 6 8 101214 pH pH

c) d) 1 1

.5 +++ .5 +++ Am Am AmO2(cr) AmO2(cr) AmHCO++ 3 + AmCO3 AmCO3OH.0.5 H2O(cr) -

Am(CO3)2 Eh (volts) Eh (volts) 0 0 --- Am(CO3)3 AmCO3OH.0.5 H2O(cr) Am(OH)3(cr) Am(OH)3(cr) µ µ

–.5 –.5 25°C 25°C 0 2 4 6 8 10 12 14 02468101214 pH pH

e) 1

.5 AmO (cr) Am+++ 2

++ AmHCO3 AmCO+ 3 - no graph - Am(CO )- Eh (volts) 0 3 2

Am(OH)3(am) µ

–.5 25°C 0 2 4 6 8 10 12 14 pH

Figure 2: Eh-pH diagram of americium (Am-C-S-O-H) for the BC reference porewater system. The assumed activity of Am = 10-8. Diagram a) LLNL / NAGRA/PSI TDB, b) ANDRA / NEA / MOLDATA TDB, c) MOLDATA TDB solid phases included, d) same as diagram c, but [Am] = 10-3 and [Na] = 1.3 × 10-2 (most stable solid shown), e) same diagram as d (least stable solid shown, more stable phases suppressed in calculation). Code: The Geochemist's Workbench- 8.10.

Speciation and solubility calculations -8 - Considering an Am activity of 10 , Am(CO3)2 represents the predominant aqueous species under BC conditions, independently of which database is used for the calculation (Figure 2 a, b). Solubility calculations were performed for all phases comprised in Table 2. As can be seen therein, the solubilities of the considered solid phases range between 1.25 × 10-3 M -8 -8 [Am(OH)3(am)] and 2.41 × 10 M [AmCO3OH × 0.5 H2O(cr)]. This means that at [Am] = 10 , solubility of Am would not be controlled by any of these solid phases and would only be solubility limited at Am-concentrations higher than 2.4 × 10-8. Besides this, the solubility calculations reveal that under the reducing BC conditions, americium oxides are not stable and highly soluble. Thus, these phases can be excluded as potential solubility limiting phases for Am.

Considering [Am] = 10-3 (and [Na] = 1.3 × 10-2), the following precipitation sequence could be anticipated (Figure 2 d, e):

1) Am(OH)3 (am) (least stable phase) → 2) Am(OH)3 (cr) → 3) AmCO3OH (am,hyd)

→ 4) Am2(CO3)3 (am) → 5) Am(CO3)1.5(cr) → 6) NaAm(CO3)2 × 5 H2O (cr)

→ 7) AmCO3OH × 0.5 H2O (cr)(most stable phase)

-6 As can be seen in Table 2, the solubility of AmCO3OH(am,hyd) was calculated to be 3.82 × 10 M. In Table 3, the species distribution and their percent fractions in equilibrium with this solid are - + illustrated. Speciation is dominated by carbonate complexes, such as Am(CO3)2 (87%), AmCO3 3- (8%) and Am(CO3)3 (4%). Furthermore, it can be seen that in presence of carbonate ligands, the neutral hydrolysis species Am(OH)3(aq) becomes only important at higher pH values (pH > 12).

Numerous solubility studies have been performed on the solid Am(OH)3. According to Guillaumont et al. (2003), the use of the notation of "crystalline" and "amorphous" to describe a specific solid phase would correspond to an "oversimplified model". This is explained as follows. As the solubility of Am(OH)3(s) depends strongly on its crystallinity and particle size, which both can change with time due to ageing and may also be affected by self-irradiation (in case of 241Am

presence), the solubility data that were (re)evaluated in the recent NEA review (Vol. 5), are quite scattered and vary over several orders of magnitude. Thus, data selected by the authors for Am(OH)3(am) and Am(OH)3(cr) cover the possible solubility range and "magnitude of the effect of an incomplete knowledge of the surface state of the solid" (as solubility is determined by the surface characteristics) (NEA Vol. 5, p. 348). As can be seen in Table 2, the calculated solubilities for the crystalline and amorphous Am(OH)3 under BC conditions are quite high and vary over two orders of magnitude.

Table 2: Solubility of Am in the BC reference porewater system at 25°C, pH 8.355 and Eh -281 mV. Code: The Geochemist's Workbench - 8.10. Database: MOLDATA Solubility controlling phases Solubility, [Am], mol/l -3 Am(OH)3(am) 1.25 × 10 -5 Am(OH)3(cr) 6.20 × 10 -6 AmCO3OH(am,hyd) 3.82 × 10 -6 Am2(CO3)3 (am) 2.52 × 10 -6 Am(CO3)1.5(cr) 2.35 × 10 -7 NaAm(CO3)2 × 5 H2O (cr) 7.65 × 10 -8 AmCO3OH × 0.5 H2O (cr) 2.41 × 10

Am2O3(cr) very soluble

AmO2(cr) very soluble

Source data: NEA, except Am(CO3)1.5(cr), which was copied from the NAGRA/PSI TDB to MOLDATA

The reaction constants of the Am solids comprised in Table 2 and in MOLDATA are the following: + 3+ Am(OH)3(am) + 3 H ↔ Am + 3 H2O log K = 16.9 + 3+ Am(OH)3(cr) + 3 H ↔ Am + 3 H2O log K = 15.6 + 3+ - Am2(CO3)3(am) + 3 H ↔ 2 Am + 3 HCO3 log K = -2.37 3+ - AmCO3OH(am,hyd) + 2 H+ ↔ Am + HCO3 + H2O log K = 4.13 + 3+ - NaAm(CO3)2 × 5 H2O + 2 H ↔ Am + Na+ + 2 HCO3 + 5 H2O log K = -0.35 + 3+ - AmCO3OH × 0.5 H2O + 2 H ↔ Am + HCO3 + 1.5 H2O log K = -11.0

Table 3: Species distribution of Am in equilibrium with AmCO3OH(am,hyd). Database: MOLDATA. Code: The Geochemist's Workbench - 8.08. Aq. Species [mol/L] Fraction [%] - -6 Am(CO3)2 3.32 × 10 87 + -7 AmCO3 3.17 × 10 8 3- -7 Am(CO3)3 1.51 × 10 4 2+ -8 AmSiO(OH)3 2.96 × 10 1 Total 3.82 × 10-6 100

Experimental data In-house measured solubility data on Am are scarce. However, within the TRANCOM II project (Maes et al., 2004), the solubility of Am was determined in synthetic Boom Clay water (SBCW) and measured to correspond to 2 × 10-8 M after ultrafiltration at 30 kD.

Liu et al. (2008) performed several solubility experiments with Eu(OH)3(s) in presence and absence of organic matter (OM) and under geochemical conditions representative for BC (pH ~8.3, glovebox atmosphere: 0.4% CO2). The Eu solubility in absence of OM was determined in Synthetic Boom Clay Water (SBCW = 0.014 M NaHCO3 as background electrolyte) from undersaturation direction. After 2 weeks of reaction time, phase separation was performed by filtration at 0.45µm and by ultrafiltration at 30 kD. XRD analysis of the solid phase after phase separation revealed that the initial Eu(OH)3(s) had transformed during the experiment to EuOHCO3(s). The lowest measured Eu solubility after ultrafiltration (30 kD) corresponded to 4.5 × 10-8 M. Eu-concentrations measured after 0.45µm filtration were around 10-15 times higher (4.5 × 10-7 – 6.75 × 10-7). However, also after ultrafiltration some samples contained Eu concentrations in solution as high as 5.0 × 10-7 M. The latter higher solubility was interpreted to be due to the presence of very small inorganic colloids of Eu(OH)x(CO3)y. -8 The calculated solubility using the NAGRA/PSI TDB of EuCO3OH(cr) corresponds to 3.01 × 10 M, which is slightly lower than the lowest measured Eu-concentration (after UF). The calculated solubility for EuCO3OH(cr) using MOLDATA is around two orders of magnitude higher, i.e. 2.83 × 10-6 M than the one calcuated with the NAGRA/PSI TDB. This higher solubility can be explained by the different speciation in the equilibrium solution, i.e. the presence of mixed 2- - hydroxo-carbonate complexes, i.e. EuOHCO3(aq), EuOH(CO3)2 and Eu(OH)2CO3 (copied all from the LLNL TDB), which represent the dominant aqueous species in the MOLDATA calculations and which are comparativeley more stable than the simple Eu-carbonate complexes, - + i.e. Eu(CO3)2 and EuCO3 comprised in the NAGRA/PSI TDB. It should be mentioned that in the - + ANDRA TDB other thermodynamic data for EuOHCO3(s), as well as Eu(CO3)2 and EuCO3 are included (Spahiu and Bruno, 1995), due to which the calculated solubility is again different, i.e. 8.95 × 10-8 M. Comparison of the experimentally measured Eu-concentrations and calculated solubilities using the NAGRA/PSI and ANDRA TDB reveals quite good agreement, the solubility is overpredicted by using MOLDATA. This suggests that the mixed hydroxide carbonate complexes for Eu copied from LLNL should be re-evaluated. With respect to the presence of colloids, a clear filtration effect was observed, evidencing the presence of quite large (> 30 kD – < 0.45µm), but also very small inorganic colloids (< 30 kD). In the presence of NOM, the Eu-concentrations increased with increasing HA concentration and a similar filtration effect as in the NOM-free experiments was observed, revealing that Eu(III)-HA complexes and/or colloids may (also) lead to higher "apparent" solubilties.

Role of (organic) colloids/polynuclear species By analogy with Eu and as suggested by the solubility experiments (see above), it is assumed that inorganic and organic colloids may lead to "apparent" solubilities of AmOHCO3, which are two to three orders higher than the thermodynamic solubility. As stated by Kim (1991), Am(III) colloids are however much less stable compared to tetravalent Th and Pu colloids, and prone to sorption on solid surfaces, such as experimental vessels. The effect of Am(III)-NOM association/colloid formation on the solubility and transport behaviour of Am has also been put forward in the newly developed phenomenological model (Bruggeman et al., in prep.).

Solubility data provided to PA Source Range (SR) All phases comprised in Table 2 were included in the source range.

Best Estimate (BE)

The selection of AmOHCO3(am,hyd) as most relevant solubility limiting phase (BE) is mainly based on experimental data, i.e. the above described solubility experiments performed within the frame of the TRANCOM II project (Maes et al., 2004) and also by Liu et al. (2008). With respect to the former experiments, also the analogy approach is called in to justify the phase selection. As already discussed above, the calculation of Pourbaix diagrams together with the solubility calculations enable to anticipate a formation sequence of the minerals taking into account "indirectly" also precipitation kinetics. Although Am(OH)3(am) is predicted to be kinetically the most probable phase, experimental studies have shown that this phase - in presence of carbonates - seems to convert quite fast to the mixed hydroxy-carbonate solid. Therefore, rather the latter solid is considered to represent the solubility controlling phase of the expert range.

Expert range (ER)

The expert range (ER) is represented by the solubility of the BE, i.e. AmOHCO3(am,hyd) including its thermodynamic uncertainties.

Data retained in the DCF’s: -6 BE: Thermodynamic solubility of AmCO3OH(am,hyd): 3.82×10 M. ER: BE including thermodynamic uncertainty.

SR: Lower limit (LL): Thermodynamic solubility of AmCO3OH × 0.5 H2O(cr): 2.4 x 10-8 M. -3 Upper limit (UL): Thermodynamic solubility of Am(OH)3(am): 1.25 × 10 M.

Eh/pH/pCO2 sensitivity of the solid phases

Guillaumont et al. (2005) and Duro et al. (2006) showed that the solubilities of Am(OH)3(am) and Am(OH)3(cr) are highly pH dependent with a clear decreasing trend as pH increases. The solubility (and stability) of Am-carbonates and mixed Am-hydroxy-carbonates is of course strongly dependent on the partial pressure of CO2. For further information it is referred to publications of (Runde et al., 1992; Runde et al., 2000b, Spahiu and Bruno, 1995).

Remarks The LLNL and NAGRA/PSI TDB’s comprise Am-data from Silva et al. (1995, NEA Vol. 2). In MOLDATA and the ANDRA TDB, mainly the updated americium data from NEA (2005, NEA Vol. 9) were incorporated. But also data from additional sources were considered. ++ The Am(III) bicarbonate complex AmHCO3 , which is covering a small Eh-pH space under acid conditions (see diagram Figure 2 b) is lacking in the LLNL and NARGRA/PSI TDB. In the latter database this bicarbonate species was not incorporated, as Silva et al. (1995) doubted about the existence of this complex and claimed further experimental work to prove its existence.

Am(CO3)1.5(cr) is the only solid comprised in MOLDATA that was copied from the NAGRA/PSI TDB. All other selected Am-solids in MOLDATA are from NEA.

5.3 Beryllium (Be) General Be belongs to the group of alkaline earth metals with the atomic number 4. Be is a scarce element on the Earth and in the universe. It is naturally occurring in combination with other elements in minerals and gemstones, such as bertrandite [Be4Si2O7(OH)2], beryl (Al2Be3Si6O18), chrysoberyl (Al2BeO4) and phenakite (Be2SiO4). Aquamarine and emerald are precious forms of beryl. Beryllium and its salts are toxic and it is corrosive to tissue and can cause a chronic life- threatening allergic disease called berylliosis. Beryllium has one stable isotope, i.e. 9Be and eleven known radioisotopes. Being one of the lightest known structural metals has contributed to Be being used in a wide variety of both nuclear and non-nuclear applications. Be, due to its unique combination of structural, chemical and neutron absorption cross-section characteristics has been successfully used in nuclear reactors as neutron reflector and neutron moderator (i.e. reduces the energy of neutrons). Be is also used in other nuclear applications, for example as "window" for x- rays and gamma rays. Be intimately mixed with high-energy alpha-emitters has also been successfully used to produce neutron sources. Be has been added to the safety relevant radioelements, as future fusion power plants will generate large quantities of neutron irradiated Be. At SCK•CEN, different research projects have been studying the Be behaviour and waste management issues, such as helium generation, irradiation induced swelling, vitrification of Be waste, corrosion tests of Be waste, creep properties of irradiated Be, waste acceptance criteria, etc. (Druyts and Van Iseghem, 2003; Sannen et al., 2002; Scibetta et al., 2007).

a) b)

1 1

.5 .5 Be++ Be++

Bromellite Eh (volts) Eh Eh (volts) Eh 0 -- 0 BeO2 -- BeO2 µ µ

–.5 –.5

25°C 25°C 02468101214 02468101214 pH pH Figure 3: Eh-pH diagram of beryllium (Be-C-S-O-H) for the BC reference porewater system. The assumed activity of dissolved [Be] = 10-8. Diagram a) LLNL / NEA / MOLDATA TDB, b) same diagram as a, but solid phases included in calculation. Code: The Geochemist's Workbench- 8.10.

Speciation and solubility calculations It can be seen in Figure 3, that the speciation of Be under BC conditions is dominated by the 2- negatively charged BeO2 species. The only solid Be-phase comprised in MOLDATA is bromellite (BeO). The solubility of the latter phase under BC conditions was calculated to correspond to 8.468 × 10-15 M (Table 4). The species distribution in equilibrium with bromellite 2- 2+ was calculated to correspond to 95% of BeO2 and 5% of Be .

Table 4: Solubility of Be in the BC reference porewater system at 25 °C, pH 8.355 and Eh -281 mV. Database: MOLDATA. Code: The Geochemist's Workbench- 8.12. Solubility controlling phases Solubility, [Be], mol/l Bromellite (BeO) 8.5 × 10-15 Source data: NEA

The reaction constant of BeO comprised in MOLDATA is: + 2+ BeO + 2 H  Be + H2O logK = 1.1237

Role of (organic) colloids/polynuclear species No information available.

Experimental data No in-house or BC specific solubility data available.

Data retained in the DCF’s: BE: Thermodynamic solubility of Bromellite: 3.0 × 10-15 M. ER: BE ± 1 order of magnitude (arbitrarily set uncertainty range). SR: BE ± 2 orders of magnitude (arbitrarily set uncertainty range).

Remarks NEA, Nagra/PSI TDB and ThermoChimie TDB do not comprise any thermodynamic data for Be.

5.4 Calcium (Ca) General Calcium belongs to the group of alkaline earth elements, has the atomic number 20 and is fifth- most-abundant element by mass in the Earth's crust, of which it forms more than 3%. Calcium is not naturally found in its elemental state. It occurs most commonly in sedimentary rocks in the minerals calcite, dolomite and gypsum. Ca is also comprised in igneous and metamorphic rocks, mainly in the silicate minerals, such as plagioclases, amphiboles, pyroxenes and garnets. Calcium from limestone is an important element in Portland cement. The solubility of the carbonate in water containing carbon dioxide causes the formation of caves with stalactites and stalagmites and is responsible for hardness in water (Handbook of Chemistry and Physics, 1992-1993). Calcium has four stable isotopes, i.e. 40Ca, 42Ca, 43Ca and 44Ca. Ninety-seven percent of naturally occurring Ca is 40Ca. 46Ca and 48Ca have such long half-lives that they are considered also stable. 41Ca is a cosmogenic isotope with a half-life of 103,000 years.

a) b)

1 1

.5 .5 ++ Ca++ Ca Eh (volts) Eh (volts) Eh 0 0 Calcite CaCO3 µ µ

–.5 –.5

25°C 25°C 02468101214 02468101214 pH pH Figure 4: Eh-pH diagram of calcium (Ca-C-S-O-H) for the BC reference porewater system. Activity of dissolved [Ca] = 1.35 × 10-5. Diagram a) MOLDATA TDB, same diagram as a, but solid phases included in caluclation. Code: The Geochemist's Workbench- 8.08.

Speciation and solubility calculations The Eh-pH diagram for calcium is shown in Figure 4. In the equilibrium model for the BC porewater chemistry, calcite represents the solubility controlling mineral for the calcium concentration in the porewater. Therefore, the cross symbol representing the BC reference conditions plots exactly on the boundary line between Ca2+ and calcite. The solubility of the latter corresponds to 5.1 × 10-5 under BC conditions (Table 5).

Table 5: Solubility of Ca in the BC reference porewater system at 25 °C, pH 8.355 and Eh -281 mV. Database: MOLDATA. Code: The Geochemist's Workbench- 8.08. Solubility controlling phases Solubility, [Ca], mol/l -5 Calcite (CaCO3) 5.1  10 Source data: NAPSI

The reaction constant of calcite comprised in MOLDATA is: + 2+ - CaCO3calcite + H ↔ Ca + HCO3 log K = 1.84

Role of (organic) colloids/polynuclear species Not relevant.

Experimental data No in-house or BC specific solubility data available.

Data retained in the DCF’s: BE: Thermodynamic solubility of calcite: 5.1 × 10-5 M. ER: BE including thermodynamic uncertainty. SR: Taken equal to ER, as no other solubility limiting phases are considered.

Remarks 2+ The basis species Ca in MOLDATA was selected from the NEA TDB, whereas CaCO3(aq) and + + CaHCO3 as well as calcite data were taken from the NAGRA/PSI TDB. CaCl was selected from ANDRA.

5.5 Caesium (Cs) General Caesium belongs to the group of alkali metals and its atomic number is 55. Caesium has more than 30 isotopes ranging from 112Cs to 137Cs. The only naturally occurring and stable isotope is 133Cs. 134Cs (half-life ~2 years) may be formed via neutron capture of the non-radioactive 133Cs or directly by the fission of uranium. The only HLW (vitrified waste and spent fuel) relevant caesium isotope is 135Cs. 135Cs is one of the 7 long-lived fission products (LLFP, half-life = 2.3 × 106 years) with 135Xe being its predecessor. The latter represents a neutron poison, meaning it is a strong absorber of thermal neutrons. Therefore, a large part of 135Xe is transmuted to the non-radioactive 136Xe before decaying to 135Cs. 137Cs represents a so-called medium lived fission product (MLFP) and is produced in around 6 % of fissions with a half-life of 30.23 years. During its beta decay to 137mBa, 137Cs is emitting strong gamma radiation and therefore represents a strong radioactive hazard. Caesium exists in the environment in the +1 oxidation state. The aqueous speciation of caesium is among the simplest of the elements/radionuclides described in this report. The dominant species in most groundwaters and also in BC porewater is the monovalent Cs+ cation. Caesium forms weak complexes with sulphate, halides and nitrate. Although Baes and Mesmer (1976) report, that Cs may be associated with OH- ions in solution, hydrolysis data for Cs are scarce. Data for the first hydrolysis species are comprised in ThermoChimie v.5 and have been copied to MOLDATA. As can be seen in Figure 5, the formation of inorganic complexes does not play a role over the entire/illustrated Eh-pH domain.

Cs+ Eh (volts) Eh 0

–.5

25°C 02468101214 pH

Figure 5: Eh-pH diagram of caesium (Cs-C-S-O-H) for the BC reference porewater system. Assumed activity of dissolved [Cs] = 10-8. Database: MOLDATA / NEA TDB. Code: The Geochemist's Workbench - 8.08.

Role of (organic) colloids/polynuclear species Not relevant.

Experimental data Solubility data are not relevant.

Solubility data provided to PA The caesium concentration is not solubility limited.

Remarks No further remarks.

5.6 Carbon (C) General Carbon belongs to group 14 of the periodic table and its atomic number is 6. Carbon is widely distributed in nature and carbon compounds form the basis of all known life on Earth. Carbon is present in three allotropic forms: amorphous, graphite, and diamond (Handbook of Chemistry and Physics, 1992-1993). Graphite is one of the softest known materials while diamond is one of the hardest. Graphite exists in two forms: alpha and beta. These have identical physical properties, except for their crystal structure. All carbon allotropes are solids under normal conditions with graphite being the most thermodynamically stable form. Carbon is non-metallic and the most common oxidation state of carbon in inorganic compounds is +4. In the tetravalent oxidation state, C is found as carbon dioxide in the atmosphere of the Earth and dissolved in all natural waters. In carbon monoxide and other transition metal carbonyl complexes, the oxidation state of C is +2. Inorganic carbon is a component of calcium and magnesium carbonates, such as limestones and dolomites, respectively. Significant carbon quantities occur in organic deposits of coal, peat, oil and methane clathrates. Carbon forms more compounds than any other element, with almost ten million pure organic compounds described to date (Wikipedia). Carbon has seven isotopes, of which three are naturally occurring, i.e. 12C (98.9%) and 13C (1.1%), representing stable isotopes and 14C being radioactive with a half-life of about 5,700 years. The latter is mainly produced by cosmic rays and in nuclear reactors through neutron activation of 14N, which is comprised in oxide fuels, coolant water and structural materials. A detailed overview about 14C generation in nuclear power plants is given by Wang et al. (2009).

a) b) 1 1

.5 .5 CO2(aq) CO2(aq)

HCO- 3 HCO- Eh (volts) Eh Eh (volts) 3 0 -- 0 CO3 -- CO3 Methane µ µ

–.5 –.5

25°C 25°C 02468101214 02468101214 pH pH Figure 6: Eh-pH diagram of carbon (C-S-O-H) for the BC reference porewater system. - -2 Assumed activity of [HCO3 ] = 1.27 × 10 . Diagram a) LLNL / NAGRA/PSI / MOLDATA TDB, b) ANDRA / NEA TDB. Code: The Geochemist's Workbench- 8.08.

Speciation calculations

The carbonate concentration in the reference porewater is imposed by a partial pressure of CO2 (g) -2.437 - = 10 atm at pH 8.355 and 25 °C, which corresponds to a HCO3 concentration of 0.014 M (log activity = -1.8954). Figure 6 reveals that the bicarbonate species is stable up to a pH of ~10. In more alkaline conditions (pH >10), the carbonate ion becomes the dominant species. As more reducing conditions are encountered, and very close to the sulphide-sulphate boundary C(IV) is reduced to C(-II), with aqueous methane as important species.

Role of (organic) colloids/polynuclear species Not relevant.

Experimental data No in-house or BC specific solubility data available.

Solubility data provided to PA The carbon concentration is not solubility limited.

Remarks - 2- The following species incorporated in MOLDATA were taken from NEA: HCO3 and CO3 . The - mixed iron-carbonate hydroxy-species (FeCO3OH and FeCO3OH ) were selected from the ANDRA TDB. Methane data are from the NAGRA/PSI TDB.

5.7 Chorine (Cl) General Chlorine belongs to the group of halogen elements and its atomic number is 17. Under standard state conditions (25°C, 1 atm), two chlorine atoms form the diatomic molecule Cl2. Chlorine has more than 20 isotopes and two principal stable isotopes, i.e. 35Cl (75.76%) and 37Cl (24.24%). The only radioactive waste relevant isotope is 36Cl, which is a beta-emitter with a half-life of 3 × 105 years. 36Cl represents a long-lived activation product, which is produced via neutron activation of 35Cl. Most solid chlorides are soluble in water, therefore they are usually only found in dry climates, or deep underground. Common chloride minerals include halite (NaCl), sylvite (KCl), and carnallite (KMgCl3 × 6 H2O).

Cl- Eh (volts) Eh 0

–.5

25°C 02468101214 pH

Figure 7: Eh-pH diagram of chlorine (Cl-C-S-O-H) for the BC reference porewater system. Activity of dissolved [Cl] = 6.43 × 10-4. MOLDATA/NEA TDB. Code: The Geochemist's Workbench- 8.08.

Speciation calculations In aqueous media, the speciation of chlorine is generally dominated by the chloride anion, which is also the case under BC conditions.

Role of (organic) colloids/polynuclear species Not relevant.

Experimental data Solubility data are not relevant.

Solubility data provided to PA The chlorine concentration is not solubility limited.

Remarks No further remarks.

5.8 Curium (Cm) General Curium belongs to the series of actinides, more specifically the transuranic elements and has an atomic number of 96. Fourteen isotopes of Cm are known and they are all highly radioactive. 247Cm represents the most stable isotope with a half-life of 15.6 million years. Other long-lived isotopes are 248Cm (half-life: 3.40 × 105 years), 245Cm (half-life 8.51 × 103 years), 250Cm (8.01 ×103 years) and 246Cm (4.73 × 103 years). Another isotope comprised in the radioactive waste inventory is 244Cm, which is an alpha-emitter with a half-life of 18 years. Curium does not occur naturally, most of the curium isotopes are produced in nuclear reactors via neutron bombardment of uranium or plutonium. In solid compounds, curium usually exhibits valence +III and sometimes +IV, while in aqueous solutions the +III valence is dominant.

Solubility data provided to PA Taken to be identical to Am.

Data retained in the DCF’s BE/SR/ER: Taken to be identical to Am.

Remarks Curium data are not available in the present MOLDATA TDB. Although the curium data are sparse, there is a consensus that curium is a chemical analogue of americium (Bruggeman et al., 2012). The commonly cited solubility controlling phase in carbonate rich waters is CmOHCO3(cr), an analogue of AmOHCO3(cr) (Berner, 1995; Bruno et al., 1997).

5.9 Iodine (I) General Iodine belongs to the group of halogens and its atomic number is 53. Iodine has more than 30 isotopes, of which only one is stable, i.e. 127I. 131I is one of the most important radioactive isotopes of iodine (also named radio-iodine) as it represents one of the most hazardous short-lived (half-life of around 8 days) fission products. 131I can be released during nuclear weapon tests or nuclear accidents (e.g. Fukushima) due to fission of 235U (fission yield of 2.878 %), and as such is known to be able to induce significant health effects, such as thyroid cancer. The radiological inventory for category B and C waste however, comprises only the long-lived (half-life: ~1.61 × 107 years) 129I isotope. The latter represents one of the seven long-lived fission products besides 99Tc, 126Sn, 79Se, 93Zr, 135Cs and 107Pd. Iodine is rare in the solar system and its concentration in the Earth's crust is around 0.5 ppm. Due to its high volatility, iodine has accumulated in the outer envelopes of the Earth. Consequently, sedimentary rocks contain more iodine than magmatic and metamorphic rocks (Becker et al., 1972). Chile is the world’s leading iodine producing nation. Caliche is the name of the sedimentary rock (CaCO3) from which the iodine is retrieved. Iodine minerals are very rare and iodide salts often very soluble in water. Iodine concentrations in seawater range between 47–60 µg/L. For further details concerning the iodine geochemistry and its retention and migration behaviour in Boom Clay, it is referred to the Iodine Topical Report (Bruggeman et al., 2010).

a) HIO3(aq) b) - - 1 I3 1 ICl2

- - IO3 IO3

.5 .5

I- I- Eh (volts) Eh Eh (volts) Eh 0 0

µ µ

–.5 –.5

25°C 25°C 0 2 4 6 8 10 12 14 02468101214 pH pH

d) HIO3(aq) c) 1 1

- IO3

.5 .5

I2 (volts) I- Eh

Eh (volts) 0 0

µ µ –.5 –.5 - I 25°C 25°C 0 2 4 6 8 10 12 14 02468101214 pH pH

e) f) HIO3(aq) 1 - 1 ICl2

- IO3 .5 .5

- Eh (volts) Eh I 0 (volts) Eh 0

µ µ –.5 - I –.5 25°C 25°C 0 2 4 6 8 10 12 14 02468101214 pH pH Figure 8: Eh-pH diagram of iodine (I-C-S-O-H) for the BC reference porewater system. Assumed activity of dissolved [I] = 10-8. Diagram a) LLNL TDB, b) ANDRA TDB, c=e) NAGRA/PSI TDB, d) NEA TDB, e) MOLDATA TDB, f) same diagram as e, but I2(aq) suppressed in calculation. Code: The Geochemist's Workbench - 8.08.

Speciation calculations Iodine is a redox-sensitive element with oxidation states ranging between –1 and +7, but the - environmentally most important are the zero (elemental), -1 (iodide) and +5 (IO3 ; iodate) valence states. Under BC conditions, the iodide anion is predicted to be the stable iodine species when using the LLNL, ANDRA and NEA TDB’s for calculating the Pourbaix diagrams (see Figure 8 a, b, d). In contrast, using the NAGRA/PSI and MOLDATA, molecular iodine represents the prevalent iodine species (Figure 8 c, e). This is due to the fact, that on the one hand, the aqueous I2 species, which was copied from the NAGRA/PSI TDB to MOLDATA is not included in the other TDB’s and that in the database from NAGRA/PSI, I2(aq) is thermodynamically more stable than I. It is debatable, if the former species should be retained in the speciation calculations. Suppressing I2(aq) in the calculations, leads to a similar diagram (see Figure 8 f) as with the other TDB’s where iodide is predicted as dominant species under BC conditions. According to Bruggeman et al. (2010), the natural inventory of iodine in BC is composed of both I- and I(0) species with the latter having been shown to be able to bound covalently to phenolic moieties present in natural organic matter (NOM). As there is no reason to discard molecular iodine from the database, it has been retained according to the strategy of MOLDATA to strive for data completeness. The calculations of the Pourbaix diagrams and the current knowledge about the iodine geochemistry suggest however that a more detailed evaluation of the iodine thermodynamic data should be performed. It should be mentioned, that the diatomic iodine molecule I2 reacts reversibly with the iodide, − generating the triiodide anion I3 , which is soluble in water.

Role of (organic) colloids/polynuclear species Iodine (elemental) interaction with humic substances and organic iodine compounds have been described by many authors (Mercier et al., 2000; Richard and Gaona, 2011; Christiansen and Carlsen, 1989). As revealed by field observations and geochemical calculations iodine may be fixed by organic matter only under oxidizing conditions. Thus, under the (reducing) reference conditions, where iodine occurs as iodide, I-NOM interaction/colloid formation is not expected (Bruggeman et al., 2010).

Experimental data Solubility data are not relevant.

Solubility data provided to PA The iodine concentration is not solubility limited.

Remarks - - Iodine species incorporated in MOLDATA comprise the basis species I , IO3 and HIO3(aq) - - (NEA), I2 (aq) (NAGRA/PSI), IO (LLNL), as well as IO4 (ANDRA).

5.10 Molybdenum (Mo) General Molybdenum belongs to the group of transition metals and its atomic number is 42. Nowadays 35 isotopes of molybdenum are known with their atomic mass ranging from 83 to 117. There exist seven naturally occurring isotopes, i.e. 92Mo, 94Mo, 95Mo, 96Mo, 97Mo, 98Mo and 100Mo, with 92Mo and 100Mo being unstable. The unstable isotopes decay into isotopes of niobium, technetium and ruthenium. 98Mo is the most abundant isotope (24.14 %) with a half-life of about 1019 years. Molybdenum does not occur as pure metal in nature, but rather in uranium ores or in various oxidation states in different minerals. Most of these minerals are characterized by a low solubility. The waste relevant molybdenum isotope is 99Mo, which represents a fission product. In this context, molybdenum is important due to the potential formation of a variety of solid molybdates in spent nuclear fuel and their role in the retention of some other fission products. The most soluble ones are formed with alkaline cations (Cs and Rb) and more stable ones formed with alkaline earth metals, of which calcium molybdate is likely to control the solubility of molybdenum in the hyperalkaline waters resulting of the degradation of cement (Andra, 2005).

a) b) 1 1

.5 .5

-- +++ MoO4 Mo Eh (volts) Eh Eh (volts) Eh 0 0

µ µ

–.5 –.5

25°C 25°C 02468101214 0 2 4 6 8 10 12 14 pH pH

c) d) 1 1

H2MoO4(aq) H MoO (aq) - .5 2 4 HMoO4 .5 - HMoO4

Mo3O8 -- -- MoO (s) MoO Eh (volts) Eh MoO 2 4 0 +++ 4 (volts) Eh Mo 0

MoS2(s) µ µ –.5 –.5

25°C 25°C 0 2 4 6 8 101214 02468101214 pH pH

e) 1

H2MoO4(s) - HMoO4

Mo3O8

-- MoO4 - no graph - Eh (volts) Eh 0

MoO2(s) µ

–.5

25°C 02468101214 pH

Figure 9: Eh-pH diagram of molybdenum (Mo-C-S-O-H) for the BC reference porewater system. Assumed activity of dissolved [Mo] = 10-8. Diagram a) LLNL and NAGRA/PSI TDB, b) ANDRA TDB, c) MOLDATA TDB, d) same diagram as c, butsolid phases included (MoS3,s suppressed in calculation), -5 e) same diagram as d, but [Mo] = 10 (MoS2,s and MoS3,s suppressed in calculation). Code: The Geochemist's Workbench - 8.10-8.12.

Speciation and solubility calculations + 2- Molybdenum species in natural waters occur as various Mo(V, VI) oxyions (e.g. MoO2 , MoO4 ). According to the speciation calculations performed with MOLDATA (Figure 9 c), the molybdate 2- anion MoO4 represents the dominant aqueous species under BC conditions. Furthermore, it can be seen that Mo is very soluble and mobile at higher Eh conditions. Solubility calculations were performed for elemental molybdenum, Mo-sulfide, different -8 molybdenum oxides and Ca-molybdate. At [Mo] = 10 , MoS2(s) is predicted to be the most stable solid. As can be seen in Table 6, the solubility of this phase is very low (2 × 10-13 M) under BC conditions. At higher Mo-concentrations, i.e.  ~2 × 10-6 M, Mo-oxide (tugarinovite) becomes stable and represents the kinetically more likely phase to form. Elemental Mo, molybdite and ilsemannite are characterized by very high solubilities under BC conditions. As can be seen in Figure 9 d, tha latter two oxides are stabilized under more acidic and oxidizing conditions.

Table 6: Solubility of Mo in the BC reference porewater system at pH 8.355 and Eh -281 mV. Database: MOLDATA. Code: The Geochemist's Workbench - 8.08. Solubility controlling phases Solubility, [C], mol/l -4 Calcium molybdate (CaMoO4,s) 8.9 × 10 -6 Tugarinovite (MoO2,s) 1.7 × 10 -13 MoS2 (s) 2.0 × 10

Molybdite (MoO3) very soluble

Ilsemannite (Mo3O8,s) very soluble Mo(s) very soluble Source data: all solid phase data were taken from the ANDRA TDB

The reaction constants of the Mo solids comprised in Table 6 and MOLDATA are the following: 2- 2+ CaMoO4(s) ↔ MoO4 + Ca logK = -7.9 2- + MoO2(s) + H2O + 0.5 O2(aq) ↔ MoO4 + 2 H logK = 13.12

2- 2- + MoS2(s) + 3 H2O + 4.5 O2(aq) ↔ MoO4 + 2 SO4 + 6 H logK = 249.64 2- MoO3(s) + H2O ↔ MoO4 + 2 H+ logK = -11.98 2- Mo3O8(s) + 3 H2O + 0.5 O2(aq) ↔ 3 MoO4 + 6 H+ logK = -20.57 2- 2- + MoS3(s) + 4 H2O + 6 O2(aq) ↔ MoO4 + 3 SO4 + 8 H logK = 346.77

Role of (organic) colloids/polynuclear species No information available.

Experimental data No in-house or BC specific data available.

Solubility data provided to PA Source Range (SR) Solubility calculations revealed that the Mo-oxides molybdite and ilsemannite, as well as elemental Mo are very soluble under BC conditions and can be excluded as solubility limiting phases with respect to PA considerations. Calcium molybdate has been reported to represent a potential solubility limiting phase for Mo, but rather in hyperalkaline waters resulting from cement degradation (Andra, 2005). Therefore, it can be also excluded as solubility controlling solid under undisturbed BC conditions. Thus, only tugarinovite and MoS2(s) have been retained in the source range of potential solubility limiting solids.

Best Estimate (BE)

MoO2(s) is considered to represent the most probable solid (ER) controlling the aqueous Mo concentration, if they exceed 10-6 M.

Expert range (ER)

The expert range (ER) is represented by the solubility of the BE, i.e. MoO2(s) including its thermodynamic uncertainties.

Data retained in the DCF’s: -6 BE: Thermodynamic solubility of MoO2(s): 1.7 × 10 M. ER: BE including thermodynamic uncertainty. -13 SR: Lower limit (LL): Thermodynamic solubility of MoS2(s): 2.0 x 10 M. Upper limit (UL): Equal to BE.

Eh/pH/pCO2 sensitivity of the solid phases

The stability of MoO2(s) is reported by ANDRA (i.e. Dossier 5) to be very Eh and pH sensitive.

Remarks The NEA TDB does not contain data on Mo. Therefore, Mo species included in MOLDATA were selected mainly from the ANDRA TDB. The basis species in MOLDATA is the molybdate anion 2- MoO4 .

Mo has a similar ionic radius as Fe3+ (0.064Å), due to which it may replace some Fe in the structure of Fe minerals. Molybdenum replacement of Fe in Fe oxides is highly dependent on the pH of the system (Goldberg et al. 1996).

5.11 Neptunium (Np) General Neptunium belongs to the group of actinides and with its atomic number of 93 it represents the first transuranic element. The isotope 237Np (half-life of 2.14 × 106 years) is obtained as a by- product from nuclear reactors in the production of plutonium. Trace quantities of the element are actually found in nature due to transmutation reactions in uranium ores produced by the neutrons which are present (Handbook of Chemistry and Physics, 1992-1993). In total, 19 neptunium radioisotopes have been characterized, with the most stable being 237Np with a half-life of 2.14 million years, 236Np with a half-life of 154,000 years, and 235Np with a half-life of 396.1 days. All of the remaining radioactive isotopes have half-lives less than 4.5 days (Wikipedia). Np exists in five oxidation states, i.e. + III, +IV, +V, +VI and +VII, but only the tetravalent and pentavalent states are important in natural waters (Langmuir, 1997).

++ a) b) NpO2 1 1

-- (NpO2)(CO3)2 + ---- NpO + NpO (CO ) 2 (NpO )(CO )---- NpO2 2 3 3 2 3 3 .5 .5 (NpO )(CO )(OH)-- Np(OH)+++ 2 3 2 ++++ - Np NpO2CO3 - (NpO2)(CO3) +++ --- ++ NpOH NpO (CO ) Np(OH)2 2 3 2 NpO (CO ) OH---- ++ NpO2OH(aq) 2 3 2 Np(OH)2 Eh (volts)

Eh (volts) +++ 0 +++ Np(OH)+ 0 Np + Np 3 Np(OH)3

Np(OH) Np(OH)µ 4(aq) µ 4

–.5 –.5 25°C 25°C 02468101214 02468101214 pH pH

++ ++ NpO2 d) NpO c) 2 1 1

NpO (CO )-- -- 2 3 2 NpO2(CO3)2 + + ---- NpO NpO NpO2(CO3)3 2 ---- 2 NpO2(CO3)3 .5 .5 ) - +++ ++ ts NpOH NpSO NpO2(OH)3 l ++++4 - - Np NpO CO NpO2CO3 2 3 vo

+++ ( NpOH ++ NpO (CO ) OH---- Np(OH)2 2 3 2 Eh (volts) 0 +++ ---- Eh +++ Np NpO2(CO3)2OH 0 Np

Np(OH) (aq) µNp(OH)4 µ 4

–.5 –.5 25°C 25°C 0 2 4 6 8 10 12 14 02468101214 pH pH

++ ++ NpO2 NpO2 f) e) 1 1 + NpO2 -- -- NpO2(CO3)2 NpO2(CO3)2 - + NpO2CO3 ---- NpO2 ---- NpO2(CO3)3 NpO2(CO3)3 .5 .5 - +++ NpO2(OH)3 NpOH - -- - NpO2(OH)3 NpO2(OH)4 NpO CO -- ++ 2 3 NpO (OH) Np(OH)2 2 4

Eh (volts) Eh +++ NpO (cr) Eh (volts) Eh ---- 2 +++ NpO (CO ) OH 0 Np 0 Np + 2 3 2 Np(OH)3 µ µNp(OH)4(aq)

–.5 –.5

25°C 25°C 02468101214 0 2 4 6 8 10 12 14 pH pH

++ g) NpO2 1

NpO (CO )-- + 2 3 2 NpO2 ---- NpO2(CO3)3 .5 +++ NpOH - - NpO (OH) NpO2CO3 2 3 ++ NpO (OH)-- Np(OH)2 2 4 ---- - no graph - Eh (volts) Eh NpO (CO ) OH 0 +++ + 2 3 2 Np Np(OH)3

NpOµ 2(am,hyd)

–.5

25°C 0 2 4 6 8 101214 pH

Figure 10: Eh-pH diagram of neptunium (Np-C-S-O-H) for the BC reference porewater system. Assumed activity of dissolved Np = 10-8. Diagram a) LLNL TDB, b) ANDRA TDB, c) NAGRA/PSI TDB, d) NEA TDB, e) MOLDATA TDB, f) same diagram as e, but stability fields of minerals/solid phases included, g) same diagram as f, but NpO2(cr) suppressed in calculations. Code: The Geochemist's Workbench - 8.08.

Speciation and solubility calculations According to MOLDATA (and also the other TDB’s), the Np(IV) hydrolysis species Np(OH)4(aq) is predicted to be the dominant aqueous species under undisturbed Boom Clay conditions (Figure 10 e). Including solid phases into the calculations, shows that NpO2(cr) occupies a very large stability field and represents the most stable phase under BC conditions (Figure 10). However, considering kinetic aspects, the amorphous, hydrated neptunium oxide (NpO2,am,hyd) would most probably precipitate first and with time/ageing transform into the more stable and crystalline NpO2. The Np(V) solid Np2O5(cr) is only stable under more oxidizing, i.e. atmospheric pO2(g) conditions (Duro et al., 2006). The Np(V)O2OH phases are characterized by very high solubilites under the reference conditions. They are reported to become more stable under more alkaline and oxidizing conditions (Berner, 2002).

Table 7: Solubility of Np in the BC reference porewater system at 25°C, pH 8.355 and Eh -281 mV. Database: MOLDATA. Code: The Geochemist's Workbench - 8.08. Solubility controlling phases Solubility, [Np], mol/l -18 NpO2 (cr) 1.8  10 -9 NpO2 (am,hyd) 2.0 × 10

NpO2OH (am, fresh) very soluble

NpO2OH(am, aged) very soluble Source data: all solid phase data were taken from the NEA TDB. Remark: NpO2 × 2 H2O is thermodynamically equivalent to NpO2(am,hyd) and was thus not considered in the calculations.

The reaction constants comprised in MOLDATA of the minerals mentioned in Table 7 are: + 4+ NpO2(cr) + 4 H ↔ Np + 2 H2O log K = -9.75 + 4+ NpO2 (am,hyd) + 4 H ↔ Np + 2 H2O log K = -0.70 + 4+ NpO2OH (am, fresh) + 4 H ↔ Np + 2.5 H2O + 0.25 O2(aq) log K = -5.98 + 4+ NpO2OH (am, aged) + 4 H ↔ Np + 2.5 H2O + 0.25 O2(aq) log K = -6.58

Experimental data No in-house or BC specific data available.

Role of (organic) colloids/polynuclear species As the other tetravalent actinides, Np(IV) - due to its high effective charge - has a strong tendency to hydrolyse and to form strong aqueous complexes. As can be seen in Figure 10 e, hydrolysis of Np starts at very acidic pH. At higher concentrations, Np(IV) may undergo oligomerization and generate very stable intrinsic colloids (Runde et al., 2000a; Guillaumont et al., 2005; p. 294). For further details, it is referred to paragraph 4.3.8.

Np-NOM colloid association has been put forward in the newly developed phenomenological model (Bruggeman et al., in prep.).

Solubility data provided to PA Source Range (SR) Neptunium concentrations under BC conditions are expected to be controlled by Np(IV) solids, i.e. NpO2(cr) or NpO2(am,hyd), respectively. As can be seen in Figure 10 f and g, these solids are stable over a quite large Eh (and pH) range, from very reducing to slightly oxidizing conditions. Based on Ostwald's rule, the amorphous phase is kinetically favoured and therefore should precipitate first. On the long-term, this phase may convert to the more stable crystalline solid. The Np-hydroxides, i.e. NpO2OH (am, fresh) and NpO2OH(am, aged) were not retained due to their high solubility under BC conditions.

Best Estimate (BE)

Based on the reasoning above, NpO2(am,hyd) is considered as the most probable solubility limiting solid for Np.

Expert Range (ER)

The expert range (ER) is represented by the solubility of the BE, i.e. NpO2 (am,hyd) including its thermodynamic uncertainties.

Data retained in the DCF’s: -9 BE: Thermodynamic solubility of NpO2(am,hyd): 2.0 × 10 M. ER: BE including its thermodynamic uncertainty. -18 SR: Lower limit (LL): Thermodynamic solubility of NpO2(cr): 1.8 × 10 M. Upper limit (UL): Equal to ER.

Remarks The Np data incorporated in the NAGRA/PSI TDB were taken from the NEA review Vol. 4 (Lemire et al., 2001). The Np data comprised in the LLNL TDB are even older, dating from Lemire et al. (1984). In MOLDATA, the most recent NEA data from Guillaumont et al. (2005, Vol. 5) as well as Np-data from ThermoChimie v.5 were incorporated.

Since the review performed by Lemire et al. (2001), no new experimental data for NpO2(cr) have appeared and all thermodynamic quantities are well established. While in the review of Lemire et al. (2001) a ∆fG°m value for NpO2(am,hyd) was selected, the authors of the recent review were reluctant in selecting a value. The reason for this is, that the selection strategy by NEA with respect to amorphous solids has changed within the last years and this solid is considered to be chemically not well (enough)-defined, meaning "it could be a hydroxide, amorphous hydrous oxide or oxihydroxide" (p. 298) and thus, only reaction data are given (p. 88).

Trivalent aqueous Np species are only stable in the presence of strong reductants (but easily oxidized by air), due to which they can be neglected under BC conditions. For completeness they have been incorporated into MOLDATA. Trivalent Np species in MOLDATA comprise the 2+ + + aqueous hydroxides Np(OH) , Np(OH)2 , Np(OH)3, as well as the carbonate species Np(CO3) - 3- 2+ 3- Np(CO3)2 and Np(CO3)3 . The species Np(OH) and Np(CO3)3 represent data that were selected from Lemire et al. (2001). The higher hydrolysis products were taken from Allard et al. (1980) and the remaining data represent estimated values. For more details, also with respect to trivalent solid phases, it is referred to the documentation of the ThermoChimie v.5.

5.12 Nickel (Ni) General Nickel is a silvery-white transition metal with an atomic number of 28. Nickel has five stable isotopes with 58Ni being the most prevalent one and representing two thirds of natural nickel. The other four stable isotopes are 60Ni (26 %), 61Ni (1.1 %), 62Ni (3.6 %) and 64Ni (0.9 %). Concerning radioactive isotopes, only two, i.e. 59Ni and 63Ni are of potential health and environmental concern. Both are produced by neutron activation of nuclear reactor components, such as various steel alloys and/or the spent fuel hardware. Therefore, these isotopes are present in wastes resulting from reprocessing of spent nuclear fuel. 59Ni decays with a half-life of 75,000 years by electron capture and 63Ni decays with a half-life of 96 years by emitting a beta-particle. In nature, Ni is present in various ores, such as niccolite (nickel arsenide, NiAs), garnierite (mixture of the Ni-Mg-hydrosilicates serpentine, talc, sepiolite, chlorite and smectite), pentlandite [(Fe,Ni)9S8] and to a lesser extent in soils. Nickel minerals comprise gaspéite (NiCO3), millerite (β-NiS), vaesite (NiS2), pentlandite [(Fe,Ni)9S8], violarite (FeNi2S4) and trevorite (NiFe2O4). Besides this, Ni is often found in meteorites. Ni is redox-sensitive and occurs in several oxidation states, i.e. 0, +II, and +IV. However, the tetravalent oxidation state is not stable in solution, it only appears with a formal +IV valence in some solid phases (ThermoChimie documentation).

a) b) 1 1

.5 .5 Ni++ Ni++ + Ni(HCONi(CO3) 3) Eh (volts)

Eh (volts) Eh -- 0 0 Ni(CO3)2 Ni(OH)2(aq) - - Ni(OH)3 Ni(OH)3

µ Ni(HS)2(aq)µ

–.5 –.5

25°C 25°C 02468101214 02468101214 pH pH d) c) 1 1

.5 .5 ++ Ni Ni++

NiCO3 NiCO3(aq) Eh (volts) Eh Eh (volts) Eh 0 0 - Ni(OH)2 Ni(OH)3 - Ni(OH)3 µ NiHS+µ Ni(HS)2 –.5 –.5 Ni(l) 25°C 25°C 0 2 4 6 8 10 12 14 02468101214 pH pH

e) f) 1 1

.5 .5 ++ Ni++ Ni + Ni(HCO3) -- Trevorite Ni(CO ) (volts) Eh Eh (volts) Eh 3 2 0 + 0 Ni(HCO ) - 3 Ni(OH)3 -- Ni(CO3)2 NiS(beta) µ Ni(HS)2(aq)µ Violarite –.5 –.5 Ni(l) 25°C 25°C 0 2 4 6 8 10 12 14 0 2 4 6 8 101214 pH pH

g) h) 1 1

.5 .5 ++ Ni Ni++

Trevorite Trevorite Eh (volts) Eh Eh (volts) Eh 0 + 0 Ni(HCO3) Ni(CO )-- 3 2 NiCO3(cr) NiS(beta)µ µ –.5 –.5 Ni(cr) 25°C 25°C 0 2 4 6 8 10 12 14 02468101214 pH pH

i) j) 1 1

.5 .5 ++ ++ Ni3O4(s) Ni Ni3O4(s) Ni

+ Ni(HCO3) NiCO3(cr) Eh (volts) Eh Eh (volts) Eh 0 -- 0 Ni(CO3)2 Ni(OH)2(beta) Ni(OH)2(beta) µ Ni(HS)2(aq)µ

–.5 –.5 Ni(cr) Ni(cr) 25°C 25°C 0 2 4 6 8 10 12 14 0 2 4 6 8 10 12 14 pH pH Figure 11: Eh-pH diagram of nickel (Ni-C-S-O-H) for the BC reference porewater system. Assumed dissolved activity of [Ni] = 10-8. Diagram a) LLNL TDB, b) ANDRA TDB, c) NAGRA/PSI TDB, d) NEA TDB, e) MOLDATA TDB, f) same diagram as e, but solid phases included (high-T sulphides suppressed in calculation), g) same as diagram f, but violarite suppressed in calculation, h) same diagram as g, but [Ni] = 10-4 and NiS(beta) suppressed in calculation, i) same diagram as g, but [Ni] = 10-4, but trevorite suppressed in calculation, j) -3 same diagram as I, but [Ni] = 10 and NiCO3(s) suppressed in calculation. Code: The Geochemist's Workbench - 8.08.

Speciation and solubility calculations Speciation diagrams for Ni as function of pH and Eh were calculated using the individual thermodynamic databases as well as MOLDATA and the results are illustrated in Figure 11 (a-f). Assuming a nickel activity [Ni] = 10-8, the aqueous speciation under BC conditions is calculated 2- (MOLDATA TDB) to be dominated by carbonate complexes of the composition Ni(CO3)2 .

Solubility calculations reveal that Ni-oxide (NiO,cr), Ni-hydroxide [β-Ni(OH)2], Ni-silicate (Ni2SiO4), as well as the Ni-carbonates NiCO3(cr) and NiCO3 × 5.5 H2O(cr) are characterized by quite high solubilities under BC conditions, while the sulfides millerite (β-NiS), violarite -8 (FeNi2S4) and vaesite (NiS2, cr) are less soluble (Table 8). At [Ni] = 10 , only violarite would be oversaturated and able to precipitate under BC conditions. At slightly higher Ni-activitites, Millerite (β-NiS) would become (over)saturated and could control the aqueous Ni-concentration. -4 At Ni-activities  10 , trevorite and gaspéite (NiCO3,cr) are stabilized (Figure 11 h). Trevorite (NiFe2O4) represents a solid solution type of mineral belonging to the spinel group of minerals, which according to Bruno et al. (2001 b) are generally not formed at low temperatures. Suppressing trevorite in the calculation, leaves NiCO3(cr) as solubility controlling solid. In Table 9, the aqueous species distribution in equilibrium with gaspéite is given. As already mentioned 2- above, Ni(CO3)2 dominates Ni-speciation under the reference conditions. At an even higher Ni- -3 activity (10 ), Ni(OH)2(beta) appears in the diagram when the Ni-carbonate is suppressed in the calculation. It should be mentioned however, that such high Ni-activity seems very unprobable in the far-field and would be similar to the activity of the major cations and anions. Olivine was included for completeness in the calculations, but is considered to be not a relevant solubility limiting phase. Bunsenite is characterized by a very high solubility under BC conditions. The same is the case for NiCO3 × 5.5 H2O(cr). Therefore, the former two phases were not considered in the evaluation of potential solubility limiting phases.

Table 8: Solubility of Ni in the BC reference porewater system at 25 °C, pH 8.355 and Eh -281 mV. Database: MOLDATA. Code: The Geochemist's Workbench- 8.08. Solubility controlling phases Solubility, [Ni], mol/l -2 Bunsenite NiO (cr)1 3.0  10 1 -3 Ni2SiO4 (oliv) 2.7  10 1 -4 Nickel hydroxide (Ni[OH]2,beta) 7.2 × 10 2 -5 Trevorite (NiFe2O4) 1.8 × 10 1 -5 Gaspéite (NiCO3,cr) 2.6 × 10 1 -7 Vaesite NiS2 (cr) 6.2 × 10 Millerite (β-NiS)1 1.8  10-8 2 -9 Violarite (FeNi2S4) 9.5 × 10 1 NiCO3 × 5.5 H2O(cr) very soluble Source data: 1NEA TDB, 2ANDRA TDB

The reaction constants of the Ni solids comprised in Table 8 and MOLDATA are the following: + 2+ NiO(cr) + 2 H ↔ Ni + H2O log K = 12.48 + 2+ Ni2SiO4 + 4 H ↔ 2 Ni + Si(OH)4(aq) log K = 19.42 + 2+ β-Ni(OH)2 + 2 H ↔ Ni + 2 H2O log K = 11.03 + 2+ 2+ NiFe2O4 + 6 H ↔ Ni + 2 Fe + 3 H2O + 0.5 O2(aq) log K = -7.52 + 2+ - NiCO3(cr) + H ↔ Ni + HCO3 log K = -0.67 2+ 2- + NiS2(cr) + 3.5 O2(aq) + H2O ↔ Ni + 2 SO4 + 2 H log K = 215.60

2+ 2- β-NiS + 2 O2(aq) ↔ Ni + SO4 log K = 128.15 2+ 2- + 2+ FeNi2S4 + H2O + 7.5 O2(aq) ↔ 2 Ni + 4 SO4 + 2 H + Fe log K = 472.82

Table 9: Species distribution in equilibrium with Ni(OH)2(beta). Database: MOLDATA. Code: The Geochemist's Workbench- 8.08. Aqueous species Ni [mol/l] Percentage [%] 2- -5 Ni(CO3)2 2.563 × 10 98.9 -7 NiCO3(aq) 1.608 × 10 0.6 + -7 Ni(HCO3) 1.348 × 10 0.5 Ni2+ 1.276 × 10-7 0.5

Experimental data Up to now, no solubility experiments have been performed at SCK•CEN. Within the frame of the European Project CATCLAY, migration as well as sorption experiments are currently in progress and the latter set of data will be complemented by solubility experiments.

Role of (organic) colloids/polynuclear species Nickel is prone to form oligomer hydroxo-complexes and the following two main species have 3+ 4+ been recognized, i.e. Ni2OH and Ni4(OH)4 . Due to the formation of these oligonuclear species, the onset of hydrolysis reaction of Ni(II) is strongly concentration dependent (Gamsjäger et al., 2005, p. 91). Association of Ni with OM under BC conditions has been acknowledged by Bruggeman et al. (in progress).

Solubility data provided to PA Source Range (SR) Except vaesite, bunsenite and olivine, all other minerals comprised in Table 8 were retained in the source range. NiCO3(s) and β-Ni(OH)2 are reported to be the most stable phases under ambient conditions and therefore have been retained in the group of minerals delimiting the source range (Hummel et al. 2002). Violarite (FeNi2S4) is a supergene sulfide mineral, which is formed by alteration and oxidation of primary sulfide assemblages (i.e. pentlandite-pyrrohtite-pyrite) in nickel ore mineralisations (Wikipedia). Thus, this phase has been included in the SR. As millerite (β-NiS) may also form under low temperature conditions, and represents an alteration product of other nickel minerals, it was also retained in the SR. Trevorite (NiFe2O4) represents a solid solution type of mineral belonging to the spinel group of minerals, which according to Bruno et al. (2001 b) are generally not formed at low temperatures. Hummel et al. (2002) even did not include this solid in their database. Despite the fact, that trevorite formation does not seem very probable (but cannot completely be excluded), it has been kept in the group of SR minerals.

Excluded phases: As olivine [(Mg,Fe,Ni)SiO3] and sulphide minerals such as vaesite (NiS2) and α-NiS, Ni3S2(s) and Ni3S4(s) are known or described (Gamsjäger et al., 2005) to form only under elevated temperature conditions, these minerals were not considered as relevant solubility controlling phases for Ni in the far-field. Bunsenite (NiO,cr) is reported to be formed either upon dehydration of Ni(OH)2(cr) at temperatures as high as 440-558 K (Hummel et al., 2002), or by hydrothermal decomposition of gaspéite (NiCO3,cr). Due to this, and its very high solubility, it is thought that NiO(cr) can be neglected as solubility limiting phase.

Best Estimate (BE)

Based on the reasoning above, Ni(OH)2(beta) is considered to represent the most probable solubility limiting solid for Ni.

Expert Range (ER)

The expert range (ER) is represented by the solubility of the BE, i.e. Ni(OH)2(beta) including its thermodynamic uncertainties.

Data retained in the DCF’s: -4 BE: Thermodynamic solubility of Ni(OH)2(beta): 7.2 × 10 M. ER: BE including thermodynamic uncertainty.

SR: Lower limit (LL): Thermodynamic solubility of violarite (FeNi2S4): 9.5 × 10-9 M. Upper limit (UL): taken equal to ER(UL).

Eh/pH/pCO2 sensitivity of the solid phases As suggested by the Pourbaix diagrams of Figure 11, NiCO3(cr) between pH 6-9 seems quite Eh independent. Only under very oxidizing conditions, Ni3O4 is stabilized. Changes in pCO2 are supposed to have a larger effect than Eh. Ni-hydroxide (β-Ni[OH]2) is expected to be stabilized under more oxidizing and alkaline conditions, while under more reducing and acid conditions, it would become less stable/more soluble.

Remarks The thermodynamic nickel data selected for MOLDATA were mainly taken from the NEA TDB (Gamsjäger et al., 2005; Vol. 6). The differences in the speciation diagrams calculated with the NEA and MOLDATA TDB result mainly from the fact, that two other aqueous carbonate + 2- complexes, i.e. Ni(HCO3) and Ni(CO3)2 were included into MOLDATA, originating from the ANDRA TDB. Apart from the two above mentioned nickel carbonate complexes, also an aqueous nickel thiosulphate (NiS2O3) species was copied from the ANDRA TDB and added to MOLDATA. NAGRA/PSI and NEA did not incorporate this latter species in their databases and only nickel + sulphides, i.e. (Ni[HS]2; NAPSI), (NiHS ; NEA, NAPSI) and sulphate species, i.e. (NiSO4,aq; 2- NEA, NAPSI) and Ni(SO4)2 ; NAPSI) are comprised therein.

The data of the solid polydymite (Ni3S4,s) were also copied from the ANDRA TDB. The latter solid was not incorporated in the NEA compilation due to the lack of experimental investigations concerning the thermodynamic properties. Nickel sulphides were discarded in the NAGRA/PSI TDB due to an evaluation performed by Thoenen (2002), revealing that no reliable solid Ni- sulphide data were available. It should be menioned that the latter evaluation was performed before the NEA review of Ni data was published in 2005. In contrast to all other TDB’s, the LLNL TDB does not contain any aqueous Ni carbonate species, due to which in the first diagram (Figure 11 a) the free Ni2+ ion represents the dominant cation in solution.

5.13 Niobium (Nb) General Niobium belongs to group 5 of the periodic table and is a transition metal with the atomic number 41. Niobium has physical and chemical properties similar to those of the element tantalum, due to which they are difficult to distinguish (Wikipedia). Large deposits of niobium have been found associated with carbonatites (carbon-silicate rocks), as a constituent of pyrochlore [(Na,Ca)2Nb2O6(OH,F)]. Extensive ore reserves are found in Canada, Brazil, Nigeria, Zaire, and in Russia. The largest deposit is hosted within a carbonatite intrusion at Araxá, Minas Gerais Brazil (Handbook of Chemistry and Physics, 1992-1993) . Additionally, niobium is often incorporated in niobate-tantale salts. Niobium has one naturally occurring stable isotope, i.e. 93Nb. At least 32 radioisotopes have been synthesized, ranging in atomic mass between 81 and 113. The most stable of the former is 92Nb with a half-life of 34.7 million years, followed by 94Nb (half-life: 20,300 years) and 91Nb with a half-life of 680 years. The isotope of major concern that is comprised in the SF/HLW waste inventory is 94Nb. The latter is produced through neutron activation of 93Nb and has a half-life of 20,000 years. 94Nb is found in metallic reactor components, but also as reactor coolant impurity/radionuclide generated through corrosion of stainless steel and Inconel surfaces in nuclear power plants. Niobium forms oxides with oxidation states +V (Nb2O5), +IV (NbO2), and +III (Nb2O3), and also with the rarer oxidation state +II (NbO). a) b) 1 1

++ Nb(OH)+3 .5 Nb(OH)4 .5 Nb(OH)5 HNbO3(aq) - Nb(OH)6 Eh (volts) Eh Eh (volts) Eh 0 0 - -- NbO3 Nb(OH)7

µ µ –.5 –.5

25°C 25°C 0 2 4 6 8 10 12 14 0 2 4 6 8 101214 pH pH

c) 1 d) 1

++ Nb(OH)3 .5 .5 Nb2O5(s) Nb2O5(s) - Nb(OH)6 Eh (volts) Eh Eh (volts) Eh 0 -- 0 -- Nb(OH)7 Nb(OH)7 µ µ

–.5 –.5

25°C 25°C 0 2 4 6 8 10 12 14 0 2 4 6 8 10 12 14 pH pH Figure 12: Eh-pH diagram of niobium (Nb-C-S-O-H) for the BC reference porewater system. Assumed activity of dissolved [Nb] = 10-8. Diagram a) LLNL TDB (thermo.com.R7beta) b) ANDRA

and MOLDATA TDB, c) same diagram as b, but solid phases included, d) same diagram as c, but [Nb] = 10-5. Code: The Geochemist's Workbench - 8.08.

Speciation and solubility calculations - Speciation calculations using MOLDATA indicate that the niobiate ion Nb(OH)6 is the prevalent species under undisturbed BC conditions (Figure 12 b). In Table 11, the species distribution in - 2- equilibrium with Nb2O5(s) is summarized. Besides Nb(OH)6 , Nb(OH)7 represents the second most important aqueous species. At [Nb] = 10-8, no phase is predicted to be stable under BC conditions. By increasing the niobium activity by 3 orders of magnitude in the calculations, Nb2O5(s) appears in the Pourbaix diagram and as can be seen in Figure 12 d is stable between pH = 0-9 over the entire Eh range. NbO2(s) is very soluble under BC conditions.

Table 10: Solubility of Nb in the BC reference porewater system at 25 °C, pH 8.355 and Eh -281 mV. Database: MOLDATA. Code: The Geochemist's Workbench- 8.08. Solubility controlling phases Solubility, [Nb], mol/l 2 -6 Nb2O5 (s) 2.4 × 10 3 NbO2(cr) very soluble Source data: 2ANDRA TDB, 3NAGRA/PSI

The reaction constants of the Nb minerals comprised in Table 10 and MOLDATA are the following: - + Nb2O5(s) + 7 H2O ↔ 2 Nb(OH)6 + 2 H log K = -28.38 - + NbO2(cr) + 3.5 H2O + 0.25 O2(aq) ↔ Nb(OH)6 + H log K = 16.68

Table 11: Species distribution of Nb in equilibrium with Nb2O5(cr). Database: MOLDATA. Code: The Geochemist's Workbench- 8.08. Aqueous species Nb [mol/l] Percentage [%] - -6 Nb(OH)6 1.65 × 10 69 2- -7 Nb(OH)7 7.52 × 10 31 Total 2.36 × 10-6 100

Experimental data No in-house or BC specific data available.

Role of (organic) colloids/polynuclear species Nb(V) polymeric/polynuclear species are reported to form only at total niobium concentrations exceeding 10-3 M (ThermoChimie documentation).

Solubility data provided to PA Source Range (SR)/Best Estimate (BE) Based on the restricted thermodynamic data for Nb, the solubility assessment for Nb is also limited. The solubility limiting phase for Nb considered by the scientific society dealing with waste issues corresponds to Nb2O5(s). Within the frame of a large experimental program of ANDRA, Nguyen-Trung et al. (1995) and Peiffert et al. (1997) performed solubility studies of Nb(V) oxides (amorphous and crystalline) under different geochemical conditions to get insight

into the effect of pH, ionic strength and temperature changes. Besides this, the influence of dissolved chloride, sulphate and carbonate ions on their solubility was evaluated. The obtained results enabled ANDRA to compile an important set of stability constants, which were also incorporated in MOLDATA. Besides ANDRA, also JNC, NIREX and NAGRA/PSI consider the Nb-pentoxide as the solubility limiting phase in their systems. As can be seen in Table 10, the solubility of this phase under BC conditions is quite high. However, Nb2O5(s) is considered to be the most relevant phase to be retained to control the aqueous Nb concentrations. Due to its very soluble character under BC conditons, NbO2 was neither retained in the source nor in the expert range.

Expert range (ER)

An uncertainty evaluation for Nb2O5(s) was not possible. Therefore, the lower and upper limits of the expert range were taken from the previous DCF’s (Marivoet et al., 1999).

Data retained in the DCF’s: -6 BE: Thermodynamic solubility of Nb2O5: 2.4 × 10 M. ER: Lower and Upper limits taken from previous DCF’s. SR: Taken equal to ER.

Eh/pH/pCO2 sensitivity of the solid phases Nguyen et al. (1995) observed that niobium pentoxide is only stable up to pH ~ 9 and may then transform into sodium niobiate Na8Nb6O19 × 13 H2O.

Remarks The LLNL (thermo.comv8.r6+.dat) and NEA TDB do not comprise data for niobium. Therefore, an alternative LLNL TDB, the so-called thermo.com.R7beta.dat TDB was used to calculate the speciation diagram. ANDRA dedicated a quite exhaustive experimental program to the study of niobium. Therefore, and due to the fact that no thermodynamic data are provided by NEA for this element, most of the - data selected in MOLDATA have been copied from the ANDRA TDB. The niobiate ion Nb(OH)6 represents the basis species of Nb(V).

5.14 Palladium (Pd) General Palladium belongs to the group of platinum (Pt) metals (besides rhodium, ruthenium, iridium and osmium) and its atomic number is 46. The most stable and waste relevant isotope is 107Pd with a half-life of 6.5 million years. It is produced during nuclear fission, but only in low percentages. Due to its low decay energy (30 KeV), Pd does not represent a very hazardous isotope. Pd forms rare minerals, such as cooperite [(Pt,Pd,Ni)S] and polarite [(Pd)Bi,Pb),cr]. Additionally, Pd sulphide and arsenide minerals are known (Hummel et al., 2002). Examples are Vysotskite (PdS,cr), PdS2(s) and Pd4S(s), the latter being high-T solids (Hummel et al., 2002).

a) b) 1 1

.5 PdCl2(aq) .5 PdCl2

Pd(OH)2

Eh (volts) Eh PdO(aq) Eh (volts) Eh 0 0 Pd(OH)--- Pd(OH)43 µ µ

–.5 –.5

25°C 25°C 0 2 4 6 8 10 12 14 0 2 4 6 8 10 12 14 pH pH

c) d) 1 1

.5 PdCl2 .5 PdCl2

Pd(OH)2 Pd(OH)2 Eh (volts) Eh 0 Eh (volts) 0 - Pd(OH) Pd(OH)--- 3 Pd(OH)34 µ µ

–.5 –.5

25°C 25°C 0 2 4 6 8 101214 02468101214 pH pH

1 f) 1 PdCl2 PdCl2

.5 PdO(s) .5 PdO(s)

-- -- Pd(OH)4 Pd(OH)4 Eh (volts) Eh Pd(cr) (volts) Eh 0 0 Pd(cr) PdS(s) µ µ

–.5 –.5

25°C 25°C 02468101214 02468101214 pH pH

g) h) 1 1

.5 PdCl2 .5 PdCl2

Pd(OH)2(s)

PdO(s) (volts) Eh

Eh (volts) Eh 0 0 Pd(OH)--- -- Pd(OH)43 Pd(OH)4 µ µ –.5 –.5 25°C 25°C 02468101214 02468101214 pH pH Figure 13: Eh-pH diagram of palladium (Pd-C-S-O-H) for the BC reference porewater system. Assumed activity of dissolved [Pd] = 10-8. Diagram a) LLNL TDB, b) ANDRA TDB c) NAPSI TDB, d) MOLDATA TDB, e) same as diagram d, but solid phases included, f) same diagram as e, but PdS suppressed in calculation, g) same diagram as f, but Pd(cr) suppressed in calculation, h) [Pd] = 10-5 PdS(s), Pd(cr) and PdO(s) suppressed in calculation. Code: The Geochemist's Workbench - 8.08/8.10.

Speciation and solubility calculations Palladium occurs in 3 oxidation states, i.e. 0, +2 and +4. The hydrolysis of the Pd2+ ion is reported to start at very low pH values, i.e. pH ~0.7 (Hummel et al., 2002). Besides this, Pd has a tendency to form polynuclear species and colloids, due to which the hydrolysis behaviour of Pd is controversely discussed. -8 Speciation calculation performed with MOLDATA at [Pd] = 10 show that, Pd(OH)2(aq) is the dominant species under BC conditions (Figure 13 d). The same speciation is predicted by calculations using the ANDRA or NAGRA/PSI TDB, respectively (Figure 13 b, c). It should be mentioned that the aqueous species PdO(aq) was suppressed in the calculations, as the reaction constant was found to be wrong due to erroneous data in the LLNL TDB from which this species was copied to MOLDATA (Wang, pers. communication). Solubility calculations were performed for Pd(IV) oxide and hydroxide, as well as for elemental Pd and Pd sulphide. The results are summarized in Table 12. Elemental Pd and PdS(s) are characterized by an extremely low solubility under BC conditions. According to ANDRA (Dossier

5), these phases represent quite unprobable solubility limiting phases and it seems more reasonable to retain the oxide and hydroxide phases in SA evaluations.

Table 12: Solubility of Pd in the BC reference porewater system at 25 °C, pH 8.355 and Eh -281 mV. Database: MOLDATA. Code: The Geochemist's Workbench- 8.08. Solubility controlling phases Solubility, [Pd], mol/l -6 Pd(OH)2 (s) 4.0 × 10 PdO (s) 1.5  10-10 Pd (cr) 2.2 × 10-30 (insoluble) PdS (s) 7.9 × 10-35 (insoluble) Source data: all solid phase data were taken from the ANDRA TDB

The reaction constants for the Pd solids comprised in Table 12 and MOLDATA are the following: + 2+ Pd(OH)2(s) + 2 H ↔ Pd + 2 H2O log K = -1.61 + 2+ PdO(s) + 2 H ↔ Pd + H2O log K = -6.02 + 2+ Pd(cr) + 0.5 O2(aq) +2 H ↔ Pd + H2O log K = 9.96 2+ 2- PdS(s) + 2 O2(aq) ↔ Pd + SO4 log K = 91.42

Based on the solubility and speciation calculations, the following precipitation/formation sequence could be anticipated:

At [Pd] = 10-8: 1) PdO(s) kinetically favoured phase → 2) Pd(cr) → 3) PdS(s) most stable phase and at [Pd] = 10-5:

1) Pd(OH)2(s) kinetically favoured phase → 2) PdO(s) → 3) Pd(cr) → 4) PdS(s) most stable phase

Experimental data No in-house or BC specific data available.

Role of (organic) colloids/polynuclear species Pd(II) is reported to tend to form hydroxo polynuclear complexes and colloids (ThermoChimie documentation). Pd-NOM association has been put forward in the newly developed phenomenological model (Bruggeman et al., in prep.)

Solubility data provided to PA Source Range (SR) All phases reported in Table 12 have been retained to delimit the solubility source range (SR).

Best Estimate (BE)

Pd(OH)2(s) has been judged to represent the most likely solubility/concentration limiting solid for Pd.

Expert Range Thermodynamic data for Pd were copied from ThermoChimie v.5. No uncertainty values are provided for the standard molar Gibbs energies of formation of Pd(OH)2(aq) and the solid hydroxide. The same is the case for the reaction constant. Therefore, it was not possible to calculate an uncertainty and an arbitrary value of 1 order of magnitude was attributed to the solubility of Pd(OH)2(s).

Data retained in the DCF’s: -6 BE: Thermodynamic solubility of Pd(OH)2(s): 4.0 × 10 M. ER: BE ± 1 order of magnitude. SR: Lower limit (LL): Thermodynamic solubility of PdS(cr): 7.9 × 10-35 M. Upper limit (UL): BE including uncertainty of one order of magnitude.

Eh/pH/pCO2 sensitivity of the solid phases As revealed by the diagrams of Figure 13 e-h, Pd solids are remarkably stable over the entire Eh- pH range. PdS would become unstable at higher Eh and pH conditions (compared to the reference conditions) and replaced by Pd(cr). The latter phase would then - at higher redox - be oxidized to PdO(s). Pd(OH)2(s) stability seems pH and Eh independent in the pH range 2-12.

Remarks No thermodynamic data for palladium are incorporated in the NEA TDB. In the LLNL TDB only the first hydrolysis species of Pd is incorporated (i.e. PdOH+). In the NAGRA/PSI TDB only the second and third hydrolysis species are considered, while ANDRA 2-x gives the entire set of hydrolysis constants (i.e. Pd(OH)x with x = 1, 2, 3 and 4). Thermochemical data for PdO(s) were not included into the NAGRA/PSI TDB, as precipitation of this phase from aqueous solution at 25°C was not observed. According to the review of Hummel et al. (2002), the formed phase correponded rather to PdO × H2O(precip) or Pd(OH)2(s). Pd sulphides are also not included in the database, as PdS2(s) and Pd4S(s) represent only high-T phases and also none of the available data concerning the low-T stable PdS(cr) from Pearson et al. (1992) have been evaluated to be reliable. The possible incorporation of Cl anions into the Pd(OH)2(s) structure is reported by ANDRA (ThermoChimie documentation). Thus, mixed hydroxo-chloride solids of Pd may be included as relevant solubility limiting phases in waters of higher salinity.

5.15 Plutonium (Pu) General Plutonium (atomic number = 94) is a transuranic radioactive element and represents the actinide of major environmental concern. By far of greatest importance is the isotope 239Pu, with a half-life of 24,100 years, produced in extensive quantities in nuclear reactors from natural uranium:

238U  239U  239Np  239Pu

Neutrons of the fission of 235U are captured by 238U to form 239U. Through two beta decays, and 239Np as intermediate, 239Pu is synthesized. Twenty isotopes of plutonium have been characterized of which 244Pu is the longest-lived one with 80.8 million years. Alpha particle emission, which is the release of high-energy helium nuclei, is the most common form of radiation given off by 238 239 plutonium (Wikipedia). The most stable plutonium isotopes are: Pu (t1/2 = 87.7 years), Pu 4 240 3 241 242 5 (2.41 × 10 years), Pu (t1/2 = 6.56 × 10 years), Pu (t1/2 = 14.3 years), Pu (t1/2 = 3.74 × 10 years) and 244Pu (8.00 × 107 years). In aqueous solutions, plutonium can exist in five oxidation states, i.e. the tri-, tetra-, penta- hexa- and heptavalent state. It has been shown that Pu(III), Pu(IV), Pu(V) and Pu(VI) can coexist in unequal quantities in the same solution. This can be explained by the similar redox potentials between the Pu(IV), Pu(V) and Pu(VI) states (Choppin, 2003). Disproportio-nation reactions may also influence the oxidation states of Pu. The tendency of plutonium to hydrolyze follows the effective charge of the ion in the order of (Kim, 1986):

+4 2+ +3 + Pu > PuO2 > Pu > PuO2

Pu(III) is however unstable in aqueous solutions, as it is easily oxidized and becomes less stable as pH increases (Lemire et al., 2001). Therefore, under neutral or alkaline conditions (and in absence of carbonates), the trivalent state is much less likely to be present than under strong acid and anoxic conditions. The speciation of pentavalent Pu is relatively simple since it is only a very weak complexant (Choppin, 2003) and begins to hydrolyse between pH 7-8. In oxic waters, penta- + 2+ as well as hexavalent plutonium form dioxo-cations, such as Pu(V)O2 and Pu(VI)O2 . If the concentration is relatively high, Pu(V) can also form dimeric hydroxides (e.g. [PuO2OH], + [(PuO2)2(OH) ].

a) ++ b) PuO++++ ++ 1 Pu 2 1 PuO2

Pu(OH)++ + 2 PuO+ 2CO3(aq) PuO2 PuO2 -- -- PuO2(CO3)2 PuO2(CO3)2 .5 ---- .5 + PuO2(CO- 3)3 Pu(OH)3 PuO2CO3 PuO2OH(aq) PuO2(OH)2(aq) +++ PuO OH(aq) +++ Pu 2 Pu - Pu(CO3)2(OH)2 2 Eh (volts) Eh Eh (volts) Eh 0 0 Pu(OH)4(aq) Pu(CO )+ 3 Pu(OH)4(aq) - µ Pu(COµ3)2 PuOH++ --- –.5 –.5 Pu(CO3)3

25°C 25°C Pu(OH)3 0 2 4 6 8 10 12 14 0 2 4 6 8 10 12 14 pH pH

c) ++ d) ++ 1 PuO2 1 PuO2

PuO CO Pu(OH)++ + 2 3 2 PuO+ 2CO3(aq) PuO2 -- PuO2 -- PuO2(CO3)2 PuO2(CO3)2 ---- .5 PuO (CO- ) ---- 2 3 3 .5 + PuO2(CO- 3)3 PuO2CO3 Pu(OH) PuO2CO3 PuO (OH) 3 PuO (OH) (aq) PuO2 OH2 2 2 +++ 2 PuO2OH(aq) Pu Pu+++ Eh (volts) Eh Eh (volts) Eh 0 0 Pu(OH)4 Pu(OH)4(aq)

µ++ µ++ PuOH PuOH –.5 –.5

25°C 25°C 0 2 4 6 8 101214 0 2 4 6 8 10 12 14 pH pH

f) e) PuO++ ++ 1 2 ++ 1 PuO2 Pu(OH)2 Pu(OH)++ + 2 PuOPuO2 2CO3(aq) + PuO (CO )-- Pu(OH)3 2 3 2 ---- .5 .5 PuO2(CO3)3 PuO2(OH)2 Pu+++ Pu+++ PuO2(cr) Eh (volts) Eh Eh (volts) Eh PuO (am,hyd) 0 0 2

+ Pu(CO3) Pu(CO )+ µ µ - 3 Pu(CO3)2 --- –.5 –.5 Pu(CO3)3 25°C 25°C 0 2 4 6 8 101214 02468101214 pH pH g) ++ 1 PuO2

Pu(OH)++ 2 PuO+ 2CO3(s) PuO2

- ---- .5 + PuOPuO2(CO(OH)3)3 Pu(OH)3 PuO2CO32 2

Pu+++ - Pu(CO3)2(OH)2 2

Eh (volts) Eh - no graph - 0 + Pu(OH) (aq) Pu(CO3) 4 Pu(CO )(OH)(s) 3 µ

–.5 Pu(OH)3(cr) 25°C 0 2 4 6 8 10 12 14 pH Figure 14: Eh-pH diagram of plutonium (Pu-C-S-O-H) for the BC reference porewater system. Assumed dissolved activity of [Pu] = 10-8. Diagram a) LLNL TDB, b) ANDRA/ MOLDATA TDB, c) NAGRA/PSI TDB, d) NEA TDB, e) MOLDATA TDB solid phases included, f) MOLDATA TDB -4 PuO2(cr) suppressed, g) MOLDATA TDB with [Pu] = 10 , but PuO2(cr) and PuO2(am,hyd) suppressed in calculation. Code: The Geochemist's Workbench - 8.08.

Speciation & solubility calculations According to the speciation calculations using MOLDATA, mono-nuclear Pu(III) carbonate 3- - complexes, i.e. Pu(CO3)3 and Pu(CO3)2 are predicted to represent the dominant aqueous species under BC conditions (Figure 14 b). Results of the solubility calculations are summarized in Table 13. Using a Pu activity of 10-8, the crystalline Pu-oxide represents the most stable solid phase. However, taking into account the Ostwald principle, rather the amorphous phase, i.e. PuO2(am,hyd) would precipitate first and with time/ageing convert to the crystalline form. At higher Pu activities (e.g. 10-4), the mixed Pu- hydroxy-carbonate solid becomes stable and would be the solubility controlling mineral. Pu(OH)3(cr) occupies the stability field under more alkaline conditions.

At [Pu] = 10-4 the following precipitation sequence can be anticipated:

1) PuCO3OH(s) → 2) PuO2(am,hyd) → 3) PuO2(cr)

Table 13: Solubility of Pu in the BC reference porewater system at 25°C, pH 8.355 and Eh -281 mV. Database: MOLDATA. Code: The Geochemist's Workbench - 8.08. Solubility controlling phases Solubility, [Pu], mol/l 1 -4 Pu(OH)3(cr) 1.4 × 10 2 -5 PuCO3OH(s) 1.6 × 10 1 -8 PuO2 (am,hyd) 1.3 × 10 1 -14 PuO2 (cr) 2.6  10 Source data: 1NEA TDB, 2ANDRA TDB

The reaction constants of the Pu solids comprised in Table 13 and MOLDATA are the following: + 4+ Pu(OH)3(cr) + 4 H + 0.25 O2(aq) ↔ Pu + 3.5 H2O log K = 19.6 + 4+ - PuCO3OH(s) + 0.25 O2(aq) +3 H ↔ Pu + HCO3 + 1.5 H2O log K = 8.39 + 4+ PuO2(am,hyd) + 4 H ↔ Pu + 2 H2O log K = -2.33 + 4+ PuO2(cr) + 4 H ↔ Pu + 2 H2O log K = -8.03

Table 14, the calculated species distribution in equilibrium with PuO2(am,hyd) is illustrated. It can 3- - be seen, that 70 % of the species are Pu(CO3)3 and 27% are of the form Pu(CO3)2 . According to calculations using the LLNL, NAGRA/PSI and NEA TDB, Pu(OH)4(aq) is the prevalent aqueous Pu species (Figure 14 a, c, d).

Table 14: Species distribution of Pu in equilibrium with PuO2(am,hyd). Database: MOLDATA. Code: The Geochemist's Workbench - 8.08. Aqueous species Pu [mol/L] Fraction [%] III 3- -9 Pu (CO3)3 9.23 × 10 70 III - -9 Pu (CO3)2 3.53 × 10 27 III + -10 Pu CO3 3.37 × 10 3 Total 1.33 × 10-8 100

Role of (organic) colloids/polynuclear species It should be mentioned that intrinsic/real colloid formation plays an important role in the behaviour of Pu(IV). The latter is known to hydrolyse even in dilute acidic solutions and at concentrations exceeding 10-6 M will undergo polymerization to form very stable intrinsic colloids. This colloid generation process has been the reason for difficulties in determining accurate thermodynamic solubility constants and in predicting the soluble Pu(IV) concentrations in natural waters (Runde, 2000). Colloids can vary in size and the distribution is not uniform and depends on ageing. The fraction of larger particles is increasing with time (Silva and Nitsche, 1995). Inorganic colloid formation (i.e. Pu-NOM colloid association) has been put forward in the newly developed phenomenological model (Bruggeman et al., in prep.).

Experimental data No in-house or BC specific data available.

Solubility data provided to PA Source Range (SR) All phases mentioned in Table 13 have been retained in the source range (SR).

Best Estimate (BE)

Taking the Ostwald Step Rule as guide, PuO2(am,hyd) is considered to represent the most likely phase to control Pu-concentrations in the far-field.

Expert range (ER)

An unrealistic high uncertainty was calculated for PuO2(am,hyd) due to the high uncertainty given 3- -1 for the molar Gibbs energy of formation of Pu(CO3)3 (± 60.576 kJ mol , see ANNEX II) for further calculation details), representing the predominant aqueous species under BC conditions. Therefore, an arbitrarily chosen uncertainty value of 2 orders of magnitude was attributed to the solubility value of the BE.

Data retained in the DCF’s: -8 BE: Thermodynamic solubility of PuO2(am,hyd): 1.3 × 10 M. ER: BE ± 2 log units. -14 SR: Lower limit: Thermodynamic solubility of PuO2(cr): 2.6 × 10 M. -4 Upper limit (UL): Thermodynamic solubility of Pu(OH)3(cr): 1.4 × 10 M.

Eh/pH/pCO2 sensitivity of the solid phases

As can be seen in Figure 14 e, PuO2(cr) is characterized by a very large stability field. At pH < 6 (and BC Eh), the stability of PuO2(cr) is however decreasing or in other words, solubility would increase. The amorphous, hydrated oxide is less stable and more sensitive to changes in Eh and pH, which is even more the case for PuCO3OH(s). As suggested by the Pourbaix diagram (Figure 14 f), increasing Eh and pH (compared to reference conditions) tends to stabilize the PuO2(am,hyd) solid. With respect to PuCO3OH(s), pCO2 changes may additionally affect the stability/solubility of this phase. Under more alkaline conditions, the Pu-hydroxide becomes the more stable phase.

It should however be stressed that the above mentioned statements represent only evaluations, and more specific calculations should be performed to enable better "predictions" about the influence of changing conditions/parameters on the solubility of the different phases.

Remarks Due to the fact that trivalent Pu (and Np) species are easily oxidized and only stable in the presence of strong reductants, the determination of thermodynamic data, e.g. hydrolysis constants is difficult and experimental data are scarce. Results of the few conducted experiments were evaluated to be generally unreliable. Only one hydrolysis constant was proposed in the NEA review of 2001 (Vol. 4) and no data were added during the last review performed in 2003 (Guillaumont et al., 2003)

3+ + + Pu + H2O ↔ Pu(OH)2 + H log β = -6.9 ± 0.3

Furthermore, the NEA TDB does not contain Pu(III) carbonate complexes. The reason for the reluctance to incorporate these species is also related to the easy oxidation of Pu(III) to Pu(IV) in basic aqueous solutions, lacking experimental studies and evidence concerning the identity of such complexes and their stability.

In contrast to NEA, ANDRA recommends the incorporation of trivalent Pu carbonate complexes 2n-3 with stoichiometries of An(CO3)n with n = 1, 2 and 3. The selected data were taken from (Yui et al. 1999, JNC TDB) and the values estimated by analogy to Am carbonates.

According to Guillaumont et al. (2005), also the problem to provide qualitative thermodynamic data for Pu(IV) hydroxide complexes is persisting and even "exacerbating", which among other reasons is due to the phenomenon/tendency of tetravalent actinides to form colloids. It has been shown by Knopp et al. (1999) that when the Pu(IV) concentration exceeds the solubility of Pu(OH)4(am,hyd), colloids represent the dominant species in solution.

The question remains whether plutonium under undisturbed Boom Clay conditions prevails as Pu(III) or Pu(IV).

5.15.1 The redox behaviour of tri- and tetravalent (Np and) Pu The only trivalent hydrolysis species comprised in the NEA TDB are Np(OH)2+ and Pu(OH)2+. The ANDRA TDB additionally comprises data for the second and third hydrolysis products + Np/Pu(OH)2 and Np/Pu(OH)3. In MOLDATA all these trivalent Pu-species have been incorporated. Data for tetravalent hydrolysis products incorporated in MOLDATA have been taken from the ANDRA TDB. For details concerning the source of these data, it is referred to the documentation of ThermoChimie v.5. In Figure 15 (a and b) below, it can be seen that the stability field of Np3+ is restricted to a very small Eh-pH domain, while the Pu3+ cation covers a much bigger stability field, which reflects well the much higher stability of Pu(III) compared to Np(III). In these diagrams carbonate speciation was neglected to better illustrate this fact. If taking into account carbonate speciation, Np(OH)4(aq) stays the predominant species for Np under BC conditions, while for Pu the - 3- carbonate species Pu(CO3)2 and Pu(CO3)3 are predicted to dominate speciation in BC porewater and not anymore Pu(OH)4(aq).

++ a) NpO2 b) ++ 1 1 PuO2

Pu(OH)++ 2 + PuO2 + NpO2 .5 .5 + PuO2(OH)2(aq) +++ Pu(OH) NpOH - 3 PuO2OH(aq) NpO2(OH)3 ++ NpO (OH)-- +++ Np(OH)2 2 4 Pu NpO2OH(aq) Eh (volts) Eh Eh (volts) Eh +++ - 0 Np + NpO2(OH)2 0 Np(OH)3 Pu(OH)4(aq)

Np(OH)4(aq) µ PuOHµ++ –.5 –.5 + Pu(OH)2 25°C 25°C Pu(OH)3 0 2 4 6 8 101214 0 2 4 6 8 10 12 14 pH pH

++ NpO2 c) d) ++ 1 1 PuO2 ++ Pu(OH) PuO CO (aq) -- 2 PuO+ 2 3 NpO2(CO3)2 2 -- + PuO2(CO3)2 NpO2 ---- NpO2(CO3)3 .5 ---- .5 + PuO2(CO- 3)3 +++ Pu(OH)3 PuO2CO3 NpOH - PuO (OH) (aq) - NpO (OH) 2 2 NpO CO 2 3 +++ PuO OH(aq) ++ 2 3NpO (OH)-- Pu 2 Np(OH)2 2 4 - Pu(CO3)2(OH)2 2 Eh (volts) Eh ---- (volts) Eh +++ NpO (CO ) OH 0 Np + 2 3 2 0 Np(OH)3 Pu(CO )+ 3 Pu(OH)4(aq) - µNp(OH)4(aq) Pu(COµ3)2 --- –.5 –.5 Pu(CO3)3

25°C 25°C Pu(OH)3 0 2 4 6 8 10 12 14 0 2 4 6 8 10 12 14 pH pH Figure 15: Comparison of Np(III)/Np(IV) and Pu(III)/Pu(IV) speciation. Assumed activites of dissolved [Np] and [Pu] =10-8 (MOLDATA TDB). In diagrams a and b, carbonate complexation was neglected, while in diagrams c and d, the influence of carbonate complexation was taken into account.

The higher stability of Pu3+ compared to Np3+ is well reflected by the following equations and equilibrium reactions:

Np4+/Np3+ equilibrium:

4+ 3+ + Np + 0.5 H2O ↔ Np + 0.25 O2(aq) + H Log25°C: -17.8010

Equilibrium equation: 4+ 3+ + log K = - log a[Np ] + log a[Np ] + 0.25 × log a[O2(aq)] + log a[H ]

- 0.5 × log a[H2O]

Pu4+/Pu3+ equilibrium:

4+ 3+ + Pu + 0.5 H2O ↔ Pu + 0.25 O2(aq) + H Log25°C: -3.8024

Equilibrium equation: 4+ 3+ + log K = - log a[Pu ] + log a[Pu ] + 0.25 × log a[O2(aq)] + log a[H ]

- 0.5 × log a[H2O]

5.16 Protactinium (Pa) General Protactinium belongs to the group of actinides and its atomic number is 91. It is one of the rarest and most expensive naturally occurring elements. Pa can be present in uraninite (pitchblende) ore deposits (e.g. Jachymov, Czech Republic or Republic of Kongo) in amounts ranging between 0.3 and 3 ppm. The longest-lived and most abundant (nearly 100 %) naturally occurring isotope of protactinium is 231Pa with a half-life of 32,783 years, which is a decay product of 235U. Due to its long half-life, it is a major contributor to the long-term radiotoxicity of spent nuclear fuel (Wikipedia). 233Pa represents a decay product of 233Th and is formed via neutron irradiation of the latter. Pa occurs in the oxidations states +II, +III, +IV and +V. Tetravalent Pa is only stable in very acidic media and easily oxidized to Pa(V), which is known to be prone to hydrolysis (Berner, 1995; Gascoyne, 1992).

a) b) 1 1

++ PaO(OH)++ PaO(OH) .5 .5 + + PaO(OH)2 PaO(OH)2 Eh (volts) Eh Eh (volts) Eh 0 0 Pa2O5(s) Pa(OH)5(aq)

µ µ

–.5 –.5

25°C 25°C 02468101214 02468101214 pH pH Figure 16: Eh-pH diagram of protactinium (Pa-C-S-O-H) for the BC reference porewater system. Assumed activity of [Pa] = 10-8. Diagram a) ANDRA/MOLDATA TDB (2010_MOLDATA_nov_b O2.dat) b) ANDRA/MOLDATA TDB solid phases included. Code: The Geochemist's Workbench - 8.12.

Speciation & solubility calculations Thermodynamic data for Pa are still scarce, but were comprised in ThermoChimie v.5 from which they were copied to MOLDATA. More recent data have become available in the meantime (Trubert et al., 2002; Trubert et al., 2003) and were taken into account in the calculations presented here. Based on these data, the prevalent aqueous species under BC conditions is calculated to be Pa(OH)5(aq) (see Figure 16 a). Three solid phases are comprised in MOLDATA, i.e. Pa(cr), PaO2(s) and Pa2O5(s). The former two solid were calculated to be very soluble under BC conditions and therefore not included in the solubility assessment (see below). The solubility calculated for Pa2O5(s) corresponds to to 9.8 × 10-10 (Table 15).

Table 15: Solubility of Pa in the BC reference porewater system at 25°C, pH 8.355 and Eh -281 mV. Database: MOLDATA. Code: The Geochemist's Workbench - 8.08. Solubility controlling phases Solubility Pa [mol/l] -10 Pa2O5 (s) 9.8 × 10

PaO2 (s) very soluble Pa(cr) very soluble Source data: ANDRA TDB

The reaction constant for the Pa2O5(s) solid comprised in MOLDATA is the following: + 4+ Pa2O5(s) + 8 H ↔ 2 Pa + 4 H2O log K = -55.43 + 4+ PaO2(s) + 4 H ↔ Pa + 2 H2O log K = 166.78 + 4+ Pa(cr) + 4 H + O2(aq) ↔ Pa + 2 H2O log K = 184.73

Experimental data No in-house or BC specific data available.

Role of colloids/polynuclear species Strong tendency of Pa(V) towards hydrolysis, colloid and polymer formation is reported by Trubert et al. (2003). Pa-NOM association has been put forward in the newly developed phenomenological model by Bruggeman et al. (in prep.).

Solubility data provided to PA Source Range (SR)

As the solubility assessment is based only on one solid phase, i.e. Pa2O5(s), the source range was delimited on the one hand by data extracted from migration experiments, i.e. the measured constant concentration (LL: 1×10-11 M) and on the other hand by data comprised in the former DCF (UL: 1 × 10-5 M).

Best Estimate (BE)

Thermodynamic solubility of Pa2O5(s).

Expert Range (ER) An unrealistic high uncertainty has been attributed to the molar Gibbs energy of formation of -1 Pa2O5(s) (± 35.531 kJ mol , see ANNEX II for further calculation details). Therefore, an arbitrarily chosen uncertainty value of 1 order of magnitude was attributed to the solubility value of the BE.

Data retained in the DCF’s: -10 BE: Thermodynamic solubility of Pa2O5(s): 9.8 × 10 M ER: BE ± one order of magnitude. SR: Lower limit (LL): Taken equal to constant concentration from migration experiment: 1×10-11 M. Upper limit (UL): Former DCF reports solubilities up to 10-5 (based on Finnish data).

Eh/pH/pCO2 sensitivity of the solid phases

As can be seen in Figure 16 b, the solubility of Pa2O5(s) is independent of Eh. According to Yui et al. (1999), the solubility of Pa2O5 is also independent of pH in the range from 6-11, but increases at acidic pH values.

Remarks Pa shows "unique" chemical properties and behaviour, due to which analogies with other elements/actinides are controversely discussed. According to Trubert et al. (2003), Pa(IV) shows little resemblance with Nb(IV) and Ta(IV), but any with pentavalent actinides. Guillaumont (1966) considers that Zr(IV) and Hf(IV) can be taken as good analogues for the acid-base properties of Pa(IV), while Kraus and Nelson (1950) consider analogy with U(IV) and Th(IV) more appropriate. With respect to pentavalent Pa, consensus seems to exist that no good analogues exist.

5.17 Radium (Ra) General Radium belongs to the group of earth alkaline elements and its atomic number is 88. Radium is a naturally occurring radionuclide that is part of the uranium and thorium decay series and 4 isotopes, i.e. 228Ra, 226Ra, 224Ra and 223Ra are known to exist in nature, which are all radioactive. In addition to their own radiological properties, three of these radium isotopes present additional environmental and health concerns due to the fact that they decay to radon. The radium isotope 226 3 considered in performance assessment calculations is Ra (t1/2= 1.6 × 10 years), representing the 230 4 238 daughter product of Th (t1/2 = 7.52 × 10 years), which itself is part of the U (4n + 2) decay 228 223 chain/series. The other radium isotopes, such as Ra (t1/2 = 5.75 a), Ra (t1/2 = 11.68 d) and 224 Ra (t1/2 = 3.64 d) are irrelevant with respect to long-term waste management considerations as they represent short-lived isotopes. The chemical properties of radium are similar to those of 2+ + + barium. Radium dissolves as Ra and may form following complexes RaOH , RaCl , RaCO3(aq), + RaHCO3 , and RaSO4(aq).

a) b) 1 1

.5 .5

++ + Ra RaHCO3 Eh (volts) Eh Eh (volts) Eh 0 0

µ µ

–.5 –.5

25°C 25°C 0 2 4 6 8 10 12 14 0 2 4 6 8 101214 pH pH

c) d) 1 1

.5 .5 ++ Ra++ Ra Eh (volts) Eh Eh (volts) Eh 0 0 RaCO3(aq) RaCO3 µ µ

–.5 –.5

25°C 25°C 0 2 4 6 8 10 12 14 0 2 4 6 8 10 12 14 pH pH Figure 17: Eh-pH diagram of radium (Ra-C-S-O-H) for the BC reference porewater system. Assumed activity of dissolved [Ra] = 10-8. Diagram a) LLNL TDB, b) ANDRA TDB, c) NAGRA/PSI TDB, d) MOLDATA TDB. Code: The Geochemist's Workbench - 8.08/8.10

Speciation & solubility calculations Aqueous speciation under the BC reference conditions is dominated by the divalent radium cation (Figure 17 d, Table 17). At higher pH values (> 9.5), the free Ra cation is replaced by the -8 RaCO3(aq) complex. At [Ra] = 10 , radium is not solubility limited and only at higher Ra concentrations, radium concentrations might be controlled by either Ra carbonate or sulphate (Table 16). Considering the possibility of co-precipitation of Ra or its incorporation into a solid solution, the above mentioned concentration threasholds could be lower as these phases are known to reduce the solubility of the pure endmembers (for details see below).

Table 16: Solubility of Ra in the BC reference porewater system at 25 °C, pH 8.355 and Eh -281 mV. Database: MOLDATA. Code: The Geochemist's Workbench- 8.10 Solubility controlling phases Solubility, [Ra]/[Ba], mol/l -5 RaCO3(s) 6.7 × 10 -6 RaSO4(s) 7.0 × 10 Source data: ANDRA TDB

The reaction constants for the Ra-solids comprised in (Table 16) and MOLDATA are the following: + 2+ - RaCO3(s) + H ↔ Ra + HCO3 log K = 2.03 2+ 2- RaSO4(s) ↔ Ra + SO4 log K = -10.26

Table 17: Species distribution of Ra in equilibrium with RaSO4(s). Database: MOLDATA. Code: The Geochemist's Workbench-8.12. Aqueous species [mol/l] Percentage [%] Ra2+ 6.615 × 10-6 94.1 + -7 RaHCO3 2.183 × 10 3.1 -7 RaCO3(aq) 1.663 × 10 2.4 -8 RaSO4(aq) 3.115 × 10 0.4

Role of solid solutions comprising Ra (e.g. Ra-barite solid solution [(Ba,Ra)SO4]) Most (pure) radium salts are insoluble, especially the sulphate (RaSO4,s) and carbonate (RaCO3,s) salts. According to Langmuir and Riese (1985), Ra concentrations in natural waters, and waters associated with uranium mining and nuclear waste disposal are however rarely high enough to reach saturation with a pure radium solid. Therefore, maximum radium concentrations are rather limited by adsorption or solid solution formation. The latter process has been extensively studied in the last years by many authors and it is well known that co-precipitation of Ra isotopes with barite is a phenomenon that should be taken into account when discussing the solubility limit of this radionuclide. The ionic radius of radium (1.43 Å) is very similar to the one of barium (1.36 Å) due to which it can be accommodated in the barite lattice. Besides this, barite commonly contains minor amounts of strontium (Sr) and lead (Pb). Different studies have shown that the solubility limit for radium in many natural and anthropogenic environments is controlled by barite precipitation. Thus, barite represents an excellent scavenger for this radionuclide. Berner and Curti (2002) calculated the retention of radium in a nuclear waste repository in Opalinus Clay as host rock by using the Gibbs Energy Minimization (GEM) approach. Their results showed that the the total radium concentration in the porewater was 4.8 × 10-8 M, if controlled by the pure phase -12 RaSO4(s), while the solubility limit was lowered by few orders of magnitude (8.2 × 10 ) when considering co-precipitation of radium with barite.

Kulik et al. (2004) showed that the radium concentration in solution is dependent on the Ba/Ra activity ratio in the precipitated solid solution. The higher the ratio, the more the solubility is decreased compared to the pure phase. It should be mentioned that the sulphate concentration in undisturbed BC porewater is quite low, i.e. 0.02 mmol/L. Barium concentrations have been determined by De Craen et al. (2004) to range between 10-60 µg/L (7.3 × 10-8 - 4.4 × 10-7 mol/L). A significant decrease in Ra solubility compared to pure RaSO4(s) can be expected only at Ba/Ra activity ratios >100, which seems however not to be achievable under undisturbed BC conditions. Thus, for trace concentrations of Ra, the "host mineral" solid solution can be considered and must be close to saturation with respect to the groundwater in order to limit the Ra solubility. It should be bared in mind that sulphate concentrations may also be affected by microbial 2- activity/perturbation. The presence of sulphate reducing bacteria could diminish the SO4 concentration in solution, leading to the dissolution of the Ra solid phase (either solid solution or pure phase) and to a remobilization/enhanced mobility of radium. Also the formation of a binary solid solution between the two end-members BaCO3 (witherite) and RaCO3 should be considered as a potential process controlling Ra solubility. As already mentioned above, the ionic radii of Ra2+ and Ba2+ are similar due to which Ra can be accommodated also in the witherite lattice.

Ra can also be incorporated in strontianite and form a (Sr,Ra)CO3 solid solution. Besides the formation of so-called binary solid solutions, also ternary systems exist (e.g.: [Ba,Sr,Ra]CO3 or [Ba,Sr,Ra]SO4). For further details concerning the formation of the above described sulphate and carbonate solid solutions, it is referred to Kulik and Curti (2005) and Bruno et al. (2007).

Experimental data No in-house or BC specific data available.

Role of (organic) colloids/polynuclear species Not relevant.

Solubility data provided to PA Source Range (SR)

The source range for Ra comprises the Ra-phases RaSO4(s) and RaCO3(s). The former phase represents the LL of the SR, while the UL is equal to the one of the ER.

Best Estimate (BE) Radium carbonate is considered to represent the most relevant solubility controlling solid for Ra in the far-field.

Expert Range (ER)

The Expert Range is delimited by the uncertainties calculated for RaCO3(s), representing the BE.

Data retained in the DCF’s: -5 BE: Thermodynamic solubility of RaCO3(s): 6.7 x 10 M. ER: BE including thermodynamic uncertainty.

-6 SR: Lower limit (LL): Thermodynamic solubility of RaSO4(s): 7.0 x 10 M. Upper limit (UL): Taken equal to ER.

Eh/pH/pCO2 sensitivity of the solid phases

The solubility of RaCO3(s) would be affected by changes in pH as well as pCO2.

Remarks There are no Ra data in the NEA TDB. The Ra speciation diagrams calculated with the NAGRA/PSI and MOLDATA TDB’s look the same, but thermodynamic data of the RaCO3(aq) species are slightly different. Besides this, + MOLDATA comprises the RaHCO3 species, which is not present in the NAGRA/PSI TDB.

+ It should be mentioned, that in MOLDATA the logβ for RaHCO3 was changed according to the ThermoChimie 7b version, due to which the ANDRA diagram and MOLDATA diagram are different.

5.18 Rubidium (Rb) General Rubidium belongs to the alkali metal group and in total 26 isotopes of rubidium (atomic number: 37) are known. Naturally occurring rubidium is made of two isotopes, i.e. 85Rb (abundance ~72%) and 87Rb (abundance 28%). The latter is radioactive and has a half-life of 4.9 × 1010 years. It decays to stable 87Sr by emission of a negative beta particle. Rubidium metal has a high reactivity towards oxidation leading to the subsequent formation of the rubidium cation Rb+, which is very stable, and is normally unreactive towards further oxidative or reductive chemical reactions. Rubidium, like sodium and potassium, is almost always in its +1 oxidation state when dissolved in water.

Speciation & solubility calculations As can be seen in Figure 18, the speciation of Rb is very simple and the uncomplexed Rb+ cation is the dominant species over the entire Eh and pH range.

Rb+ Eh (volts) Eh 0

–.5

25°C 0 2 4 6 8 10 12 14 pH Figure 18: Eh-pH diagram of rubidium (Rb-C-S-O-H) for the BC reference porewater system. Assumedactivity of dissolved [Rb] = 10-8. Diagram a) MOLDATA/NEA TDB. Code: The Geochemist's Workbench - 8.08.

Experimental data Solubility data are not relevant.

Role of (organic) colloids/polynuclear species Not relevant.

Solubility data provided to PA The rubidium concentration is not solubility limited.

Remarks No further remarks.

5.19 Samarium (Sm) General Samarium (atomic number: 62) is a rare earth metal and belongs to the group of lanthanides. Twenty-one isotopes of Sm are known. Naturally occurring samarium is composed of four stable isotopes, 144Sm, 150Sm, 152Sm and 154Sm, and three extremely long-lived radioisotopes, 147Sm (1.06 × 1011 a), 148Sm (7 × 1015 a) and 149Sm (> 2 × 1015 a), with 152Sm being the most abundant (26.75% natural abundance). 151Sm has a half-life of 90 years, and 145Sm has a half-life of 340 days. All of the remaining radioisotopes have half-lives that are less than 2 days, and the majority of these have half-lives that are less than 48 seconds. The long lived isotopes, 146Sm, 147Sm, and 148Sm primarily decay by alpha decay to isotopes of neodymium (Wikipedia). Samarium is not found free in nature, but, like other rare earth elements, is contained in many minerals, i.e. 3+ monazite [Ce(REE)PO4], bastnäsite [(REE)CO3F], samarskite (Y0.2REE0.3Fe 0.3U0.2Nb0.8Ta0.2O4) 3+ 3+ and allanite [A2M3Si3O12(OH)]. The A sites may contain Ca, Sr, REE and M sites Al , Fe , Mn3+, Fe2+, Mg2+, including additional elements, such as Th, U, Zr, P, Ba, Cr and others. The dominant oxidation state of Sm (and REE in general) in aqueous solution at 25°C is the trivalent state (Spahiu and Bruno, 1995). However, under extremely reducing conditions, Sm may also exist in the divalent state. REE(III) ions form strong complexes with carbonate, phosphate, hydroxide, fluoride and sulphate, with carbonate complexes being generally the dominant soluble species of the rare earths in natural waters (Wood, 1990; Lee and Byrne, 1992; Choppin, 1984).

a) b) 1 1

.5 .5 +++ Sm+++ Sm

+ + SmCO3 SmCO3 - - Eh (volts) Eh Eh (volts) Eh Sm(CO ) Sm(CO ) 0 3 2 0 3 2 - - SmO2 Sm(OH)4 µ µ

–.5 –.5

25°C 25°C 0 2 4 6 8 10 12 14 02468101214 pH pH

c) d) 1 1

.5 +++ .5 +++ Sm Sm

+ Sm2(CO3)3(s) SmCO3

- (volts) Eh Eh (volts) Eh Sm(CO ) SmOHCO3(s) 0 3 2 0 - Sm(OH) (s) Sm(OH)4 3 µ µ

–.5 –.5

25°C 25°C 0 2 4 6 8 10 12 14 0 2 4 6 8 10 12 14 pH pH

e) f) 1 1

.5 Sm+++ .5 Sm+++

Sm2(CO3)3(s) + Eh (volts) SmCO3 - 0 Sm(CO3)2 Eh (volts) Sm(OH)3(s) 0 Sm(OH)3(s) µ

–.5 µ

25°C –.5 02468101214 25°C pH 02468101214 pH Figure 19: Eh-pH diagram of samarium (Sm-C-S-O-H) for the BC reference porewater system. Assumed activity of dissolved [Sm] = 10-8. Diagram a) LLNL TDB, b) ANDRA, c) MOLDATA TDB, -3 d) same diagram as c, but solid phases included and [Sm] = 10 , e) same diagram as d, but SmOHCO3(s) suppressed, f) same as diagram e, but Sm2(CO3)3(s) suppressed. Code: The Geochemist's Workbench - 8.08/8.10

Table 18: Solubility of Sm in the BC reference porewater system at 25 °C, pH 8.355 and Eh -281 mV. Database: MOLDATA. Code: The Geochemist's Workbench- 8.08/8.10. Solubility controlling phases Solubility, [Sm], mol/l -2 Sm(OH)3 (am) 4.9 × 10 -4 Sm(OH)3 (s) 3.7  10 -7 Sm2(CO3)3 (s) 5.0  10 -8 SmOHCO3 (s) 8.9 × 10 Source data: ANDRA TDB

The reaction constants for the Sm-solids comprised in (Table 18) and MOLDATA are the following: + 3+ Sm(OH)3(am) + 3 H ↔ Sm + 3 H2O log K = 18.6 + 3+ Sm(OH)3(s) + 3 H ↔ Sm + 3 H2O log K = 16.5 + 3+ - Sm2(CO3)3(s) + 3 H ↔ 2 Sm + 3 HCO3 log K = -3.52 + 3+ - SmOHCO3(s) + 2 H ↔ Sm + HCO3 + H2O log K = 2.63

Speciation & solubility calculations The predicted dominant species of Sm (as for Am) under BC conditions and using MOLDATA is - the second carbonate complex, i.e. Sm(CO3)2 (Figure 19 c). Speciation over the pH range 0 - - ~11.5 using the LLNL, ANDRA and MOLDATA TDB is the same, but at pH > 11.5, Sm(OH)4 represents the prevalent species when calculating with MOLDATA and ThermoChimie v.5, while - SmO2 is the dominant species using the LLNL TDB. At [Sm] = 10-8, no solid phase is predicted to be stable. Increasing the Sm activity to [Sm] = 10-7, SmOHCO3(s) becomes stable under BC conditions. In Figure 19 d, the calculated Pourbaix -3 diagram at [Sm] = 10 is illustrated. Suppressing SmOHCO3(s) in the calculations, Sm2(CO3)3 becomes stable and when this phase is also suppressed, the least stable, but kinetically most

favourable phase, i.e. Sm(OH)3(s) is shown in the diagram. Thus the following precipitation sequence can be anticipated:

At [Sm] = 10-3:

1) Sm(OH)3(s) kinetically most favourable phase → 2) Sm2(CO3)3(s)

→ 3) SmOHCO3(s) most stable phase

At [Sm] = 10-6, the sequence would be:

1) Sm2(CO3)3(s) → 2) SmCO3OH(s)

-7 At [Sm] = 10 : → SmCO3OH(s)

- Assuming equilibrium with SmOHCO3(s), calculations predict Sm(CO3)2 to be the dominant + species (93 %) besides SmCO3 (7 %).

Experimental data No in-house or BC specific data available.

Role of (organic) colloids/polynuclear species By analogy with Am, the formation of intrinsic colloids may lead to higher apparent solubilities. Sm-NOM colloid association has been put forward by the new phenomenological model (Bruggeman et al., in prep.). For further details see also paragraph 4.3.8.

Solubility data provided to PA Solid Ln/REE carbonates and hydroxicarbonates have been proposed by Choppin (1984, 1986) to represent the solubility limiting phases in natural waters. Ln/REE oxides are easily hydrated and quite soluble in water. Due to the high solubility of solid Sm(III) sulphates, chlorides, bromides, nitrates and iodates, these phases are generally not considered being relevant as solubility limiting 3+ 3+ phases for safety assessment reasons (Spahiu and Bruno, 1995). The Ln /REE -CO2(g)-H2O system has been studied by many authors, but experimental data concerning the solubility of solid Ln/REE carbonates and hydroxicarbonates are very scattered. Main reason for this scattering, is the possibility of phase transformations between Ln/REE-hydroxides, Ln/REE-hydroxi-carbonates and/or normal carbonates. The main parameter influencing these transformations, is the partial pressure of CO2(g) of the system, but also temperature may play a role. Results of different studies suggest that Ln/REE-hydroxides [Ln/REE(OH)3)] at pCO2 > 0.1 atm transform into normal -3.5 carbonates [(Ln/REE)2(CO3)3], whereas under atmospheric pCO2 conditions (10 atm), the formation of mixed Ln/REE hydroxi-carbonates is favoured. The latter transformation has been only observed for the light REE (LREE), while for the heavy REE (HREE) this transformation process (hydrolysis) seems to be kinetically hindered or much slower. The "easier" formation of hydroxi-carbonates with LREE was already stated by Caro and co-worker at the end of the sixties (1968).

Source Range (SR)

Except Sm(OH)3(am), all phases comprised in Table 18 have been retained in the SR. The amorphous samarium hydroxide was rejected from the selection due to its high solubility under BC conditions.

Best Estimate:

SmOHCO3(s) is considered to represent the most relevant solubility controlling solid for Sm in the far-field. This phase was selected in analogy with Am.

Expert Range

SmOHCO3(s) including its thermodynamic uncertainties.

Data retained in the DCF’s: -8 BE: Thermodynamic solubility of SmOHCO3(s): 8.9 × 10 M. ER: BE including thermodynamic uncertainty. SR: Lower limit (LL): Taken equal to ER. -4 Upper limit (UL): Thermodynamic solubility of Sm(OH)3(s): 3.7  10 M.

Eh/pH/pCO2 sensitivity of the solid phases As can be seen in Figure 19 (d-f), the stability of all Sm-solids is independent of Eh, as Sm is a non-redox sensitive element and their precipitation beaviour is only controlled by pH and pCO2 (and the Sm activity of course).

Remarks Sm data are neither comprised in the NEA nor NAGRA/PSI database. Therefore, the Sm data included in MOLDATA were mainly selected from the ANDRA TDB.

5.20 Selenium (Se)

General Selenium (atomic number: 34) is a redox-sensitive element and member of the sulfur family. As such, its chemistry is similare to that of sulfur. Natural Se comprises six stable isotopes, i.e. 74Se, 76Se, 77Se, 78Se, 80Se and 82Se. 79Se is one of the 7 long-lived fission products (LLFP) besides 99Tc, 126Sn, 93Zr, 135Cs, 107Pd and 129I and is considered as the main contributor to the dose-to-man for the disposal of spent fuel and high-level waste (HLW) in Boom Clay (De Cannière et al., 2010; Marivoet et al., 1999). Recently, a new value for the half-life of 79Se, i.e. 3.27 x 105 years was determined by Jörg et al. (2010). Selenium is a rare element in the Earth’s crust and also rarely occurs in its elemental state, but can be found in sulfide ores such as pyrite, where it partially replaces the sulfur. Furthermore, Se occurs naturally in selenides (e.g. FeSe, HgSe, PbSe, ZnSe), selenates, which are analogous to sulfates and have similar chemistry, and selenites (e.g. Ag2SeO3, Na2SeO3). Se is redox-sensitive and occurs in several oxidation states, i.e. 0, –II, -I, +IV and +VI.

a) - b) - HSeO H(SeO4) 1 4 1

H (SeO ) H2SeO3(aq) 2 3 SeO-- -- .5 4 .5 SeO4 - - HSeO3 HSeO3

H Se(aq) -- Eh (volts) Eh (volts) 2 Se4 0 -- 0 H Se SeO3 2 -- SeO3

µHSe- µ HSe- –.5 –.5

25°C 25°C 02468101214 02468101214 pH pH

c) - d) HSeO- HSeO4 4 1 1

H2SeO3(aq) H2SeO3 -- SeO-- .5 SeO4 .5 4 - - HSeO3 HSeO3

H Se Eh (volts) Eh (volts) 2 Se 0 0 4 -- -- H2Se(aq) SeO SeO3 3

µHSe- µ - -- HSe Se3 –.5 –.5

25°C 25°C 02468101214 02468101214 pH pH

- - HSeO4 e) HSeO4 f) 1 1

H2SeO3(aq) H2SeO3(aq) -- -- .5 SeO4 .5 SeO4 - - HSeO3 HSeO3

Se(trigonal)

-- (volts) Eh Eh (volts) Se4 0 -- 0 H Se(aq) -- SeO3 2 SeO3 H2Se(aq) FeSe (cr) HSe- 2 µ µ - -- HSe Se3 FeSe(s) –.5 –.5

25°C 25°C 02468101214 0 2 4 6 8 101214 pH pH

- g) HSeO4 h) 1 - HSeO4 1

H2SeO3(aq) H2SeO3(aq) -- .5 SeO4 HSeO- SeO-- 3 .5 - 4 HSeO3

Eh (volts) Eh Se(trigonal) 0 -- SeO Eh (volts) Se(trigonal) H Se(aq) 3 0 -- 2 SeO3 H2Se(aq) HSe- µ Fe Se (gamma) - 3 4 HSe µ Fe Se (alpha) –.5 7 8 Fe Se(beta) –.5 25°C 1.04 Fe1.04Se(beta) 25°C 0 2 4 6 8 10 12 14 pH 02468101214 pH Figure 20: Eh-pH diagram of selenium (Se-C-S-O-H) for the BC reference porewater system. Assumed activities of dissolved [Se] = 10-8 and [Fe] = 2.2  10-7. Diagram a) LLNL TDB, b) ANDRA TDB, c) NAGRA/PSI TDB, d) NEA TDB, e) MOLDATA TDB f) same diagram as e, but solid phases included in -6 calculations, g) same diagram as f, but [Se] = 10 (FeSe and FeSe2 suppressed in calculation), g) same diagram as g, but Fe3Se4(gamma) suppressed in calculation. Code: The Geochemist's Workbench - 8.08/8.10.

Speciation & solubility calculations Based on the speciation diagrams calculated with MOLDATA (Figure 20), hydrogen selenide (HSe-) is predicted to be the dominant aqueous Se(-II) species under the reducing conditions prevailing in Boom Clay. Assuming a Se activity of 10-8, achavalite FeSe(s) is predicted to be the thermodynamically most stable phase (Figure 20 f), while kinetically ferroselite represents the more favourable Se solubility limting phase. According to Brookins (1988), the latter, i.e. Se analogue of pyrite occurs however only rarely in nature. The stabilities of the iron selenides, i.e. FeSe(s) and FeSe2 depend on the one hand on the Fe concentration present in the porewater and on the other hand on the redox conditions. This Eh dependence is clearly visible in Figure 20 f. Two elemental selenium polymorphs are comprised in MOLDATA, i.e. monoclinic and trigonal selenium. The latter represents the thermodynamically more stable phase in the temperature range from 0 to 494.2 K, while the former phase is metastable and melts at 413 K (Olin et al., 2005). As can be seen in Figure 20 f, elemental selenium is stable at more oxiding conditions, but also higher Se concentrations in solution would be required (> 2 x10-6 M, Table 19) to precipitate this phase.

As the intital speciation of Se in the source term is not known, the presence/release of Se (+VI), in 2- form of selenate (SeO4 ), cannot be ruled out. Due to its very soluble character, no solubility limit is expected for this inorganic species under undisturbed BC conditions in the absence of heavy metals, or large concentrations of earth alkaline cations, such as Ba2+ (De Cannière et al., 2010).

Selenates and selenites are characterized by high solubilities, while metal selenides are similar to sulphides characterized by low solubility products. According to Hummel et al. (2002), the role of the fomer phases as solubility limiting solids in natural systems seems questionable and no phases were proposed in their review, by referring to the NEA review and leaving it to them to deal with the problem of "controversial data".

Depending on the Se-activity, the following precipitation/formation sequence could be anticipated:

At [Se] = 10-8

1) Ferroselite (FeSe2,cr) → 2) Achavalite (FeSe,s)

At [Se]  10-6

1) Se(trig) → 2) Fe3Se4(gamma) → 3) Ferroselite (FeSe2,cr) → 4) Achavalite (FeSe,s)

Table 19: Solubility of Se in the BC reference porewater system at 25 °C, pH 8.355 and Eh -281 mV. Database: MOLDATA. Code: The Geochemist's Workbench- 8.08. Solubility controlling phases Solubility, [Se], mol/l Se (mono)1 2.4  10-6 Se (trig)1 6.5 × 10-7 1 -7 Fe3Se4 (gamma) 1.4 × 10 1 -9 Ferroselite (FeSe2,cr) 3.6  10 Achavalite (FeSe,s)2 1.1 × 10-9 Source data: 1NEA TDB, 2ANDRA TDB

The reaction constants for the selenium solids comprised in (Table 19) and MOLDATA are the following: 2- + Se(mono) + O2(aq) + H2O ↔ SeO3 + 2 H log K = 25.06 2+ 2- Fe3Se4(gamma) + 5.5 O2(aq) + H2O ↔ 3 Fe + 4 SeO3 log K = 233.16 2- + Se(trig) + O2(aq) + H2O ↔ SeO3 + 2 H log K = 24.84 2+ 2- FeSe(s) + 1.5 O2(aq) ↔ Fe + SeO3 log K = 67.52 2+ + 2- FeSe2(cr) + 2.5 O2(aq) + H2O ↔ Fe + 2 H + 2 SeO3 log K = 90.78

Experimental data The solubility of solid iron selenide (FeSe) in synthetic and interstitial Boom Clay water has been extensively studied by KU Leuven and AEA Technology (AEAT). The effect of Eh, organic matter, the presence or absence of pyrite and solid BC on the FeSe solubility was tested in different experimental set-ups. For details concerning these experiments, it is referred to the Se

Topical report (De Cannière et al., 2010). Results of these experiments revealed that FeSe appears to be very sensitive to oxidation, which may result in high apparent solubilities (8 × 10-6 to 2 × 10- 5 M, AEAT) due to the presence and dissolution of oxidation products at the FeSe surface. By applying a pre-leaching step at the beginning of the experiments, a decrease of the solubility by several orders of magnitude (5 × 10-8 to 4.5 × 10-10 M) could be observed (De Cannière, 2010). Similar experiments as for FeSe(s) were performed with elemental Se(s). The independently determined solubilities of Se(s) by AEAT and KU Leuven range between 4 × 10-8 to 3 × 10-7 M and 1.7 × 10-9 to 8 × 10-8, respectively. FeSe(s) and Se(s) solubilities appear not to be influenced by the presence of organic matter.

Role of (organic) colloids/polynuclear species Se(0) is reported to have a strong tendency to form intrinsic colloids (Wersin and Schwyn, 2004).

Solubility data provided to PA Source Range (SR)

Taking native Se and Fe3Se4(gamma) into account or not as Se(-II) solubility controlling minerals, is difficult to judge. According to Duro et al. (2006), the precipitation of native selenium under far-field (low-T) conditions is questionable and thus neglected by them as solubility limiting phase. On the other hand, Hummel et al. (2002) stress that the existence of selenide in reducing subsurface environments is well established and besides this, it is strongly associated with sulfide minerals. Thus, (Se(-II)-control by mixed phases or even solid solutions with sulfide minerals represents "a meaningful alternative to the rather exotic pure selenide minerals". Therefore, it has been decided to include all minerals comprised in Table 19 into the source range of potential Se- solubility controlling minerals.

Best Estimate (BE)

Ferroselite (FeSe2, cr) is considered to represent the most relevant solubility controlling solid for Se in the far-field.

Expert Range (ER) The Expert Range is delimited by the thermodynamic uncertainties associated to the solubility of ferroselite.

Data retained in the DCF’s: -9 BE: Thermodynamic solubility of ferroselite (FeSe2,cr): 3.6  10 M. ER: BE including thermodynamic uncertainty. SR: Lower limit (LL): Equal to ER. Upper limit (UL): Thermodynamic solubility of Se0(mono): 2.4  10-6 M.

Eh/pH/pCO2 sensitivity of the solid phases The stability of the Se-solids discussed above and the solubility of Se are highly Eh and pH dependent, which is clearly revealed by the diagrams of Figure 20. Due to the common association of Se with Fe, as well as sulphur, which are like selenium redox-sensitive elements, make predictions concerning effects of changing conditions very difficult.

Remarks The data for Se selected in MOLDATA are mainly coming from the NEA TDB (NEA review, Vol. 7). FeSe(s) was selected from the ANDRA TDB.

5.21 Silver (Ag) General Silver (atomic number 47) belongs to the group of transition metals and has twenty-eight radioisotopes. Naturally occurring silver is composed of two stable isotopes, 107Ag and 109Ag, with 107Ag being the most abundant (51.839 % natural abundance). Silver (Ag) occurs naturally in its pure form (i.e. native silver), as an alloy with gold and other metals, in antimonides, arsenides, tellurides, selenides and in sulfides, such as argentite (Ag2S) and acanthite (Ag2S), respectively. Acanthite is the low-T modification of silver sulphide and stable phase under atmospheric pressure conditions (high-T modification corresponds to argentite). Ag also forms sulfates and halides, such as chlorargyrite (AgCl,cr). Silver could also be identified as a trace component in carbonates, Fe-oxides, Mn-oxides and organic matter. In the SF and HLW inventory, the metastable 108mAg 110m (t1/2 = 418.3 years) and Ag (t1/2 = 249.9 days) isotopes are comprised, as they represent activation products. The latter isotopes are formed by neutron capture of other materials, such as structural components of the nuclear reactor, the reactor coolant, control rods or other neutron poisons or materials in the environment (Wikipedia). Under oxidizing and acidic conditions (pH < 4) silver shows a relatively high mobility. In neutral to alkaline environments silver would display a medium (moderate) mobility. Ag is accompanied by a suite of chemical elements, such as As, Se, Pb, Bi, and Sb, respectively.

a) b) 1 1

.5 .5 AgCl(aq) AgCl(aq)

Eh (volts) Eh - 0 (volts) Eh AgCO - 0 3 AgCO3 - Ag(OH)2 µ Ag(HS) µ –.5 –.5 25°C 25°C 02468101214 pH 02468101214 pH

c) d) 1 1

AgCl(aq) .5 .5 - AgCO3 - Ag(OH)2 Ag+ Eh (volts) Eh Eh (volts) Eh 0 0 Acanthite Ag(cr) µ µ

–.5 –.5

25°C 25°C 02468101214 0 2 4 6 8 10 12 14 pH pH

e) f) 1 1

.5 AgCl(aq) .5

AgCl(cr) - AgCO3 Eh (volts) Eh 0 - Eh (volts) Ag(OH)2 0 Ag(OH)(s)- Ag(OH)2 Acanthite µ µ Ag(HS) –.5 –.5

25°C 25°C 02468101214 0 2 4 6 8 101214 pH pH g)

.5 Ag+

Ag2(CO3)(s) - no graph - Eh (volts) Eh 0 Ag(OH)(s) µ Ag(HS) –.5

25°C 02468101214 pH

Figure 21: Eh-pH diagram of silver (Ag-C-S-O-H) for the BC reference porewater system. Assumed activity of dissolved [Ag] = 10-8. Diagram a) LLNL TDB, b) ANDRA and MOLDATA TDB, c) NEA TDB, d) MOLDATA solid phases included, e) same diagram as d, but Ag(cr) suppressed, -4.62 2- -8.8 f) MOLDATA TDB [AgHS] = 10 and [SO4 ] = 10 Ag(cr) and acantite suppressed g) MOLDATA TDB [Ag] = -3.6 2- -11.7 -3.3 10 , [SO4 ] = 10 and [Cl] = 10 Ag(cr), acantite and AgCl(cr) suppressed. Code: The Geochemist's Workbench - 8.08 / 8.10.

Speciation & solubility calculations Under BC conditions and using MOLDATA, the aqueous AgHS is predicted to be the dominant species (Figure 21 b, Table 21. If the latter species would be suppressed in the calculations, AgCl(aq) would be the prevalent species. According to Duro et al. (2006), these species are characterized by a high stability and in groundwaters with high chloride contents, silver speciation will be completely dominated by aqueous chloride species.

When plotting the Ag species distribution as function of total Ag concentration in solution (Figure 22), it can be seen that there are two concentration thresholds where the species distribution changes, i.e. with AgHS being dominant at total Ag concentrations < 10-4 M and AgCl(aq) being the dominant species at Ag-concentrations between > 10-4 M up to 2 × 10-4 M. At Ag- concentrations above the latter value, the Ag+ cation represents the dominant aqueous species. This changing speciation as function of total concentration is quite unique and specific for silver

and thus had to be taken into account in the calculations of the Pourbaix diagrams (for details see figure caption).

As can be seen in Figure 21 d, at [Ag] = 10-8, two silver solids are stable, i.e. crystalline silver and acanthite (Ag2S). These solids are characterized by quite low and similar solubilities (Table 20). -4 At higher Ag activities ( 5.6 × 10 ), two other solids, i.e. AgCl(cr) and Ag2CO3(s) are stabilized and appear in the Pourbaix diagram (Figure 21 f, g). Based on the performed calculations, the following precipitation/formation sequence can be anticipated:

In case [Ag] > 2.6 × 10-4:

1) Ag2CO3(s) → 2) AgCl(cr) → 3) Ag2S (acanthite) → 4) Ag(cr) most stable

At [Ag] ≤ 10-8, only the latter two phases may control the Ag concentrations in solution.

Table 20: Solubility of Ag in the BC reference porewater system at 25 °C, pH 8.355 and Eh -281 mV. Database: MOLDATA. Code: The Geochemist's Workbench- 8.08 / 8.10. Solubility controlling phases Solubility, [Ag], mol/l 2 -4 Ag2CO3(s) 5.6 × 10 Chlorargyrite (AgCl, cr)1 'horn silver' 2.5  10-5 2 -13 Acanthite (Ag2S) 5.4  10 1 Ag (cr) 3.7  10-13 Source data: 1NEA TDB, 2ANDRA TDB

The reaction constants for the silver solids comprised in (Table 20) and MOLDATA are the following: + + - Ag2CO3(s) + H ↔ 2 Ag + HCO3 log K = -0.73 AgCl(cr) ↔ Ag+ + Cl- log K = -9.75 + 2- Ag2S + 2 O2(aq) ↔ 2 Ag + SO4 log K = 102.21 + + Ag(cr) + 0.25 O2(aq) + H ↔ Ag + 0.5 H2O log K = 7.99

Table 21: Species distribution of Ag in equilibrium with AgCl(cr). Database: MOLDATA. Code: The Geochemist's Workbench- 8.08 / 8.10. Aq. species [mol/L] Fraction [%] AgHS 2.41 × 10-5 97.2 AgCl(aq) 3.31×10-7 1.3 Ag+ 3.17×10-7 1.3 - -8 AgCl2 2.41×10 0.1 - -8 AgCO3 2.04×10 0.1 Total 2.48 × 10-5 100

Figure 22: Silver speciation as function of total dissolved Ag concentration

Experimental data No in-house or BC specific data available.

Role of (organic) colloids/polynuclear species Ag-NOM colloid association has been put forward in the newly developed phenomenological model (Bruggeman et al., in prep.).

Solubility data provided to PA Source Range (SR) All phases comprised in Table 20 were retained in the SR.

Best Estimate It is assumed that AgCl(cr) represents the kinetically most probable solubility limiting phase.

Expert Range (ER) Solubility of chlorargyrite (AgCl,cr) including its thermodynamic uncertainties.

Data retained in the DCF’s: BE: Thermodynamic solubility of chlorargyrite, AgCl(cr): 2.5 × 10-5 M. ER: BE including thermodynamic uncertainties. SR: Lower limit (LL): Thermodynamic solubility for metallic Ag(cr): 3.7  10-13 M. -4 Upper limit (UL) Thermodynamic solubility of Ag2CO3(s): 5.6 × 10 M.

Eh/pH/pCO2 sensitivity of the solid phases As revealed by the Pourbaix diagrams in Figure 21, are the stabilities of all phases more or less pH and Eh dependent. Changes in pCO2 are assumed to mainly affect Ag2CO3(s).

100

Remarks No Ag data are comprised in the NAGRA/PSI TDB. In the NEA TDB, Ag is only comprised as auxiliary species. Therefore, the Pourbaix diagram calculated with the NEA TDB (Figure 21 c) is not very meaningful. Silver species incorporated in MOLDATA were mainly selected from the ANDRA TDB.

101

5.22 Strontium (Sr) General Strontium is an alkaline earth metal and its atomic number is 38. It commonly occurs in nature, mainly in the form of sulphate, i.e. celestite (SrSO4) or carbonate, i.e. strontianite (SrCO3). Natural strontium is a mixture of four stable isotopes: 84Sr (0.56%), 86Sr (9.86%), 87Sr (7.0%) and 88Sr (82.58%). Only 87Sr is radiogenic. It is produced by decay from the radioactive alkali metal 87Rb, which has a half-life of 4.88 × 1010 years. 87Sr/86Sr ratios are commonly used to determine the likely provenance areas of sediments, especially in marine and fluvial environments. The isotope of major concern with respect to radioactive waste disposal is the unstable 90Sr isotope, which has a half-life of 29.14 years and represents a medium-lived fission product (MLFP). 90Sr has a fission yield of ~4.5 % and together with 137Cs are responsible for the highest radiation produced by MLFP several years after SF cooling.

a) b) 1 1

.5 .5 ++ Sr Sr++ Eh (volts) Eh (volts) Eh 0 0 SrCO (aq) 3 Sr(CO3) µ µ

–.5 –.5

25°C 25°C 02468101214 02468101214 pH pH

c) d) 1 1

.5 .5

Sr++ Sr++ Eh (volts) Eh (volts) 0 0 SrCO3 µ µ

–.5 –.5

25°C 25°C 02468101214 02468101214 pH pH Figure 23: Eh-pH diagram of strontium (Sr-C-S-O-H) for the BC reference porewater system. Assumed activity of dissolved [Sr] = 10-8. Diagram a) LLNL TDB, b) ANDRA and MOLDATA TDB, c) NAGRA/PSI TDB, d) NEA TDB. Code: The Geochemist's Workbench - 8.08/8.10

Speciation & solubility calculations As revealed by Figure 23 (all TDB’s), the uncomplexed Sr2+ cation represents the dominant aqueous species under the reference conditons.

102

Solubility calculations were performed for strontianite and celestite and the results are summarized in Table 22. Celestite is very soluble under BC conditions, so that it was not included in the solubility assessment for the far-field (see below).

Table 22: Solubility of Sr in the BC reference porewater system at 25 °C, pH 8.355 and Eh -281 mV. Database: MOLDATA. Code: The Geochemist's Workbench- 8.08 Solubility controlling phases Solubility, [Sr], mol/l 3 -6 Strontianite (SrCO3) 8.0  10 2 Celestite (SrSO4) very soluble Source data: 2ANDRA TDB, 3NAPSI TDB

The reaction constants for the strontium solids comprised in (Table 22) and MOLDATA are the following: + 2+ - SrCO3(s) + H ↔ Sr + HCO3 log K = 1.05 2+ 2- SrSO4(s) ↔ Sr + SO4 log K = -6.64

Experimental data No in-house or BC specific solubility data available.

Solubility data provided to PA Source Range (SR)

Solubility of strontianite (SrCO3) including its thermodynamic uncertainties.

Best Estimate

Strontianite (SrCO3) is considered to represent the most likely solubility limiting phase under far- field conditions.

Expert Range (ER) Taken equal to SR.

Data retained in the DCF’s: -6 BE: Thermodynamic solubility of strontianite (SrCO3) : 7.9 × 10 M. ER: BE including thermodynamic uncertainty. SR: Taken equal to ER.

Eh/pH/pCO2 sensitivity of the solid phases

The solubility of carbonate minerals, such as SrCO3(s), siderite and/or calcite are affected by changes in pH as well as pCO2.

Remarks In the NEA TDB Sr is only comprised as auxiliary species. The Pourbaix diagram (Figure 23 d) is therefore not very meaningful.

103

5.23 Technetium (Tc) General Technetium belongs to the group of transition metals and its atomic number is 43. Technecium has no stable isotopes, all known isotopes are radioactive. 99Tc is the most common and significant isotope and produced by fission of 235U and 239Pu (fission yield ~6%) in nuclear reactors. It is one of the 7 long-lived fission products (half-life 2.14 × 105) besides 126Sn, 79Se, 93Zr, 135Cs, 107Pd, and 129I. Besides 99Tc, 98Tc (half-life of 4.2 Ma) and 97Tc (half-life: 2.6 Ma) are the most stable radioisotopes. Technetium was the first element to be produced artificially. Since its discovery, searches for the element in terrestrial material have been made without success. If it does exist, the concentration must be very small (Handbook of Chemistry and Physics, 1992-1993). Tc is not redox stable and occurs in several oxidation states, i.e. 0, IV, V (unstable), VI (unstable), and VII. The presence of tri- and tetravalent Tc species is uncertain and thermodynamic data are lacking and/or not recommended (Hummel et al., 2002). Tc(IV) and Tc(VII) are the chemically most relevant oxidation states. Important aqueous species of Tc(IV) are the neutral species TcO(OH)2, - TcO2(aq) and Tc(OH)4(aq), respectively, while the pertechnetate anion (TcO4 ) is the most significant Tc(VII) species. With respect to solid phases, Tc(IV) minerals are generally sparingly soluble, while Tc(VII) minerals are highly soluble. Rhenium (Re) is reported to be chemically very similar to Tc due to which it is often used as an analogue for Tc. For further details on the geochemical (and transport) behaviour, it is referred to the Topical Report on Tc by Bruggeman et al. (2011).

a) b) 1 1

.5 .5 TcO++ - TcO- TcO4 4 TcOOH+ (TcO)(OH)+ Eh (volts) Eh

Eh (volts) Eh +++ 0 Tc 0

TcO(OH)2(aq) TcO(OH) (aq) µ2 µ

–.5 –.5 - (TcO)(OH)3 25°C 25°C 02468101214 02468101214 pH pH

c) d) 1 1

.5 .5 - - TcO4 TcO4 TcO++ TcO++ TcO(OH)+ TcO(OH)+ Eh (volts) Eh Eh (volts) Eh 0 0 TcCO3(OH)2 TcCO3(OH)2(aq)

µ TcO(OH)µ (aq) TcO(OH)2 2 –.5 –.5 - TcO(OH)- TcO(OH)3 3 25°C 25°C 02468101214 02468101214 pH pH

104

e) f) 1 1

.5 .5 - - TcO4 TcO4 TcO++ TcO(OH)+ Eh (volts) Eh (volts) Eh 0 +++ 0 TcO (cr) Tc TcCO3(OH)2(aq) 2

µ Tc(cr) µ TcO(OH)2(aq) –.5 –.5 - TcO(OH)3 25°C 25°C 02468101214 02468101214 pH pH g) 1

.5 - TcO4 TcO++ TcO(OH)+

Eh (volts) Eh - no graph - +++ 0 Tc TcCO3(OH)2(aq)

TcO2µ:1.6 H2O(s)

–.5 - TcO(OH)3 25°C 02468101214 pH

Figure 24: Eh-pH diagram of technetium (Tc-C-S-O-H) for the BC reference porewater system. Assumed activity of dissolved [Tc] = 10-8. Diagram a) LLNL TDB, b) ANDRA TDB, c) NAGRA/PSI TDB, d) NEA TDB, e) MOLDATA TDB, f) same diagram as e, but solid phases included (sulfide phases, i.e. TcS2, TcS3, TcS7 suppressed in calculation, g) same diagram as f, but TcO2(cr) and Tc(cr) suppressed in calculation. Code: The Geochemist's Workbench -8.08/8.10.

Speciation & solubility calculations Speciation calculations using MOLDATA (Figure 24 e) show, that TcO(OH)2 is the dominant species under BC conditions. Under more oxidizing conditions, the Tc(VII) pertechnetate species - (TcO4 ) becomes stable. Solubility calculations were performed for the solid phases comprised in Table 23. Under the reference conditions, TcO2(cr) represents the most stable phase (Figure 24 f). When suppressing the latter solid in the calculation, TcO2 × 1.6 H2O (s) appears in the diagram, indicating that the hydrated oxide is the kinetically favoured phase to form.

105

Table 23: Solubility of Tc in the BC reference porewater system at 25 °C, pH 8.355 and Eh -281 mV. Code: The Geochemist's Workbench- 8.0. Database: MOLDATA Solubility controlling phases Solubility, [Tc], mol/l -9 TcO2 × 1.6 H2O (s) 4.4 × 10 Tc (cr) 2.0  10-11 -13 TcO2 (cr) 4.5  10 Source data: all solid phase data were taken from the NEA TDB.

The reaction constants for the technetium solids comprised in (Table 23) and MOLDATA are the following: - + TcO2 × 1.6 H2O + 0.75 O2(aq) ↔ TcO4 + H + 1.1 H2O log K = 26.66 - + Tc(cr) + 0.5 H2O + 1.75 O2(aq) ↔ TcO4 + H log K = 95.96 - + TcO2(cr) + 0.75 O2(aq) + 0.5 H2O ↔ TcO4 + H log K = 22.67

Experimental data The solubility of chemically preciptated hydrous Tc(IV) oxide in SBCW was determined at KU Leuven (Bruggeman, 2002) to range between 1.8 × 10-7 M to 4.6 × 10-7 M (without UF), which is around 2 orders of magnitude higher than the calculated thermodynamic solubility of TcO2 × 1.6 H2O (see Table 23). The higher solubility was referred to the possibility that either Tc(VII) was incorporated in the solid and/or a more hydrated and/or amorphous hydrous oxide was used. As no ultrafiltration was performed in the experiments, the higher "apparent" solubilities could also be related to the presence of intrinsic colloids.

Role of (organic) colloids/polynuclear species An exhaustive experimental programme on Tc was performed over the last several decades by SCK•CEN, AEA Technology and KU Leuven. The formation of pseudocolloids between Tc(IV) (monomers, polymers and/or colloids) and dissolved BC organic matter has been evidenced by different institutes (SCK•CEN, KU Leuven), techniques (size exclusion chromatography, X-ray absorption spectroscopy) and authors (Maes et al., 2003; Maes et al., 2004). Based on the different experimental results and observations, a so-called "colloid-colloid interaction model" was developed by Maes et al. (2001, 2003), which has been evolved during the last years to the current phenomenological model Bruggeman et al, in prep.), in which kinetics are taken into account to describe the inorganic and organic Tc-interaction and migration processes.

Solubility data provided to PA Source Range (SR) All phases comprised in Table 23 were retained in the source range. The upper limit of the SR was enlarged based on the experimental data described above.

Best Estimate

TcO2 × 1.6 H2O(am) is considered to represent the most likely solubility limiting phase for Tc under far-field conditions.

Expert Range (ER)

Solubility of TcO2 × 1.6 H2O (am) including its thermodynamic uncertainties.

106

Data retained in the DCF’s: -9 BE: Thermodynamic solubility of TcO2 × 1.6 H2O(s): 4.4 × 10 M. ER: BE including thermodynamic uncertainty. -13 SR: Lower limit (LL): Thermodynamic solubility of TcO2(cr): 4.5  10 M. Upper limit (UL): Defined based on experimental observations: 4.6 × 10-7 M.

Eh/pH/pCO2 sensitivity of the solid phases

Solubility of hydrous Tc oxide (TcO2 × H2O) is pH independent in the pH range 3-10, but increases beyond (Hummel et al., 2002). However, under higher pCO2 conditions (i.e. in presence of carbonates), the latter phase is reported to be more soluble (even) in the pH range 6.3 to 8.6 due to the formation of hydroxide-carbonate complexes (Hummel et al., 2002). Thus, the influence of pCO2 should be included in SA considerations.

Remarks The update of technetium data by Guillaumont et al. (2005, Vol. 5) has not resulted in any changes of selected values compared to the compilation of Rard et al. (1999, Vol. 3). In MOLDATA the cation Tc3+ from the LLNL TDB has been also incorporated, but it should be mentioned that this species is neither comprised in the ANDRA nor in the NEA TDB.

107

5.24 Thorium (Th) General Thorium belongs to the actinide series of the periodic table and its atomic number is 90. All natural thorium is found as 232Th, which decays to stable 208Pb with a half-life of 1.4 × 1010 years. The abundance of Th in nature is estimated to be 3-4 times as high as uranium. Thorium has around thirty radioisotopes with the most stable (after 232Th) being 230Th with a half-life of 75,380 years (daughter product of 238U), 229Th with a half-life of 7,340 years, and 228Th with a half-life of 1.92 years. All of the remaining radioactive isotopes have half-lives that are less than one month (Wikipedia). Thorium metal is a source of nuclear power. Unlike natural uranium, which contains ~0.7% ‘fissile’ 235U isotope, natural thorium does not contain any ‘fissile’ material and is made up of the ‘fertile’ 232Th isotope only. During the pioneering years of nuclear energy, i.e. from the mid 1950s to mid 1970s, there was considerable interest worldwide to develop thorium fuels and fuel cycles in order to supplement uranium reserves. Due to the discovery of new U-deposits and their improved availability during the 80s and 90s, the interest declined despite the fact that the feasibility to use Th in different reactor types could be demonstrated. However, in recent times, the need for proliferation-resistance, longer fuel cycles, higher burnup, improved waste form characteristics, reduction of plutonium inventories and in situ use of bred-in fissile material has led again to renewed interest in thorium-based fuels and fuel cycles in several developed countries (IAEA, 2005). In Belgian power plants however, Th-based fuels were never used and are not foreseen to be used in future. Thorium is present in several silicates and oxides, such as for example thorite (ThSiO4) and thorianite (ThO2), and occurs also in traces in different heavy minerals, e.g. monazite and zircon.

a) b) 1 1

++ Th(SO )++ ThSO4 4 .5 .5 Th++++ ++ Th(OH)2 ++ Th(OH)2 ++++ - Eh (volts) Eh Th(OH) (aq) (volts) Eh Th Th(CO )(OH) 0 4 0 3 3 Th(OH)+++ Th(OH)4 µ µ

–.5 –.5

25°C 25°C 02468101214 02468101214 pH pH

108

c) 1 d) 1

++ ThSO4 ++ .5 ThSO4 .5 ++++ Th ++ Th(OH)2 ++++ - Eh (volts) Th ThCO3(OH)3 Th(OH) (CO )-- 0 (volts) Eh 2 3 2 +++ ThOH 0 +++ Th(OH) Th(OH)4 Th(OH) (aq) µ 4 µ –.5 –.5 25°C 25°C 02468101214 pH 0 2 4 6 8 10 12 14 pH e) f) 1 1

++ Th(SO4)2 ThSO4 .5 .5 Th++++ ++ Th(OH)2

- ++++

Eh (volts) Eh ThO (cr) Eh (volts) Eh Th(CO )(OH) Th 2 0 3 3 0 Th(OH)+++ Th(OH)4(aq) µ µ

–.5 –.5

25°C 25°C 0 2 4 6 8 10 12 14 0 2 4 6 8 10 12 14 pH pH

g) h) 1 1

Th(SO4)2 Th(SO4)2 .5 .5

Eh (volts) ++++

Eh (volts) Eh ++++ Th ThO (am,hyd,aged) Th ThO2(am,hyd,fresh) 0 2 0 +++ Th(OH)+++ Th(OH) ++++ Th4(OH)12 µ µ

–.5 –.5

25°C 25°C 0 2 4 6 8 101214 02468101214 pH pH Figure 25: Eh-pH diagram of thorium (Th-C-S-O-H) for the BC reference porewater system. Assumed activity of dissolved [Th] = 10-8. Diagram a) LLNL TDB, b) ANDRA TDB, c) NAGRA/PSI TDB, d) NEA TDB, e) MOLDATA TDB, f) same diagram as e, but solid phases included (ThS(cr), ThS2, Th7S12 -4 and Th2S3 suppressed), g) same diagram as f, but [Th] = 10 and ThO2(cr) suppressed, h) same diagram as g, but ThO2(am,hyd,aged) suppressed. Code: The Geochemist's Workbench - 8.08 / 8.10

109

Speciation & solubility calculations The dominant aqueous species under the BC reference conditions is the mixed hydroxo-carbonate - species, i.e. Th(CO3)(OH)3 (see Figure 25 e). The latter species is not included in the LLNL TDB, by which ThO4(aq) is predicted to represent the dominant species under the conditions of interest. Also within the NEA TDB this mixed hydroxo-carbonate complex is not included. Using the latter 2- TDB for the speciation calculations, Th(OH)2(CO3)3 represents the prevalent species. Solubility calculations were performed for the phases comprised in Table 24. It can be seen that the solubilties vary over several orders of magnitude. Considering an activity of [Th] = 10-8, -4 thorianite represents the most stable phase (Figure 25 f). At higher Th activities (10 ), ThO2 (am,hyd,aged) and ThO2 (am,hyd,fresh) appear successively in the Pourbaix diagram (Figure 25 g, h).

Table 24: Solubility of Th in the BC reference porewater system at 25 °C, pH 8.355 and Eh -281 mV. Code: The Geochemist's Workbench- 8.08. Database: MOLDATA Solubility controlling phases Solubility, [Th], mol/l -5 ThO2 (am,hyd,fresh) 1.6 × 10 -6 ThO2 (am,hyd,aged) 2.6 × 10 -13 Thorianite (ThO2,cr) 4.7  10 Source data: all solid phase data were taken from the NEA TDB.

The reaction constants for the thorium solids comprised in (Table 24) and MOLDATA are the following: + 4+ ThO2 (am,hyd,fresh) + 4 H ↔ Th + 2 H2O log K = 9.3 + 4+ ThO2 (am,hyd,aged) + 4 H ↔ Th + 2 H2O log K = 8.5 + 4+ Thorianite (ThO2,cr) + 4 H ↔ Th + 2 H2O log K = 1.77

Experimental data Liu et al. (in prep.) studied the solubility ThO2(cr) in presence and absence of organic matter under BC conditions. Results of these experiments performed in SBCW (no OM) revealed solubilities of 5 × 10-11 M, 4 × 10-10 M and 8 × 10-8 M after 30 000 MWCO, 300 000 MWCO and 0.45µm filtration. The ThO2(cr) solubility given in the literature at near-neutral conditions (in absence of carbonates) range between ~10-17 and 10-14 M. The calculated solubility using -13 MOLDATA for ThO2(cr) under BC conditions corresponds to 4.7  10 M. The experimental "apparent" solubilities are several orders of magnitude higher, which is most probably related to the presence of amorphous "intrinsic"colloids (i.e. Th(OH)4,am). This interpretation is consistent with many observations, where experimental solubility data exceeding the thermodynamic solubility of ThO2(cr) are referred to the slow dissolution kinetics on the one hand and to the dissolution of small amounts of amorphous parts present within the crystal or at its surface, leading to solubilities that approach those of the amorphous solid. In absence of carbonates, the latter is reported to range between ~10-8 – 10-9 M, which is close to the experimental solubility after ultrafiltration (30 kD). However, the experiments were performed in presence of carbonates might result in higher solubilities. The calculated solubilties for hydrated, freshly precipitated and aged amorphous ThO2 are 1.6 ×10-5 and 2.6 × 10-6 M (in presence of carbonates), respectively. The latter values are 2-3 orders of magnitude higher than the measured Th concentration after ultrafiltration. Although the interpretation of the experiments is not yet finalized and clear, from database point of view, the - thermodynamic data for the ThCO3(OH)3 species might be overestimated (too stable), resulting in

110

too high solubilities. Further experiments are currently ongoing to get deeper insight about the effect of filtration, intrinsic colloid formation and carbonate complexation on solubility.

In presence of OM, the measured Th concentrations increased with increasing OM. Compared to the solubilites measured in absence of OM, 3-4 times higher solubilities were determined in experiments where 50-100 mg/L (or higher) OM concentrations were used. The latter two facts and same dependence of solubility on the operational size cut-off in presence of NOM, clearly point to a Th-HA/colloid like interaction.

Role of (organic) colloids/polynuclear species As described in the paragraph above, organic and inorganic colloids may increase the solubility of ThO2(cr) to values approaching the amorphous phase. Tetravalent actinides are known to have a strong tendency to hydrolyse, representing the first step to polymerization/polynucleation and intrinsic colloid formation. For further details see also sections 4.3.6 and 4.3.8, respectively. Th-NOM colloid association has been also put forward in the newly developed phenomenological model (Bruggeman et al., in prep.).

Solubility data provided to PA Source Range (SR) All phases comprised in Table 24 were retained in the source range.

Best Estimate

ThO2(am, hyd,aged) is considered to represent the most likely solubility limiting phase for Th under far-field conditions.

Expert Range (ER)

ThO2(am,hyd,aged) including its thermodynamic uncertainties. Lower limit was enlarged based on experimental data (see above).

Data retained in the DCF’s: -6 BE: Thermodynamic solubility of ThO2(am,hyd,aged): 2.6 × 10 M. ER: BE including thermodynamic uncertainty. Lower limit (LL) was enlarged to include experimental data. -13 SR: Lower limit (LL): Thermodynamic solubility of (ThO2,cr): 4.7 × 10 M Upper limit (UL): Taken identical to ER.

Eh/pH/pCO2 sensitivity of the solid phases At pH < 6 (no carbonates), amorphous Th(IV) precipitates are characterized by high solubilities, which is referred to the presence of polynuclear species and small (real) colloids. Solubility data for crystalline ThO2 are generally lower than for amorphous ThO2, but are also strongly pH dependent at values < 7 (Neck and Kim, 2001). Besides this, the pretreatment and alteration degree of the solid under consideration plays an important role on the solubility at pH between 0 and 3.

111

At pH > 6, the solubility of amorphous Th-oxides and hydroxides (ThO2 and Th(OH)4) has been reported to become pH-independent (Ryan and Rai, 1987; Neck and Kim, 2001; ThermoChimie documentation). The same behaviour was observed for crystalline ThO2 (Moon, 1989; Neck and Kim, 2001). As Th is not redox-sensitive, changing redox conditions is assumed to have no or only minor effect on the solubility of Th-oxides. The influence of different partial CO2 pressures on the ThO2(am,hyd) solubility was studied by Östhols et al. (1994). A steep increase in solubility from pH > ~5.5 was observed, which was - related to the formation of mixed Th-hydroxide-carbonate complexes (Thm[OH]4m-1CO3 ). Complementary studies performed by Altmaier et al. (2005) revealed that aqueous speciation is - 2- also dependent on the [OH ]/[CO3 ] ratio, even at carbonate concentrations of 1-2 M.

Remarks No further remarks.

112

5.25 Tin (Sn) General Tin is a main group element (i.e. group 14) with the atomic number 50. Ten stable isotopes are known with atomic masses ranging between 112 and 124. The most abundant ones are 120Sn (~33%), 118Sn (~24%) and 116Sn (14.5%). The only waste-relevant radioisotope is 126Sn, which belongs to the 7 long-lived fission products (besides 99Tc, 79Se, 93Zr, 135Cs, 107Pd, and 129I. 126Sn has a half-life of 2.3 × 105 years. Compared to 235U and 239Pu, in thermal reactors, 126Sn is only produced at low fission yields. Tin occurs in minerals, such as cassiterite (SnO2) and stannite (Cu2FeSnS4). The latter is also known as bell metal ore. Minerals with tin are almost always associated with granite rock, usually at a level of 1% tin oxide (Wikipedia). Tin shows chemical similarity to both neighbouring group 14 elements, i.e. germanium and lead. Sn occurs in two oxidation states, i.e. Sn2+ and Sn4+. Divalent Sn species are however rare and less stable than the tetravalent Sn species (Brookins, 1988).

a) b) 1 1

+ SnCl3 .5 + .5 Sn(OH)3

Sn(OH) Sn(OH)4(aq) 4

Eh (volts) Eh - 0 ++ Eh (volts) Sn(OH) Sn 0 ++ 5 Sn -- Sn(OH)6 µ µ –.5 –.5

25°C 25°C 02468101214 02468101214 pH pH

c) d) 1 1

.5 Sn++ .5 SnOH+

++ Sn(OH)2 Sn Eh (volts) Eh Eh (volts) Eh 0 0 - Sn(OH)3 µ µ

–.5 –.5

25°C 25°C 02468101214 02468101214 pH pH

113

e) f) 1 1

+ + SnCl SnCl3 + 3 Sn(OH)3 + .5 .5 Sn(OH)3

Cassiterite(tetragonal) Cassiterite(tetragonal)

- (volts) Eh Eh (volts) Eh Sn(OH) ++ - 0 ++ 5 0 Sn Sn(OH)5 Sn -- Sn(OH)-- Sn(OH)6 6 SnS(s) SnS(s) µ µ –.5 –.5 25°C 25°C 0 2 4 6 8 101214 02468101214 pH pH

g) 1

SnCl+ Sn(OH)3 + .5 3

SnO2(am) - no graph -

Eh (volts) Eh - 0 Sn++ Sn(OH)5 -- Sn(OH)6 Ottemannite SnS(s) µ

–.5

25°C 02468101214 pH

Figure 26: Eh-pH diagram of tin (Sn-C-S-O-H) for the BC reference porewater system. Assumed activity of dissolved [Sn] = 10-8. Diagram a) LLNL TDB, b) ANDRA and MOLDATA TDB, c) NAGRA/PSI TDB, d) NEA TDB (not meaningful, as Sn2+ is the only comprised aq. species), e) MOLDATA TDB solid phases included, f) same diagram as e, but [Sn] = 10-6, g) same diagram as f, but cassiterite suppressed in calculation. Code: The Geochemist's Workbench - 8.08 / 8.10.

Speciation & solubility calculations Calculations performed with MOLDATA (and the ANDRA TDB) reveal that the aqueous - speciation of tin is dominated by the Sn(IV), i.e. the pentahydroxide species Sn(OH)5 (Figure 26 b). The species distribution in equilibrium with SnO2(am) was calculated to correspond to 73 % - 2- -8 Sn(OH)5 , ~26 % Sn(OH)4(aq) and 1 % Sn(OH)6 (Table 26). At [Sn] = 10 , Sn solubility under BC conditions would not be controlled by any solid phase (Figure 26 e). At higher Sn concentrations ([Sn] ≥ 10-6), cassiterite becomes stable (also) at pH > 8 and up to pH 10 (Figure 26 f). When suppressing the latter phase in the calculation SnO2(am) appears in the diagram, revealing that this phase represents the kinetically favoured one to form/precipitate (Figure 26 g).

114

Table 25: Solubility of Sn in the BC reference porewater system at 25 °C, pH 8.355 and Eh -281 mV. Database: MOLDATA. Code: The Geochemist's Workbench- 8.08. Solubility controlling phases Solubility, [Sn], mol/l 2 -4 Calcium-stannate Ca[Sn(OH)6](s) 3.1 × 10 2 -7 SnO2 (am) 1.9 × 10 1 -8 Cassiterite (SnO2) 3.4  10 Sn(elemental,cr) very soluble SnO (tetragonal)1 very soluble Source data: 1NEA TDB, 2ANDRA TDB, 3NAPSI, 4LLNL

The reaction constants for the tin solids comprised in Table 25 and MOLDATA are the following: + 2+ 2+ Ca[Sn(OH)6] + 4 H ↔ Ca + Sn + 5 H2O + 0.5 O2(aq) logK = -29.37 + 2+ SnO2(am) + 2H ↔ Sn + H2O logK = -44.68 + 2+ SnO2(cr) + 2 H ↔ Sn + H2O + 0.5 O2(aq) logK = -45.43 + 2+ SnO(tetr) + 2 H ↔ Sn + H2O logK = 2.25

Table 26: Species distribution of Sn in equilibrium with SnO2(am). Database: MOLDATA. Code: The Geochemist's Workbench - 8.10. Aq. species Sn [mol/L] Fraction [%] - -7 Sn(OH)5 1.38 × 10 72.8 -8 Sn(OH)4(aq) 4.98 × 10 26.3 2- -9 Sn(OH)6 1.69 × 10 0.9 Total 1.89 × 10-7 100

Experimental data No in-house or BC specific solubility data available.

Role of (organic) colloids/polynuclear species Sn-NOM colloid association has been put forward in the newly developed phenomenological model (Bruggeman et al., in prep.).

Solubility data provided to PA Source Range (SR)

Based on the speciation and solubility calculations, cassiterite and SnO2(am) have been retained as potential solubility controlling phases delimiting the solubility source range. Calcium stannate has been reported to represent the solubility limiting phase under more alkaline conditions (Lothenbach et al., 2000). Due to this, the latter phase was excluded from the selection. Elemental tin and tin oxide are very soluble under BC conditions, and were therefore judged to be irrelevant solids to limit Sn concentrations in the far-field.

115

Best Estimate (BE)

SnO2(am) was considered to be the most likely solubility limiting solid for Sn under far-field conditions.

Expert range (ER)

SnO2(am) including its thermodynamic uncertainties.

Data retained in the DCF’s: -7 BE: Thermodynamic solubility of SnO2(am): 1.89×10 M. ER: BE including thermodynamic uncertainty. SR: Taken equal to ER.

Eh/pH/pCO2 sensitivity of the solid phases

Amaya et al. (1997) studied the solubility of SnO2(am) and results revealed that the solubility is constant (i.e. pH-independent) in the pH range 2-7, but starts to increase at higher pH values.

Remark The NEA TDB comprises only one aqueous species (i.e. Sn2+), two gases and three solid phases, 2+ i.e. cassiterite (SnO2), SnO (tetragonal) and Sn(cr). Due to the fact that Sn represents only an auxiliary species in the NEA TDB, the speciation diagram (Figure 26 d) is in fact not meaningful.

116

5.26 Uranium (U) General Uranium belongs to the actinide series of the periodic table with an atomic number 92. Natural uranium comprises three isotopes, i.e. 238U (99.2742%), 235U (0.7204%) and 234U (0.0054%). 238U has a half-life of 4.47 billion years and 235U a half-life of 704 million years. U is classified as a lithophile element and its abundance in granitic rocks is about double its average abundance in the Earth’s crust (2.5 ppm). The primary uranium ore mineral is uraninite (UO2), also known as pitchblende. Besides this, U occurs in numerous secondary U-minerals, e.g. carnotite (K2[UO2]2[VO4]2 × 3 H2O), autunite Ca[UO2]2[PO4]2 × 8-12 H2O), coffinite (USiO4), uranophane (Ca[UO2]2[SiO3OH]2 × 5 H2O) and tobernite (Cu[UO2]2[PO4]2 × 12 H2O). Uranium has unique nuclear properties due to which it is playing an important role in nuclear power generation. Whereas the 235U nucleus is 'fissile' that of 238U is said to be 'fertile' or ‘fissionable’, meaning that the fission of 235U may occur through neutron capture/absorption of a thermal neutron (low energy neutron), while fission, 238U requires high energy neutrons. In contrast to 235U, which is able to sustain a chain reaction of nuclear fission, 238U is incapable. The latter can however be transmuted to fissile 239Pu. In order to sustain a nuclear chain reaction, 235U needs however to be enriched to 3-4%, which is generally done through a process of isoptope separation. Uranium has a complex chemistry which is due to its occurrence in multiple valence states (+III, +IV, +V, +VI). The only relevant oxidation states are the tetra- and hexavalent ones.

a) b) 1 1

UO++ ++ 2 UO2 .5 UO2CO3(aq) .5 -- (UO2)(CO3) UO2(CO3)2 -- (UO2)(CO3)2 ---- UO2(CO3)3 ---- U++++ U++++ +++ (UO )(CO ) Eh (volts) Eh +++ (volts) Eh U(OH) 2 3 3 UOH - ++ - 0 UO2(OH)3 0 + (UO )(OH) -- U(OH)2 UO2 2 3-- UO2(OH)4 (UO )(OH) + 2 4 U(OH)3 U(OH)4µ(aq) µ

–.5 –.5 U(OH)4

25°C 25°C 02468101214 02468101214 pH pH

d) c) 1 1

++ ++ UO2 UO2 .5 .5 UO2CO3 UO2CO3(aq) -- -- UO2(CO3)2 UO2(CO3)2 ------++++ UO (CO ) U++++ UO (CO ) U 2 3 3 (volts) Eh

Eh (volts) Eh 2 3 3 +++ + +++ - 0 UOH UO - 0 UOH + UO (OH) 2 UO (OH) UO2 2 3 -- 2 --3 UO2(OH)4 UO2(OH)4 µ µ U(OH)4 U(OH) (aq) –.5 –.5 4

25°C 25°C 02468101214 02468101214 pH pH

117

e) f) 1 1

++ UO++ UO2 2 .5 .5 UO CO (aq) UO CO (aq) 2 3 -- 2 3 UO2(CO3)2 UO (CO )-- 2 3 2 ---- UO2(CO3)3 ---- ++++ +++++++ UO (CO ) U +++++ - U (volts) Eh UO (OH) Eh (volts) Eh UOH 2 3 3 U(OH)UOH 2 3 -- 0 ++ + - 0 2 UO (OH) U(OH)2 UO UO2(OH)3 2 4 2 UO (OH)-- + 2 4 U(OH)3 µ UO2(cr)µ

–.5 U(OH)4(aq) –.5 25°C 25°C 02468101214 02468101214 pH pH g) h) 1 1

UO++ ++ 2 UO2 .5 UO CO (aq) .5 UO CO (aq) 2 3 -- 2 3 -- UO2(CO3)2 UO2(CO3)2

------UO2(CO3)3 UO (CO ) ++++ ++++ 2 3 3 U +++++ - +++ -

Eh (volts) Eh U U(OH)UOH UO2.25(beta) UO2(OH)3 (volts) Eh UOH UO2(OH)3 0 2 UO (OH)-- 0 U(OH)++ -- 2 4 2 UO2(OH)4

Coffinite(cr) µ UO2.25µ(beta)

–.5 –.5

25°C 25°C 02468101214 02468101214 pH pH i) j)

1 1

UO++ ++ 2 UO2 .5 UO2CO3(aq) .5 UO CO (aq) UO (CO )-- 2 3 -- 2 3 2 UO2(CO3)2

UO (CO )---- U++++++ ---- ++++ 2 3 3 USO UO2(CO3)3 U +++ - 4 +++ - Eh (volts) Eh

UOH UO (OH) (volts) Eh U(OH) ++ 2 3 UO2 2 UO2(OH)3 0 U(OH) UO (OH)-- 0 -- 2 2 4 UO2(OH)4 UO2S2O3(aq) + U(OH)3 UO µ (beta) µ 2.3333 UO2.34(b) –.5 –.5

25°C 25°C 0 2 4 6 8 10 12 14 0 2 4 6 8 101214 pH pH

118

k) 1

++ UO2 .5 UO2CO3(aq) -- UO2(CO3)2

++++ ---- U +++ UO2(CO3)3 - no graph - Eh (volts) Eh UOH ++ + - 0 U(OH) UO UO2(OH)3 2 2 UO (OH)-- + 2 4 U(OH)3 µ UO (am,hyd) –.5 2

25°C 02468101214 pH Figure 27: Eh-pH diagram of uranium (U-C-S-O-H) for the BC reference porewater system. Assumed activity of dissolved [U] = 10-8. Diagram a) LLNL TDB, b) ANDRA TDB, c) NAGRA/PSI TDB, d) NEA TDB, e) MOLDATA TDB, f) same as diagram e, but solid phases included h) coffinite(cr) suppressed, i)UO2.25 suppressed k) UO2.34 suppressed, UO2.6667 suppressed also, as irrelevant). Code: The Geochemist's Workbench - 8.08/ 8.10.

Speciation & solubility calculations The dominant aqueous species under BC conditions calculated with MOLDATA is predicted to be 4- a uranyl species, i.e. UO2(CO3)3 (Figure 27 e). The latter represents also the prevalent species using the other TDB’s to calculate the U speciation, except when using the LLNL TDB. Then, U(OH)4(aq) is calculated to be the major aqueous complex. Although with the other TDB’s the uranyl tricarbonate species is the dominant one, it can be seen, that the reference BC porewater composition plots also in the other diagrams close to the stability field of the U(OH)4(aq) species. As revealed by (Table 28), illustrating the U species distribution in equilibrium with U(OH)4(am,hyd), the respective aqueous species does not exceed 0.3% of the total uranium concentration. The predominance of the U(VI)-carbonate complexes can be referred to their high stability.

All results of the solubility calculations are summarized in Table 27. Comprised within the solubility range delimited by crystalline and amorphous UO2 are the solubilities of uranium oxides characterized by different oxidation degrees with a stoichiometry UO2+x, as well as the one of coffinite. Taking into account coffinite as solubility limiting phase for aqueous U has been controversely discussed (ThermoChimie documentation; Duro et al., 2006; Bruno et al., 2001 a). Although the presence of coffinite has been reported from the Natural Analogues of Oklo/Bangombé and Cigar Lake (Cramer and Smellie, 1994; Fayek et al, 1997; Janeczek and Ewing, 1992; Jensen et al., 2000), its synthesis under laboratory conditions has failed up to now (Robit-Pointeau, 2006).

Table 27: Solubility of U in the BC reference porewater system at 25 °C, pH 8.355 and Eh -281 mV. Database: MOLDATA. Code: The Geochemist's Workbench- 8.08. Solubility controlling phases Solubility, [U], mol/l 2 -6 Coffinite (USiO4,am) 5.6 × 10 1 -6 UO2 (am,hyd) 1.0  10 1 -10 UO2.3333 (beta) 5.5  10

119

1 -11 UO2.25 (beta) 5.2  10 2 -12 Coffinite (USiO4,cr) 1.5  10 1 -13 Uraninite (UO2,cr) 4.5  10 Source data: 1NEA TDB, 2ANDRA TDB, 3NAPSI, 4LLNL

The reaction constants for the uranium solids comprised in (Table 27) and MOLDATA are the following: + 2+ USiO4(am) + 2 H + H2O + 0.5 O2(aq) ↔ UO2 + Si(OH)4(aq) log K = 32.45 + 2+ UO2(am,hyd) + 2 H + 0.5 O2(aq) ↔ UO2 + H2O log K = 35.45 + 2+ UO2.3333(beta) + 2 H + 0.3334 O2(aq) ↔ UO2 + H2O log K = 20.26 + 2+ UO2.25(beta) + 2 H + 0.375 O2(aq) ↔ UO2 + H2O log K = 22.22 + 2+ USiO4(cr) + 2 H + H2O + 0.5 O2(aq) ↔ UO2 + Si(OH)4(aq) log K = 25.90 + 2+ Uraninite (UO2,cr) + 2 H + 0.5 O2(aq) ↔ UO2 + H2O log K = 29.10

Table 28: Species distribution of U in equilibrium with UO2(am,hyd). Database: MOLDATA. Code: The Geochemist's Workbench- 8.08. Aqueous species U [mol/l] Percentage [%] 4- -9 UO2(CO3)3 9.87 × 10 98.7 2- -9 UO2(CO3)2 9.75 × 10 1.0 -9 U(OH)4(aq) 3.10 × 10 0.3 Total 1.00 × 10-6 100

Experimental data

Delécaut (2004) performed different solubility experiments on UO2(am) in SBCW, as well as in RBCW. Uranium concentrations measured in SBCW after ultrafiltration are reported to correspond to 2.9 × 10-8 M, while in suspensions filtered at 0.45µm the U-concentrations were around two orders of magnitude higher, i.e. 1.6 × 10-6 M. This higher concentration was interpreted to be due to the presence of inorganic U-colloids with a size between 0.45 µm and 2 nm. The measured U-concentration after 0.45µm filtration is very close to the calculated UO2(am,hyd) solubility (see Table 27). The experiments revealed further, that UO2(am) solubility was not influenced by the presence of HA.

Role of (organic) colloids/polynuclear species As mentioned above, Delecaut (2004) evidenced the presence of "real" colloids with a size between 2nm and 0.45µm, which enhanced the solubility of UO2(am) by 3 orders of magnitude. U-NOM colloid association has been put forward in the newly developed phenomenological model (Bruggeman et al., in prep.). A solubility increasing effect for UO2(am) by these colloids (> 2nm) and U(IV)-HA (< 2nm) complexes was not observed (see also sections 4.3.6 and 4.3.8).

Solubility data provided to PA Source Range (SR) All phases comprised in Table 27 have been retained to delimit the solubility source range for uranium. The upper limit of the source range has been enlarged due to the experimental

120

observations done by Delécaut (2004). UO2(am) is considered to represent – based on the Ostwald principle – the kinetically most probable phase to controll the U solubility in the far-field.

Best Estimate

UO2(am,hyd) was judged to represent the most likely solubility limiting solid for U in the far- field.

Expert Range (ER)

UO2(am, hyd) including the thermodynamic uncertainty.

Data retained in the DCF’s: -6 BE: Thermodynamic solubility of UO2(am, hyd): 1.0  10 M. ER: BE including thermodynamic uncertainty. -13 SR: Lower limit (LL): Thermodynamic solubility of (UO2,cr): 4.5  10 M.

Upper limit (UL): Taken equal to ER. Range includes experimental data of UO2(am) solubility measured by Delecaut (2004) in SBCW after filtration of 0.45µm (1.6 × 10-6 M).

Eh/pH/pCO2 sensitivity of the solid phases As uranium represents a redox sensitive element, the stability of all solid U-phases is Eh and also often pH/pCO2 dependent.

Remarks In the NAGRA/PSI TDB, the incorporated U data were mainly taken from Grenthe et al. (1992). As the latter review does not comprise any mixed valence oxides (U3O8, U3O7 and U4O9), they are also not included in the NAGRA TDB.

121

5.27 Zirconium (Zr) General Zirconium belongs to the group of transition metals and its atomic number is 40. Zirconium has five natural isotopes, i.e. 90Zr, 91Zr, 92Zr, 94Zr and 96Zr of which three are stable (91Zr, 92, Zr, 94Zr). 96Zr is the longest-lived isotope with 2.4 × 1019 years. 90Zr is with an occurrence of 51.45%. With respect to radioactive isotopes, 93Zr represents the longest-lived isotope with a half life of 1.53 million years and decaying to 93Nb. 93Zr is also one of the 7 long-lived fission products. Due to the corrosion resistance and the very low absorption cross-section for thermal neutrons, large amounts of zirconium are contained in nuclear reactors usually as fuel rod cladding (i.e. zircalloy) and as coating on nuclear fuel parts. Zr is abundant in the sun, stars and meteorites. Analysis of lunar rocks has shown that zirconium is a common element on the surface of the moon. Most common natural zirconium minerals are zircon (ZrSiO4) and baddeleyite (ZrO2). Upon alteration Zr is most mikely incorporated in Zr(OH)4, representing the preferred Zr(IV) solid (Brookins, 1988).

a) b) ZrOH3+ 1 1 2+ Zr(OH)2 + Zr(OH)3 Zr(OH)+++ .5 .5

Zr(OH)4(aq)

Zr(OH) (aq) (volts) Eh

Eh (volts) 4 0 0 Zr(OH)- Zr(OH)6 25 µ µ –.5 –.5 25°C 25°C 02468101214 0 2 4 6 8 10 12 14 pH pH

c) 1 d) 1

++ +++ Zr(OH) .5 ZrOH .5 2

Zr(OH)4

Zr(OH)4(aq) Eh (volts) Eh - (volts) Eh 0 Zr(OH)5 0 -- Zr(OH)6 µ µ

–.5 –.5

25°C 25°C 02468101214 02468101214 pH pH

122

e) f) 1 1

++ Zr(OH)++ Zr(OH) .5 2 .5 2

Zr(OH)4(aq) ZrO2(monoclinic) Eh (volts) Eh Eh (volts) Eh 0 0 -- Zr(OH)6 µ µ

–.5 –.5

25°C 25°C 0 2 4 6 8 101214 02468101214 pH pH

g) h) 1 1

+++ Zr3(OH)9 .5 .5 + Zr4(OH)15

Zr(OH)4(am,aged) Eh (volts) Eh Eh (volts) Eh Zr(OH)4(am,fresh) 0 0

µ µ

–.5 –.5

25°C 25°C 02468101214 02468101214 pH pH Figure 28: Eh-pH diagram of zirconium (Zr-C-S-O-H) for the BC reference porewater system. Assumed activity of dissolved [Zr] = 10-8. Diagram a) LLNL TDB, b) ANDRA TDB c) NAGRA/PSI TDB, d) NEA TDB, e) MOLDATA f) same as diagram e, but solid phases included; (Zr(SO3)2(s), ZrS3(cr), ZrS2(cr), -4 ZrS1.5(cr) suppressed in calculation, g) same diagram as f, but [Zr] = 1.5 × 10 and ZrO2 (monoclinic) suppressed, h) same as diagram g, but also Zr(OH)4(am,aged) suppressed. Code: The Geochemist's Workbench - 8.08 / 8.10.

Table 29: Solubility of Zr in the BC reference porewater system at 25 °C, pH 8.355 and Eh -281 mV. Database: MOLDATA. Code: The Geochemist's Workbench- 8.08. Solubility controlling phases Solubility, [Zr], mol/l 1 -4 Zr(OH)4 (am,fresh) 1.11 × 10 1 -8 Zr(OH)4 (am,aged) 1.82 × 10 1 -10 Baddeleyite (ZrO2, mono) 6.43  10 1 -14 Zircon (ZrSiO4) 8.56  10 2 CaZrO3 (s) very soluble Source data: 1NEA TDB, 2ANDRA TDB, 3NAPSI, 4LLNL

The reaction constants for the uranium solids comprised in (Table 29) and MOLDATA are the following:

123

+ 4+ Zr(OH)4(am,fresh) + 4 H ↔ Zr + 4 H2O log K = -3.24 + 4+ Zr(OH)4(am,aged) + 4 H ↔ Zr + 4 H2O log K = -5.55 + ZrO2(mono) + 4 H ↔ Zr(OH)4(aq) + 4 H2O log K = -7.00 + 4+ ZrSiO4(cr) + 4 H ↔ Zr + Si(OH)4(aq) log K = -14.62 + 2+ 4+ CaZrO3(s) + 6 H ↔ Ca + Zr + 3 H2O logK = 19.58

Speciation & solubility calculations In the environment, Zr always occurs in the oxidation state +IV. Due to its high charge, small size (ionic radius = 0.72 Å in octaheadral coordination) and hard metal character, Zr4+ hydrolyses already at pH between 0 and 1 (Hummel et al., 2002) resulting in the fact that the free Zr4+ aqua ion does not exist under natural conditions. Under BC conditions, Zr(OH)4(aq) represents the dominant aqueous species, as revealed by calculations using the LLNL, ANDRA, NEA and MOLDATA TDB, respectively (Figure 28 a, b, d and e). In contrast, using the NAGRA/PSI TDB, - Zr(OH)5 is predicted to be the prevalent aqueous species (Figure 28 c) in the near-neutral and alkaline pH range.

Solubility calculations have been performed for all solid phases comprised in Table 29.

Experimental data No in-house or BC specific data available

Role of (organic) colloids/polynuclear species Zirconium has a strong tendency to form polymeric species and colloids (Baes and Mesmer, 1976). Zr-NOM colloid association has been put forward in the newly developed phenomenological model (Bruggeman et al., in prep.).

Solubility data provided to PA Source Range (SR) Zircon has been excluded from the SR as potential solubility controlling phase under ambient T- conditions. Due to the high solubility of CaZrO3(s), also this phase was rejected. All other phases comprised in Table 29 have been retained to delimit the solubility source range for zirconium.

Best Estimate

Zr(OH)4(am, aged) was judged to represent the most likely solubility limiting solid for Zr in the far-field.

Expert Range (ER)

Zr(OH)4(am, aged) including thermodynamic uncertainty.

Data retained in the DCF’s: -8 BE: Thermodynamic solubility of Zr(OH)4(am,aged): 1.82×10 M. ER: BE including thermodynamic uncertainty. SR: Lower limit (LL): Taken equal to ER.

124

-4 Upper limit (UL): Thermodynamic solubility of Zr(OH)4(am,fresh): 1.11 × 10 M.

Eh/pH/pCO2 sensitivity of the solid phases Zirconium is not redox-sensitive. Thus, Eh is not expected to have a major effect on the solid phase stability. The influence of pH on the Zr solubility seems to be more important, but is controversely discussed (Berner, 2002).

Remarks No remarks.

125

6 ANNEX I: Summary of solubility calculations for phases comprised in the SR & ER ranges

Solubility [M] BE and Element Solid controlling phases Remarks (MOLDATA) SR(UL/LL)

-3 Ac(OH)3(am) 1.247 × 10 UL Taken identical to Am -6 Actinium Ac AcCO3OH (am,hyd) 3.818 × 10 BE Taken identical to Am -8 AcCO3OH x 0.5 H2O(cr) 2.413 × 10 LL Taken identical to Am -3 Am(OH)3(am) 1.247 × 10 UL -5 Am(OH)3(cr) 6.197 × 10 -6 Am2(CO3)3(am) 2.515 × 10 -6 Americium Am Am(CO3)1.5(cr) 2.348 × 10 -6 AmCO3OH(am,hyd) 3.818 × 10 BE -7 NaAm(CO3)2 x 5 H2O(cr) 7.645 × 10 -8 AmCO3OH x 0.5 H2O(cr) 2.413 × 10 LL UL: BE ± 2; Beryllium Be Bromellite (BeO) 8.468 × 10-15 BE LL BE: ± 2 UL=ER: 8.1 × 10-5 M Calcium Ca CaCO (s) 5.133 × 10-5 BE 3 calcite LL=ER: 3.2 × 10-5 M Cesium Cs not limited not limited not limited Chlorine Cl not limited not limited not limited -3 Cm(OH)3(am) 1.247 × 10 UL Taken identical to Am -6 Curium Cm CmCO3OH(am,hyd) 3.818 × 10 BE Taken identical to Am -8 CmCO3OH x 0.5 H2O(cr) 2.413 × 10 LL Taken identical to Am Iodine I not limited not limited not limited -6 MoO2(s)tugarinovite 1.661 × 10 UL/BE Molybdenum Mo -13 MoS2(s) 1.966 × 10 LL -8 -9 UL=ER: 3.3 × 10 NpO2(am,hyd) 2.021 × 10 BE Neptunium Np (thermodyn. uncertainty) -18 NpO2(cr) 1.790 × 10 LL -4 Ni(OH)2(beta) 7.179 × 10 UL/BE -5 NiFe2O4(cr)trevorite 1.833 × 10 Nickel Ni NiCO (cr) 2.608 × 10-5 BE 3 gaspéite -8 β-NiSmillerite 1.763 × 10 -9 FeNi2S4violarite 9.522 × 10 LL UL=ER: 1 × 10-8 M -6 -3 Niobium Nb Nb2O5(cr) 2.363 × 10 BE LL=ER: 1.0 × 10 M (former DCF) UL=ER: 4 × 10-5 M -6 Pd(OH)2(s) 3.968 x 10 BE (thermodyn. uncertainty) PdO(aq) suppressed in calc. Palladium Pd PdO(s) 1.530 × 10-10 Pd(cr) 2.232 × 10-30 PdS(cr) 7.936 × 10-35 LL -4 Pu(OH)3(cr) 1.412 × 10 UL -5 PuCO3OH(s) 1.567 × 10 Plutonium Pu -8 PuO2(am,hyd) 1.327 × 10 BE -14 PuO2(cr) 2.648 × 10 LL

126

UL:1.0 × 10-5 M (former DCF); Protactinium Pa Pa O 9.785 × 10-10 BE 2 5 LL: 1.0 × 10-11 M (migration exp.) -4 -5 UL=ER: 1.2 × 10 M RaCO3(cr) 6.688 × 10 BE Radium Ra (thermodyn. uncertainty) -6 RaSO4(s) 7.033 × 10 LL Rubidium Rb not limited not limited not limited -4 Sm(OH)3(s) 3.655 × 10 UL -7 Samarium Sm Sm2(CO3)3(s) 4.971 × 10 LL=ER: 2.0 × 10-8 M SmOHCO (s) 8.948 × 10-8 BE 3 (thermodyn. ncertainty) Se (mono) 2.360 × 10-6 UL Se (trig) 6.477 × 10-7 -7 Selenium Se Fe3Se4 (gamma) 1.413 × 10 -9 FeSe(s)achavalite 1.135 × 10 LL=ER: 6.6 × 10-12 M FeSe (cr) 3.564 × 10-9 BE 2, ferroselite (thermodyn. uncertainty) -4 Ag2CO3(s) 5.574 × 10 UL AgCl(cr) 2.477 × 10-5 BE Silver Ag -13 Ag2S(cr)acanthite 5.443 × 10 Ag(cr) 3.735 × 10-13 LL UL=ER: 1.3 × 10-5 M -6 -6 SrCO3(cr)strontianite 7.917 × 10 BE LL=ER: 4.7 × 10 M Strontium Sr (thermodun. uncertainties)

SrSO4(cr)celestite very soluble -9 -7 TcO2 x 1.6 H2O(s) 4.398 × 10 BE UL: 4.6 × 10 M (exp. data) Technetium Tc Tc(cr) 1.982 × 10-11 -13 TcO2(cr) 4.460 × 10 LL -5 ThO2(am,hyd,fresh) 1.610 × 10 UL=ER: 6.1 × 10-5 M Thorium Th ThO (am,hyd,aged) 2.550 × 10-6 BE 2 (thermodyn. uncertainty) -13 ThO2(cr) 4.70 × 10 LL UL=ER: 3.3 × 10-6 M -7 -6 SnO2(am) 1.891 × 10 BE LL=ER: 1.1 × 10 M Tin Sn (thermodun. uncertainties) -8 SnO2(tetr)cassiterite 3.371 × 10 -6 USiO4(am)coffinite 5.612 × 10 UL=ER: 2.5 × 10-6 M UO (am,hyd) 9.999 × 10-7 BE 2 (thermodyn. uncertainty) -10 Uranium U UO2.3333 (beta) 5.496 × 10 -11 UO2.25 (beta) 5.220 × 10 -12 USiO4 (cr)coffinite 1.543 × 10 -13 UO2(cr)uraninite 4.461 × 10 LL -4 Zr(OH)4(am,fresh) 1.109 × 10 UL -8 -10 Zirconium Zr Zr(OH)4(am,aged) 1.816 × 10 BE LL=ER: 3.5 × 10 M -10 ZrO2(monoclinic) 6.433 × 10

127

7 ANNEX II: Summary of Source and Expert Ranges

Solubility limiting Element BE ER* Remarks SR* Remarks phase LL UL LL UL Ac 3.8E-06 taken identical to Am 2.6E-07 5.7E-05 taken identical to Am 2.4E-08 1.2E-03 taken identical to Am

SR(LL): thermodyn. sol. of AmCO3OH×0.5 ER(LL/UL): BE ± thermodyn. H O (in line with sol. measured in Am 3.8E-06 AmCO OH(am,hyd) 2.6E-07 5.7E-05 2.4E-08 1.2E-03 2 3 uncertainty TRANCOM II); SR(UL): thermodyn. sol. of Am(OH)3(am) ER(LL/UL): BE ± 1 order of magnitude Be 8.5E-15 BeO (Bromellite) 8.5E-16 8.5E-14 8.5E-17 8.5E-13 SR (LL/UL): BE ± 2 orders of magnitude (arbitrarily chosen) ER(LL/UL): BE ± thermodyn. Ca 5.1E-05 CaCO (Calcite) 3.2E-05 8.1E-05 3.2E-05 8.1E-05 SR(LL/UL): taken equal to ER 3 uncertainty Cs not limited not limited Cl not limited not limited Cm 3.82E-06 taken identical to Am 2.6E-07 5.7E-05 taken identical to Am 2.4E-08 1.2E-03 taken identical to Am I not limited not limited ER(LL/UL): BE ± thermodyn. SR(LL): thermodyn.sol. of MoS (s); Mo 1.7E-06 MoO (s)/Tugarinovite 5.2E-06 5.2E-06 2.0E-13 1.7E-06 2 2 uncertainty SR(UL): equal to BE ER(LL/UL): BE ± thermodyn. SR(LL): thermodyn.sol. of NpO (cr); Np 2.0E-09 NpO (am,hyd) 1.2E-10 3.3E-08 1.8E-18 3.3E-08 2 2 uncertainty SR(UL): equal to ER(UL) SR(LL): thermodyn. sol. of FeNi2S4 ER(LL/UL): BE ± thermodyn. Ni 2.6E-05 NiCO (cr)/Gaspéite 1.5E-05 4.4E-05 9.5E-09 7.2E-04 (violarite); 3 uncertainty SR(UL): thermodyn. sol. of Ni(OH)2(beta) ER (LL/UL): estimation not possible,

Nb 2.4E-06 Nb O (s) 1.0E-08 1.0E-03 therefore the values of the previous DCF 1.0E-08 1.0E-03 2 5 SR(LL/UL): taken equal to ER(LL/UL) (Safir-2) were taken ER (LL/UL): estimation not possible, Pd 4.0E-06 Pd(OH) (s) 4.0E-07 4.0E-05 7.9E-35 4.0E-05 2 therefore BE ± 1 order of magnitude; SR (UL): equal ER (UL) ER (LL/UL): estimation not possible, SR(LL): thermodyn. sol. of PuO (cr); Pu 1.3E-08 PuO (am,hyd) 1.3E-10 1.3E-06 therefore BE ± 2 orders of magnitude 2.6E-14 1.4E-04 2 2 SR(UL): thermodyn. sol. of Pu(OH) (cr) (arbitrarily chosen); 3

128

SR(LL) taken equal to constant concentration 9.80E- ER (LL/UL): estimation not possible, from migration experiment; Pa 9.8E-10 Pa O (cr) 9.8E-09 1.0E-11 1.0E-05 2 5 11 therefore BE ± 1 order of magnitude; SR(UL): former DCF reports solubilities up to 10-5 (Finnish data).

ER(LL/UL): BE ± thermodyn. SR(LL): thermodyn. solubility of RaSO ; Ra 6.7E-05 RaCO (s) 3.8E-05 1.2E-04 7.0E-06 1.2E-04 4 3 uncertainty SR(UL): taken equal to ER(UL) Rb not limited not limited ER(LL/UL): BE ± thermodyn. SR(LL): taken equal to ER(LL) Sm 8.9E-08 SmOHCO3(s) 2.0E-08 4.1E-07 2.0E-08 3.7E-04 uncertainty SR(UL): thermodyn.solubility of Sm(OH)3(s) ER(LL/UL): BE ± thermodyn. SR(LL): taken equal to ER(LL); Se 3.6E-09 FeSe (cr)/Ferroselite 6.6E-12 1.9E-06 6.6E-12 2.4E-06 2 uncertainty SR(UL): thermodyn. solubility of Se(mono) ER(LL/UL): BE ± thermodyn. SR(LL): thermodyn. sol. of metallic Ag(cr); Ag 2.5E-05 AgCl(cr)/Chlorargyrite 2.1E-05 3.0E-05 3.7E-13 5.6E-04 uncertainty SR(UL): thermodyn. sol. of Ag2CO3(s) ER(LL/UL): BE ± thermodyn. Sr 7.9E-06 SrCO (s)/Strontianite 4.7E-06 1.3E-05 4.7E-06 1.3E-05 SR(LL/UL): taken equal to ER(LL/UL) 3 uncertainty ER(LL/UL): BE ± thermodyn. SR(LL): thermodyn. sol. of TcO (cr); Tc 4.4E-09 TcO × 1.6 H O(s) 1.4E-09 1.4E-08 4.5E-13 4.6E-07 2 2 2 uncertainty SR(UL): derived from exp. observations ER(LL/UL): BE ± thermodyn. uncertainty, but ER(LL) was expanded to SR(LL): thermodyn. sol. of ThO (cr); Th 2.6E-06 ThO (am,hyd,aged) 1.0E-11 6.1E-05 include experimental data of ThO (s) 4.7E-13 6.1E-05 2 2 2 SR(UL): taken equal to ER(UL) solubility in SBCW after 30 and 300 kDa UF. ER(LL/UL): BE ± thermodyn. Sn 1.9E-07 SnO (am) 1.1E-08 3.3E-06 1.1E-08 3.3E-06 SR(LL/UL): taken equal to ER(LL/UL) 2 uncertainty ER(LL/UL): BE ± thermodyn. SR(LL): thermodyn. sol. of UO (cr); U 1.0E-06 UO (am,hyd) 4.1E-07 2.5E-06 4.5E-13 2.5E-06 2 2 uncertainty SR(UL): taken equal to ER(UL) SR(LL): taken equal to ER(LL); ER(LL/UL): BE ± thermodyn. Zr 1.8E-08 Zr(OH) (am,hyd) 3.5E-10 9.3E-07 3.5E-10 1.1E-04 SR(UL): thermodyn. sol. of 4 uncertainty Zr(OH)4(am,fresh)

*Effect of NOM not taken into account

129

8 ANNEX III: Calculation of thermodynamic uncertainties

The uncertainties associated to the solubilities of all phases arise from the uncertainties on the respective solubility products and the formation constant of the dominant aqueous species under BC conditions. Propagating the formal uncertainties of each relevant equilibrium produces the “total” uncertainty for a specific phase. This simple error propagation method is in accordance with the NEA guidelines (Silva et al., 1995):

…

In general - within a first step - the uncertainties associated to the solubility limiting phases were calculated from the standard molar Gibbs energies of formation fGm° of the reactants and products involved in the equilibria under consideration.

If the reaction constants of the equilibria under consideration are given in the source (e.g. NEA reviews or ThermoChimie v.5 documentation) and using the same dominant aqueous species as the one predicted under BC conditions, these reaction constants and associated uncertainties were also considered in the evaluations.

Americium The uncertainties associated to the solubilities of the americium phases arise from the uncertainties on the - respective solubility products and the formation constant of Am(CO3)2 , the latter representing the dominant aqueous complex under BC conditions. Propagating the formal uncertainties of the relevant equilibria produces the uncertainties for the potential solubility limiting phases, calculated as follows:

AmCO3OH(am,hyd) NEA

- The equilibria and associated uncertainties for AmCO3OH(am,hyd) and Am(CO3)2 given by Guillaumont et al. (2003) are (Vol. 5., p. 122):

3+ 2- - ° AmCO3OH(am,hyd) ↔ Am + CO3 + OH log10 K = -20.20 ± 1.0 3+ 2- - ° Am + 2 CO3 ↔ Am(CO3)2 log10 K = 12.90 ± 0.6

1.0 0.6  = ± 1.17 The resulting equation is: 2- - - ° AmCO3OH(am,hyd) + CO3 ↔ Am(CO3)2 + OH log10 K = -7.30 ± 1.17

Calcium The uncertainty for calcite was calculated only from the uncertainty on the standard molar Gibbs energies 2+ - of formation of the aqueous species (i.e. Ca , HCO3 ), as no uncertainty for calcite was given within the NAPSI documentation.

130

CaCO3(s) NAPSI

2+ - ° CaCO3(s) + H+ = Ca + HCO3 log10 K = 1.84

° ∆ CaCO3(s) : no data given by NAPSI ° 2+ -1 ∆ of Ca : ± 1.05 kJ mol ° - -1 ∆ of HCO3 : ± 0.251 kJ mol

1.05 0.251  = ± 1.08 kJ mol-1 / 2.303 RT (5.708) = ± 0.2

The resulting equation is: 2+ - ° CaCO3(s) + H+ = Ca + HCO3 log10 K = 1.84 ± 0.2

Molybdenum

MoO2(s)tugarinovite ANDRA

Reaction given by ANDRA: 2- + - ° MoO2(s)tugarinovite + H2O ↔ MoO4 + 4 H + 2e log10 K = 29.88 ± 0.5

Neptunium

The total uncertainty of NpO2(am,hyd) arises from the uncertainty on the solubility product and the formation constant of Np(OH)4(aq), the latter representing the dominant aqueous complex under BC conditions. Propagating the formal uncertainties of the relevant equilibria gives:

NpO2(am,hyd) NEA

Reactions given by NEA (Vol. 5, p. 88 & 296-298): - 4+ NpO2 (am,hyd) + 2 H2O ↔ 4 OH + Np log10 K° = -56.7 ± 0.5 4+ - Np + 4 OH = Np(OH)4(aq) log10 K° = 47.7 ± 1.1

0.5 1.1  = ± 1.21 The resulting equation is:

NpO2 (am,hyd) + 2 H2O ↔ Np(OH)4(aq) log10 K° = -9.01 ± 1.21

Nickel The uncertainties for the Ni-phases would normally be calculated from the uncertainty on the solubility 2- product and the formation constant of Ni(CO3)2 , which represents the dominant aqueous complex under

131

2 BC conditions. No data are however given for Ni(CO3)2 -, which makes an uncertainty evaluation difficult. Propagating the formal (available) uncertainties of the relevant equilibria gives:

-Ni(OH)2 NEA

+ 2+ β-Ni(OH)2 + 2 H ↔ Ni + 2 H2O log10 K° = 11.03 2+ - 2- Ni + 2 HCO3 ↔ Ni(CO3)2 + 2 H+ log10 K° = -10.59 - 2- β-Ni(OH)2 + 2 HCO3 ↔ Ni(CO3)2 + 2 H2O log10 K° = -0.44

° -1 ∆ of β-Ni(OH)2 : ± 1.4 kJ mol ° 2- ∆ of Ni(CO3)2 : no data given by ANDRA ° -1 ∆ of H2O : ± 0.041 kJ mol ° - -1 ∆ of HCO3 : ± 0.251 kJ mol

1.4 2 0.041 2 0.251  = ± 1.49 kJ mol-1 / 2.303 RT (5.708) = ± 0.26

Reaction (constant) given by NEA (Vol. 6, p. 115 Tab. V-7). + 2+ β-Ni(OH)2 + 2 H ↔ Ni + 2 H2O log10 K° = 11.03 ± 0.2

Niobium

Niobium data comprised in MOLDATA were copied from the ANDRA TDB ThermoChimie v.5. No - uncertainties are given for niobium pentoxide neither for the basis species Nb(OH)6 . Therefore, it is not possible to calculate an uncertainty based on the standard molar Gibbs energy of formation data. To estimate an uncertainty by analogy with another similar phase (e.g. Pa2O5) is also not possible as data for this phase are also lacking.

Nb2O5(s) ANDRA

- + Nb2O5(s) + 7 H2O ↔ 2 Nb(OH)6 + 2H log10 K° = -28.38 ± ?? No reaction (constant) is given by ANDRA.

Palladium

Pd(OH)2(s) ANDRA

+ 2+ Pd(OH)2(s) + 2 H ↔ Pd + 2 H2O log10 K° = -1.61 2+ + Pd +2 H2O ↔ Pd(OH)2(aq) + 2 H log10 K° = -3.79

Pd(OH)2(s) ↔ Pd(OH)2(aq) log10 K° = -5.40

132

° ∆ of Pd(OH)2(s) : no data given by ANDRA ° ∆ of Pd(OH)2(aq) : no data given by ANDRA ??

No reaction (constant) is given by ANDRA.

Plutonium Pu(III) carbonate complexes are not recommended by NEA, “due to the facility with which Pu(III) oxidizes to Pu(IV) in basic solutions” and due to the lack of experimental studies to determined their identity and stability. While in the Np/Pu review Vol. 4 (Lemire et al., 2001), it is however referred to a work of + Cantrell (1988), who estimated the stability constants at I = 0 and 25°C for PuCO3 (log β = 7.5) and - Pu(CO3)2 (log β = 12.4), these species are not comprised anymore in the discussions of the recent review (Vol. 5, Guillaumont et al., 2003). In ThermoChimie v. 5 (and MOLDATA), the standard molar Gibbs energies of formation for the above 3- mentioned carbonate complexes and additionally Pu(CO3)3 are given and were estimated by analogy to Am carbonates. Due to this correlation, the uncertainties of these data are extremely high (~6-60 kJ/mol). Despite this fact, uncertainty calculations were however performed, but the results are considered to be unrealistically high. Therefore, if available the reaction constants and associated uncertainties were copied from the respective sources of the solids.

PuCO3OH(s) ANDRA

+ 4+ - PuCO3OH(s) + 0.25 O2(aq) +3 H ↔ Pu + HCO3 + 1.5 H2O log10 K° = 8.39 4+ - 3- + Pu + 0.5 H2O + 3 HCO3 ↔ Pu(CO3)3 + 0.25 O2(aq) + 4 H log10 K° = -18.38 - 3- + PuCO3OH(s) + 2 HCO3 ↔ Pu(CO3)3 + H + H2O log10 K° = -9.99

° ∆ of PuCO3OH(s) : no data given by ANDRA ° 3- -1 ∆ of Pu(CO3)3 : ± 60.576 kJ mol ° - -1 ∆ of HCO3 : ± 0.251 kJ mol ° -1 ∆ of H2O : ± 0.041 kJ mol

60.576 2 0.251 0.041  = ± 60.58 kJ mol-1 / 2.303 RT (5.708) = ± 10.61 (!!!)

No reaction (constant) is given by ANDRA.

Protactinium

The uncertainties for Pa2O5(s) is calculated from the uncertainty on the solubility product and the formation constant of Pa(OH)5(aq), which represents the dominant aqueous complex under BC conditions. Propagating the formal uncertainties of the relevant equilibria gives:

Pa2O5(s) ANDRA

133

+ 4+ Pa2O5(s) + 8 H = 2 Pa + 4 H2O + 0.5 O2(aq) log10 K° = -55.43 4+ + Pa + 4.5 H2O + 0.25 O2(aq) ↔ Pa(OH)5(aq) + 4 H log10 K° = 18.69

Pa2O5(s) + 5 H2O ↔ 2 Pa(OH)5(aq) log10 K° = -18.06

° -1 ∆ of Pa2O5(s) : ± 35.531 kJ mol (!!!) ° ∆ of Pa(OH)5(aq) : no data available ° -1 ∆ of H2O : ± 0.041 kJ mol

35.531 5 0.041  = ± 35.53 kJ mol-1 / 2.303 RT (5.708) = ± 6.22 (!!!)

Radium Radium data (minerals and Ra2+) were copied from ThermoChimie v.5 to MOLDATA, therefore uncertainties - if given - were also taken from this TDB.

RaCO3(s) ANDRA

+ 2+ - RaCO3(s) + H ↔ Ra + HCO3 log K = 2.03

° ∆ of RaCO3(s) : no data given by ANDRA ° 2+ -1 ∆ of Ra : ± 1.382 kJ mol ° - -1 ∆ of HCO3 : ± 0.251 kJ mol

1.382 0.251  = ± 1.40 kJ mol-1 / 2.303 RT (5.708) = ± 0.25

The resulting equation is: + 2+ - RaCO3(s) + H ↔ Ra + HCO3 log K = 2.03 ± 0.25

Reaction given by ANDRA: 2+ 2- RaCO3(s) ↔ Ra + CO3 log K = -8.3 (no uncertainty is given)

Strontium

SrCO3(s)strontianite NAPSI

+ 2+ - SrCO3(s)strontianite + H ↔ Sr + HCO3 log10 K° = 1.05

° -1 ∆ of SrCO3(s) : ± 1.00 kJ mol (Busenberg et al., 1984)

134

° 2+ -1 ∆ of Sr : ± 0.781 kJ mol ° - -1 ∆ of HCO3 : ± 0.251 kJ mol

1.00 0.781 0.251  = ± 1.29 kJ mol-1 / 2.303 RT (5.708) = ± 0.23

The resulting equation is: + 2+ - SrCO3(s)strontianite + H ↔ Sr + HCO3 log10 K° = 1.05 ± 0.23

Original reaction (Busenberg et al., 1984) 2+ 2- SrCO3(cr) ↔ Sr + CO3 log10 K° = -9.271 ± 0.02

Selenium

FeSe2(cr)ferroselite NEA (Vol. 7, p. 48, 328, 641)

2+ + 2- FeSe2(cr)ferroselite + 2.5 O2(aq) + H2O ↔ Fe + 2 H + 2 SeO3 log10 K° = 90.78 2- + - SeO3 + H ↔ HSe + 1.5 O2(aq) log10 K° = -75.45 - 2+ FeSe2(cr) + H2O = 2 HSe + 0.5 O2(aq) + Fe log10 K° = -60.11

° -1 ∆ of FeSe2(cr) : ± 15.0 kJ mol ° - -1 ∆ of HSe : ± 2.024 kJ mol ° 2+ -1 ∆ of Fe : ± 1.00 kJ mol ° -1 ∆ of H2O : ± 0.041 kJ mol ° ∆ of O2(aq) : no data

15.0 0.041 1.00 2 2.024  = ± 15.57 kJ mol-1 / 2.303 RT (5.708) = ± 2.73

The resulting equation is: - 2+ FeSe2(cr) + H2O = 2 HSe + 0.5 O2(aq) + Fe log10 K° = -60.11 ± 2.73

No reaction constant and associated uncertainty is given by NEA (p. 328 f.).

Samarium

SmOHCO3(s) ANDRA

+ 3+ - SmOHCO3(s) + 2 H ↔ Sm + HCO3 + H2O log10 K° = 2.63 3+ - - + Sm + 2 HCO ↔ Sm(CO3)2 + 2 H log10 K° = -7.85

135

- - SmOHCO3(s) + HCO3 ↔ Sm(CO3)2 + H2O log10 K° = -5.23

° ∆ of SmOHCO3(s) : ± no data given by ANDRA ° - ∆ of Sm(CO3)2 : ± 3.781 ° - -1 ∆ of HCO3 : ± 0.251 kJ mol ° -1 ∆ of H2O : ± 0.041 kJ mol

3.781 0.251 0.041  = ± 3.79 kJ mol-1 / 2.303 RT (5.708) = ± 0.66

The resulting equation is: - - SmOHCO3(s) + HCO3 ↔ Sm(CO3)2 + H2O log10 K° = -5.23 ± 0.66

The reaction constant given by Andra is: + 3+ 2- SmOHCO3(s) + H ↔ Sm + CO3 + H2O log10 K° = -7.00 ± no uncertainty given

Silver The solubility of AgCl(cr) was calculated to correspond to 2.48 × 10-5 M with Ag(HS)(aq) representing the dominant aqueous complex. Thus, the total uncertainty is arising from the uncertainty on the solubility product/constant of AgCl(cr) and the uncertainty associated to the formation constant of AgHS(aq) and is calculated as follows:

AgCl(cr) ANDRA

+ - AgCl(cr) ↔ Ag + Cl log10 K° = -9.75 + 2- + Ag + SO4 + H ↔ Ag(HS)(aq) + 2 O2(aq) log10 K° = -124.23 2- + - AgCl(cr) + SO4 + H ↔ Ag(HS) + Cl + 2 O2(aq) log10 K°= -133.98

° -1 ∆ of AgCl(cr) : ± 0.098 kJ mol ° ∆ of Ag(HS)(aq) : no data given by ANDRA ° - -1 ∆ of Cl : ± 0.117 kJ mol ° 2- -1 ∆ of SO4 : ± 0.418 kJ mol ° ∆ of O2(aq) : no data

0.098 0.418 0.117  = ± 0.44 kJ mol-1 / 2.303 RT (5.708) = ± 0.08

The resulting equation is: 2- + - AgCl(cr) + SO4 + H ↔ Ag(HS) + Cl + 2 O2(aq) log10 K°= -133.98 ± 0.08

Equations given by ANDRA:

136

+ - AgCl(cr) ↔ Ag + Cl log10 K° = -9.75 ± 0.04 + - Ag + HS ↔ Ag(HS)(aq) log10 K° = 14.05 ± ?

Technetium The measured uncertainty for the reaction below is given by Guillaumont et al. (2005, see Vol. 5, p.129):

TcO2 × 1.6 H2O (s) NEA

TcO2 × 1.6 H2O ↔ TcO(OH)2 + 0.6 H2O log10 K° = -8.4 ± 0.5

Tin

SnO2(am) ANDRA

The equations given by ANDRA with Sn(IV): + 4+ SnO2(am) + 4 H ↔ 2 H2O + Sn log10 K° = -6.77 ± 0.73 4+ - Sn + 5 H2O ↔ Sn(OH)5 + 5 H+ log10 K° = -8.50 ± 1.00

0.73 1.00 σ 1.24

The resulting equation is: - + SnO2(am) + 3 H2O = Sn(OH)5 + H log10 K° = -15.27 ± 1.24

Thorium

ThO2 (am,hyd,aged) NEA

+ 4+ ThO2 (am,hyd,aged) + 4 H ↔ Th + 2 H2O log10 K° = 8.5 4+ - - + Th + 3 H2O + HCO3 ↔ Th(CO3)(OH)3 + 4 H log10 K° = -12.26 - - ThO2(am,hyd,aged) + H2O + HCO3 ↔ Th(CO3)(OH)3 log10 K° = -3.76

° ∆ of ThO2 (am,hyd,aged) : no data given by NEA ° - -1 ∆ of Th(CO3)(OH)3 : ± 7.868 kJ mol ° - -1 ∆ of HCO3 : ± 0.251 kJ mol ° -1 ∆ of H2O : ± 0.041 kJ mol

7.868 0.041 0.251  = ± 7.87 kJ mol-1 / 2.303 RT (5.708) = ± 1.38

137

The resulting equation is: - - ThO2(am,hyd,aged) + H2O + HCO3 ↔ Th(CO3)(OH)3 log10 K° = -3.76 ± 1.38

Reaction given by NEA (Vol. 11, p. 51): 4+ - ThO2(am,hyd,aged) + 2 H2O ↔ Th + 4 OH log10 K° = -47.5 ± 0.90

Uranium

UO2(am,hyd) NEA

+ 2+ UO2(am,hyd) + 2 H + 0.5 O2(aq) ↔ UO2 + H2O log10 K° = 35.45 2+ - 4- + UO2 + 3 HCO3 ↔ UO2(CO3)3 + 3 H log10 K° = -9.14 - 4- UO2(am,hyd) + 0.5 O2(aq) + 3 HCO3 ↔ UO2(CO3)3 + H + H2O log10 K° = 26.31

° ∆ of UO2(am,hyd) : ± no data given by NEA ° 4- -1 ∆ of UO2(CO3)3 : ± 2.116 kJ mol ° - -1 ∆ of HCO3 : ± 0.251 kJ mol ° -1 ∆ of H2O : ± 0.041 kJ mol ° ∆ of O2(aq) : no data

2.116 0.041 3 0.251  = ± 2.25 kJ mol-1 / 2.303 RT (5.708) = ± 0.39

The resulting equation is: - 4- + UO2(am,hyd) + 0.5 O2(aq) + 3 HCO3 ↔ UO2(CO3)3 H + H2O log10 K° = 26.31 ± 0.39

Zirconium

Zr(OH)4(am,aged) NEA (Vol. 8, p. 127)

+ 4+ Zr(OH)4(am,aged) + 4 H ↔ Zr + 4 H2O log10 K° = -5.55 ± 0.2 4+ + Zr + 4 H2O ↔ Zr(OH)4(aq) + 4 H log10 K° = -2.19 ± 1.70

0.2 1.7  = ± 1.71

The resulting equation is:

Zr(OH)4(am,aged) ↔ Zr(OH)4(aq) log10 K° = -7.74 ± 1.71

138

Summary of uncertainty calculations

2- - - ° AmCO3OH(am,hyd) + CO3 ↔ Am(CO3)2 + OH log10 K = -7.30 ± 1.17 2+ - ° CaCO3(s) + H+ = Ca + HCO3 log10 K = 1.84 ± 0.2* 2- + - ° MoO2(s)tugarinovite + H2O ↔ MoO4 + 4 H + 2e log10 K = 29.88 ± 0.5

NpO2(am,hyd) + 2 H2O ↔ Np(OH)4(aq) log10 K° = -9.01 ± 1.21 - 2- + NiCO3(cr)gaspéite + HCO3 ↔ Ni(CO3)2 + H log10 K° = -11.26 ± 0.23* - + Nb2O5(s) + 7 H2O ↔ 2 Nb(OH)6 + 2H log10 K° = -28.38 ± ??

Pd(OH)2(s) ↔ Pd(OH)2(aq) log10 K° = -5.40 ± ?? - 3- + PuCO3OH(s) + 2 HCO3 ↔ Pu(CO3)3 + H + H2O log10 K° = -9.99 ± 10.61 (!!!)

Pa2O5(s) + 5 H2O ↔ 2 Pa(OH)5(aq) log10 K° = -18.06 ± ?? + 2+ - RaCO3(s) + H ↔ Ra + HCO3 log10 K° = 2.03 ± 0.25* + 2+ - SrCO3(s)strontianite + H ↔ Sr + HCO3 log10 K° = 1.05 ± 0.23 - 2+ FeSe2(cr)ferroselite + H2O = 2 HSe + 0.5 O2(aq) + Fe log10 K° = -60.11 ± 2.73 - - * SmOHCO3(s) + HCO3 ↔ Sm(CO3)2 + H2O log10 K° = -5.23 ± 0.66 2- + - AgCl(cr) + SO4 + H ↔ Ag(HS) + Cl + 2 O2(aq) log10 K° = -133.98 ± 0.08*

TcO2 × 1.6 H2O(s) ↔ TcO(OH)2 + 0.6 H2O log K = -8.4 ± 0.5 + - ° SnO2(am) + 3 H2O ↔ H + Sn(OH)5 log10 K = -15.27 ± 1.24 - - ° ThO2(am,hyd,aged) + H2O + HCO3 ↔ Th(CO3)(OH)3 log10 K = -3.76 ± 1.38* - 4- UO2(am,hyd) + 0.5 O2(aq) + 3 HCO3 ↔ UO2(CO3)3 + + ° H + H2O log10 K = 26.31 ± 0.39* ° Zr(OH)4(am,aged) ↔ Zr(OH)4(aq) log10 K = -7.74 ± 1.71

*: probably higher uncertainty (due to partly lacking data)

139

9 ANNEX IV: Technical Note on Th-data, i.e. binary Th-carbonate and ternary Th-hydroxo-carbonate complexes

MOLDATA comprises 5 aqueous mixed Th-hydroxo-carbonate species, 3 of them were copied from the NEA TDB (Rand et al., 2008; Vol. 11) and the other 2 were copied from the ANDRA TDB (ThermoChimie v.5).

Table 30: Mixed Th-hydroxo-carbonate complexes comprised in MOLDATA (Source TDB’s: NEA and ThermoChimie v.5) Complex NEA TDB Reference

2- (141) Th(OH)4(CO3) 2- (122) Th(OH)2(CO3)2 Rand et al. (2008) 5- (114) Th(OH)(CO3)4

ANDRA TDB (ThermoChimie v.5)

- (131) Th(OH)3(CO3) Östhols et al. (1994) 4- (123) Th(OH)2(CO3)3 Riglet-Martial and Capdevila (1999)

4- Besides this, MOLDATA comprises two “pure” Th-carbonate species, namely Th(CO3)4 , i.e. the 6- (104)-complex, copied from ANDRA and Th(CO3)5 , i.e. the (105)-complex copied from NEA.

++ ThSO4 .5 Th++++ ++ Th(OH)2

- Eh (volts) Eh Th(OH) (CO ) 0 3 3 Th(OH)+++ Th(OH)4(aq) µ

–.5

25°C 02468101214 pH

Figure 29: Eh-pH diagram of thorium (Th-C-S-O-H) for the BC reference porewater system. Assumed activity of dissolved [Th] = 10-8. Database: MOLDATA. Code: The Geochemist's Workbench - 8.10

- As can be seen in Figure 29, the Th(OH)3(CO3) species, corresponding to the (131)-complex is predicted to represent the dominant species under BC conditions. Therefore, the correctness of this species is of particular importance, as it is also influencing the Th solubility.

140

The complexation constants of the (131) and (123) species comprised in ThermoChimie v.5 and MOLDATA were determined by Östhols et al. (1994) and Riglet-Martial and Capdevila (1999), respectively. Recently, new data for these species (as well as for other hydroxo-carbonate complexes) have become available (Altmaier et al., 2005; Altmaier et al., 2006). Although the publications from Altmaier et al. (2005, 2006) comprise complexation constants for more than 10 Th-hydroxo-carbonate complexes and 5 binary Th-carbonate complexes, only 3 of the former and 1 binary species were retained within the recent NEA review (Rand et al., 2008; Vol. 11). The authors justify their selection on the fact, that the thorough evaluation of all available ThO2(am,hyd) solubility data in carbonate solution (19 sets of experimental data from different groups of authors) under widely varying conditions (pH = 4.5-13, ionic strengths varying from I = 0.1 M – 4 M) revealed that the experimental data could be well described with only a few ternary complexes, with the most important ones being the (114) and (122) and, while the (121), (131) and (141) complexes are only of minor significance.

6- The retained binary complex corresponds to the penta-carbonate complex Th(CO3)5 . The latter will however not be discussed within this Technical note.

Determination of equilibrium constants In order to determine the stoichiometry and equilibrium constants of thorium complexes with carbonate, solubility studies are performed. But generally only the sums logK°s,1yz = (logKsp + logβ°1yz) of the reactions:

2- 4-y-2z - Th(OH)4(am) + zCO3  Th(OH)y(CO3)z + (4-y)OH logK°s,1yz

are determined from the solubility data. In order to derive the formation constants logβ°1yz, knowledge of the solubility product of the solubility limiting solid that was used in the experiment(s) is required. In the experiments performed in carbonate solutions by Altmaier et al. (2005), the solid is considered to correspond to:

4+ - Th(OH)4(am)  Th + 4 OH logKsp = -47.8

Based on the upper reactions, the formation constants of the ternary Th(IV) complexes

4+ - 2- 4x-y-2z xTh + yOH + zCO3  Thx(OH)y(CO3)z logβ°1yz could be derived using the following equation:

logβ°1yz = logK°s,1yz - logKsp

In Table 31 the solubility constants measured by Altmaier et al. (2005) and the derived formation constants are summarized.

141

Table 31: Solubilty constants of Th(OH)4(am) in carbonate solution (log10K°s,1yz) and derived formation constants (logβ°1yz)

Solubility logβ°1yz (I=0) Complex Reaction constants (I=0) (Altmaier et al., log10K°s,1yz 2005 2- + - 111 Th(OH)4(am) + CO3 = ThOH(CO3) + 3 OH < -26.2 < 21.6 2- - - 112 Th(OH)4(am) + 2 CO3 = ThOH(CO3)2 + 3 OH < -18.4 < 29.4 2- 3- - 113 Th(OH)4(am) + 3 CO3 = ThOH(CO3)3 + 3 OH < -14.0 < 33.8 2- 5- - 114 Th(OH)4(am) + 4 CO3 = ThOH(CO3)4 + 3 OH -12.0 ± 0.2 35.8 2- - 121 Th(OH)4(am) + CO3 = Th(OH)2(CO3) + 2 OH -17.1 ± 0.3 30.7 2- 2- - 122 Th(OH)4(am) + 2 CO3 = Th(OH)2(CO3)2 + 2 OH -10.8 ± 0.2 37.0 2- 4- - 123 Th(OH)4(am) + 3 CO3 = Th(OH)2(CO3)3 + 2 OH < -9.9 < 37.9 2- 6- - 124 Th(OH)4(am) + 4 CO3 = Th(OH)2(CO3)4 + 2 OH < -13.3 < 34.5 2- - - 131 Th(OH)4(am) + CO3 = Th(OH)3(CO3) + OH -9.3 ± 0.5 38.5 2- 3- - 132 Th(OH)4(am) + 2 CO3 = Th(OH)3(CO3)2 + OH < -8.4 < 39.4 2- 5- - 133 Th(OH)4(am) + 3 CO3 = Th(OH)3(CO3)3 + OH < -10.9 < 36.9 2- 2- 141 Th(OH)4(am) + CO3 = Th(OH)4(CO3) -7.2 ± 0.3 40.6 2- 4- 142 Th(OH)4(am) + 2 CO3 = Th(OH)4(CO3)2 < -9.1 < 38.7

As can be seen in Table 31, only upper limits for the formation constants are given, except for the complexes marked in bold. This is another reason why NEA did not incorporate these complexes in their database. The (121)-complex was not incorporated in the NEA TDB (although not only an upper limit was given), as it was reported to be a species of only minor significance (Altmaier et al. 2005, 2006).

The formation constants reported in Table 31 are written using the carbonate ion. In MOLDATA however, the bicarbonate ion represents the basis carbonate species. In order to be able to compare the data from Altmaier et al. (2005) with the ones comprised in MOLDATA, the formation constants need to be recalculated using the following equations:

- 2- + (1) HCO3  CO3 + H log10β° = -10.3267 - + (2) H2O  OH + H log10β° = -14. 0014

Example: The formation constant of the ternary (131)-complex is given by Altmaier et al. (2005) as follows:

4+ - 2- - (3) Th + 3 OH + CO3  Th(OH)3CO3 log10β°131 = 38.5

Reaction in MOLDATA:

4+ - - + (4) Th + HCO3 + 3 H2O  Th(OH)3CO3 + 4 H log10β*131 = -12.2614

142

Converting Reaction (3), using equations 1 and 2, results in log10β°131 equal to -13.8309. This means that there is a difference of around 1.5 log units between the data from Altmaier et al. (2005) and the data comprised in MOLDATA (i.e. Östhols et al.; 1994).

In Table 32, the results of all recalculations and the formation constants comprised in MOLDATA are summarized.

Table 32: Recalculated formation constants of ternary Th(IV) complexes

logβ°1yz (I=0) Recalculated logβ*1yz (I=0) logβ*1yz (I=0) Complex 2- - (Altmaier et al., 2005; CO3 ) (Altmaier et al., 2005; HCO3 ) (MOLDATA) 111 < 21.6 -2.7281 no data 112 < 29.4 -5.2548 no data 113 < 33.8 -11.1815 no data 114 35.8 -19.5082 -19.7083 121 30.7 -7.6295 no data 122 37.0 -11.6562 -11.8562 123 < 37.9 -21.0829 -20.8152 124 < 34.5 -34.8096 no data 131 38.5 -13.8309 -12.2614 132 < 39.4 -23.2576 no data 133 < 36.9 -36.0843 no data 141 40.6 -25.7323 -25.9322 142 < 38.7 -37.9590 no data

Comparing the data shows that except for the (131)- complex, the difference between the data of Altmaier et al. (2005) and the MOLDATA corresponds to ~0.2 log units (~0.3 log units for the 123-complex).

As mentioned at the beginning, the (114), (122) and (114) complexes were copied from NEA to MOLDATA. The difference of the 0.2 log units can be referred to a slightly different solubility product used by NEA to recalculate consistent formation constants that could be incorporated into their TDB. The procedure is explained in the following paragraph.

Consistency with NEA database As mentioned above, Altmaier et al. (2005) used the solubility product of Th(OH)4(am) = -47.8 ± 0.3 for the derivation of the formation constants of the ternary Th-complexes. In the NEA review, only the formation constants of the (114), (121), (122), (131) and (141) complexes are discussed in more detail (p. 346-359). In order to evaluate the latter data and to be consistent with the other Th-data, the NEA reviewers first recalculated the solubility constant logK°s,1yz of Altmaier et al. (2005) of -47.8 ± 0.3 based on the SIT coefficients selected in their review, which resulted in a value of logK° = -47.6 ± 0.5. The 0.2 log unit difference in the solubility product translates (also) into the formation constants, as:

logβ°1yz = logK°s,1yz - (-47.6 ± 0.5)

143

In Table 33 the results of the NEA recalculations are summarized.

Table 33: Formation constants of ternary Th(IV) complexes recalculated by NEA (see also p. 350 Vol. 11) Experimentally derived solubility NEA review logβ°1yz (I=0) Complex constants (I=0) logK°s,1yz Recalculated logβ° 1yz (Altmaier et al., 2005) (Altmaier et al., 2005) (I=0) 114 -12.0 ± 0.2 35.6 ± 0.5 35.8 ± 0.3 121 -17.1 ± 0.3 30.5 ± 0.6 30.7 ± 0.4 122 -10.8 ± 0.2 36.8 ± 0.5 37.0 ± 0.4 131 -9.3 ± 0.5 38.3 ± 0.7 38.5 ± 0.6 141 -7.2 ± 0.3 40.4 ± 0.6 40.6 ± 0.5

For comparison and future purposes, the formation constants for all ternary complexes according to the different calculation methods are summarized in Table 34.

Table 34: Recalculated formation constants logβ*1yz of Th(IV) complexes using the bicarbonate ion and NEA data (SIT & solubility product)

logβ°1yz (I=0) Recalculated logβ*1yz Recalculated logβ°1yz (I=0) logβ*1yz (I=0) Complex (Altmaier et al., 2- - (using NEA data; 2- (using NEA data; CO3 ) (MOLDATA; HCO3 ) - 2005; CO3 ) HCO3 ) 111 < 21.6 < 21.4 no data -2.9281 112 < 29.4 < 29.2 no data -5.4548 113 < 33.8 < 33.6 no data -11.3815 114 35.8 35.6 -19.7083 -19.7082 121 30.7 30.5 no data -7.8295 122 37.0 36.8 -11.8562 -11.8562 123 < 37.9 < 37.7 -20.8152 -21.2829 124 < 34.5 < 34.3 no data -35.0096 131 38.5 38.3 -12.2614 -14.0309 132 < 39.4 < 39.2 no data -23.4576 133 < 36.9 < 36.7 no data -36.2843 141 40.6 40.4 -25.9322 -25.9323 142 < 38.7 < 38.5 no data -38.1590

It can be seen, that for the (123) and (131) complexes, the formation constants that are currently comprised in MOLDATA differ by ~0.5 and ~1.8 log units. Based on the recalculated NEA values, both complexes seem to be less stable than the current values suggest.

Binary Th-hydroxo-carbonate complex As already mentioned in the beginning, MOLDATA comprises two binary Th-complexes, i.e. 4- Th(CO3)4 , which according to the specific nomenclature corresponds to the (104)-complex, and 6- Th(CO3)5 corresponding to the (105)-complex. As the latter complex was copied from NEA, only 4- the formation constant of Th(CO3)4 was re-evaluated here. Based on the paper of Altmaier et al. (2005), the formation constant was recalculated using the bicarbonate ion and corresponds to:

144

4+ - 4- + Th + 4 HCO3 = Th(CO3)4 + 4 H logβ°1yz = -11.3068

The current value in MOLDATA is -9.7421. The recalculated value, which is ~1.5 log units lower than the current value suggests that the tetra-carbonate complex seems to be less stable than formerly determined data reveal.

Conclusions

Within the frame of the NEA review, the thorough evaluation of all available ThO2(am,hyd) solubility data in carbonate solution (19 sets of experimental data from different groups of authors) under widely varying conditions (pH = 4.5-13, ionic strengths varying from I = 0.1 M – 4 M) revealed that the experimental data could be well described with only a few ternary complexes, with the most important ones being the (114) and (122), while other complexes, such as the (121), (131) and (141) complexes are of minor significance. Test calculations performed by Altmaier et al. (2005) showed that other ternary and binary complexes give no significant contributions to the solubility under the studied experimental conditions.

NEA retained 3 ternary complexes, i.e. the (114), (122) and (141) complex, as well as one binary 6- complex, i.e. Th(CO3)5 . MOLDATA additionally comprises the (131) and (123) ternary 4- complexes, as well as the binary complex Th(CO3)4 .

The more recently derived data by Altmaier et al. (2005) and current evaluation of the formation constants of the latter species suggests that their stability was overestimated by Östhols et al. (1994) and Riglet-Martial and Capdevila (1999) and that the values of the formation constants should be changed in the next MOLDATA release.

However, it is recommended that for modeling purposes, the formation constants of the (131), (123) and (104) complexes should be taken into account/used from now on already, especially for speciation and solubility calculations.

145

10 References

Allard B., Kipatsi H. and Liljenzin J. O. (1980): Expected species of uranium, neptunium and plutonium in neutral aqueous solutions. J. Inorg. Nucl. Chem., Vol. 42, p. 1015-1027.

Altmaier M., Neck, V. Müller R. and Fanghänel Th. (2005): Solubility of ThO2 · xH2O(am) in carbonate solution and the formation of ternary Th(IV) hydroxide-carbonate complexes. Radiochim. Acta., Vol. 93, p. 83-92.

Altmaier M., Neck, V. Denecke M.A., Yin R. and Fanghänel Th. (2006): Solubility of ThO2 · xH2O(am) and the formation of ternary Th(IV) hydroxide-carbonate complexes in NaHCO3-Na2CO3 solutions containing 0-4 M NaCl. Radiochimica Acta, Vol. 94, p. 495-500.

Andra (2005): Dossier 2005 Argile. Référentiel de comportement des radionucléides et des toxiques chimiques d'un stockage dans le Callovo-Oxfordien jusqu'à l'homme. Site de Meuse/Haute-Marne. Tome 2/2 : Chapitres 5 à 7. ANDRA Document Interne. 186 pages.

Amaya T., Chiba T., Suzuki K., Oda C. Yoshikawa H. and Yui M. (1997): Solubility of Sn(IV) oxide in dilute NaClO4 solution at ambient temperature. Mat. Res. Soc. Symp. Proc., Vol. 465, p. 751-758.

Appelo C. A. J. and Postma D. (2005): Geochemistry, groundwater and pollution. 2nd edition. A. A. Balkema Publishers. 683 pages.

Baes C. F. Jr. and Mesmer R. E. (1976): The hydrolysis of cations. New York, John Wiley and Sons, 498 pages.

Becker V. J., Bennett J. H. and Manuel O. K (1972): Iodine and uranium in sedimentary rocks. Chem. Geol., Vol. 9, p. 133-136.

Berner U. (1995): Estimate of solubility limits for safety relevant radionuclides. PSI-Bericht-95-07.

Berner U. (2002): Project Opalinus Clay: Radionuclide concentration limits in the cementitious near-field of an ILW repository, PSI Bericht Nr. 02-26, Nagra NTB 02-22.

Berner U. and Curti E. (2002): Radium solubilities from SF/HLW wastes using the solid solution and co-precipitation models, Paul Scherrer Institute Villingen, Switzerland, Internal Report, TM-44-02-04.

Bethke C. M. (2010): The Geochemist's Workbench Release 8.0 (four volumes), Hydrogeology Program. University of Illinois, Urbana, Illinois.

Brookins D. G. (1988): Eh-pH Diagrams for Geochemistry. Springer Verlag. 176 pages.

Brown P. L., Curti E., and Grambow B. (2005): Chemical Thermodynamics of Zirconium, Elsevier, OECD NEA, Issy-les-Moulineaux, France.

Bruggeman C., Maes N. and Vancluysen J. (2002): The interaction of technetium-99 with Boom Clay. – Fourth Progress Report. (03/2002 -08/2002).

Bruggeman C., Aertsens M., Maes N. and Salah S. (2010): Iodine retention and migration behaviour in Boom Clay. External Report, SCK•CEN-ER-119. 10/CBr/P-44. Topical Report. First Full Draft.

Bruggeman C., Aertsens M., Govaerts J., Martens E., Jacops E., Van Gompel M. and Van Ravestyn L. (2010): Technetium retention and migration in Boom Clay. External Report, SCK•CEN-ER-101. Topical Report. First Full Draft.

Bruggeman et al. (in prep.): RN retention and migration behaviour in Boom Clay. SFC-1 Level 4 report.

Bruggeman C. (2010): Colloids and colloid migration in clay-rich geological media – Case for Boom Clay. SCK•CEN report, v. 1.02.

146

Bruggeman C., Liu D.J. and Maes N. (2010): Influence of Boom Clay organic matter on the adsorption of Eu3+ by illite – geochemical modelling using the component additivity approach. Radiochim. Acta, Vol. 98, p. 597-605.

Bruggeman C., Salah S., and Maes N. (2012): Americium retention and migration behaviour in Boom Clay. Topical Report, First Full Draft, External report SCK•CEN-ER-201.

Bruno J., Duro L., Grivé M. and Merino J. (2001 a): Review of the radionuclide solubility limits for their use in PA exercises under clayey and cementitious conditions. Technical Report ANDRA C.RP.0ENQ.01.001.

Bruno J., Duro L. and Grivé M. (2001 b): The applicability and limitations of the geochemical models and tools used in simulating radionuclide behaviour in natural waters. Lessons learned from the Blind predictive Modelling exercises performed in conjunction with Natural Analogue studies. SKB Technical Report, TR-01-20.

Bruno J., Bosbach D., Kulik D. and Navrotzki A. (2007): Chemical Thermodynamics of Solid Solutions of Interest in Nuclear Waste Management. A State-of-the-art report. OECD NEA DATABANK, Issy-lès-Moulineaux, France.

Buckau G., Stumpe R. and Kim J. I. (1986): Americium colloid generation in groundwaters and its speciation by laser-induced photoacoustic spectroscopy, J. Less-Common Met., Vol. 122, p. 555-562.

Caro P., Lemaître-Blaise M. and Trombe F. (1968) : Indentification et solubilités des phases solides à l’équilibre sous une atmosphère de gaz carbonique dans les systèmes ternaires oxydes des terres rares –gaz carboniqe-eau. Comp. Rend. Ac. Sci. Paris, Serie C, Vol. 267, p. 1594-1597.

Choppin G. R. (1984): Lanthanide complexation in aqueous solutions. J. Less-Common Met., Vol. 100, p. 141-151.

Choppin G. R. (1986): Speciation of trivalent f-elements in natural waters. J. Less Common Met., Vol. 126, p. 19-26.

Choppin G. R. (2003): Actinide speciation in the environment. Radiochim. Acta, Vol. 91, p. 645-649.

Christiansen B. and Carlsen L. (1989): Iodine in the environment revisited – An evaluation of the chemical and physico-chemical processes possibly controlling the migration behaviour of iodine in the terrestrial environment, Risø-M-2791, Risø National Laboratory, Roskilde, Denmark.

Cramer J. and Smellie J. (1994): Final Report of the AECL/SKB Cigar Lake Analog Study. SKB Technical Report, TR 94-04.

De Cannière P., Moors H., Lolivier P., De Preter P. and Put M. (1996): Laboratory and in situ migration experiments in the Boom Clay. Nuclear Science and Technology Report, EUR16927EN, EC, Luxembourg.

De Cannière P., Maes A., Williams S., Bruggeman C., Beauwens T., Maes N. and Cowper M. (2010): Behaviour of Selenium in Boom Clay. External Report, SCK•CEN-ER-120, 10/PDC/P-9, 328 pages.

De Craen M., Wang L., Van Geet M. and Moors H. (2004): Geochemistry of Boom Clay pore water at the Mol site - Status 2004. SCK-CEN Report BLG-990, Mol, Belgium.

Deelman J. C (1999): Low-temperature nucleation of magnesite and dolomite - Neues Jahrbuch für Mineralogie, Monatshefte, 7, p. 289-302, Stuttgart 1999.

Delany J. M. and Lundeen S. R. (1990): The LLNL Thermochemical Database, Lawrence Livermore National Laboratory Report, UCRL-21658.

Delécaut G. (2004): The geochemical behaviour of uranium in the Boom Clay. PhD Thesis; Université Catholique de Louvain – SCK-CEN, Louvain-La-Neuve, Belgium.

De Soete H. (2010): The Geochemical Data Processor. User's Manual and Function Library. Internal SCK document.

Dierckx A. (1995): Complexation of Europium with Humic Acids. PhD thesis, N° 279, Katholieke Universiteit Leuven, Belgium.

147

Dierckx A., Put M., De Cannière P., Wang L., Aertsens M., Maes A., Vancluysen J., Verdickt W., Gielen R. Christiaens M., Warwick P., Hall A. and Van der Lee J. (1999): TRANCOM –CLAY. Transport of radionuclides due to complexation with organic matter in clay formations. Final Report, R-3388, Contract N° FI4W-CT95-0013.

Dozol M. and Hagemann R. (1993): Radionuclide migration in groundwaters: Review of the behaviour of actinides. Pure & Appl. Chem., Vol. 65, No. 5, p. 1081-1102.

Duro L., Grivé M., Cera E., Doènech C. and Bruno J. (2006): Update of a thermodynamic database for radionuclides to assist solubility limits calculation for performenace assessment. SKB Technical Report, TR-06-17, 128 pages.

Druyts F. and Van Iseghem P. (2003): Conditioning methods for beryllium waste fusion reactors. Fusion Engineering and Design, Vol. 69, Issue 1-4, p. 607-610, 22nd Symposium on Fusion Technology.

Eby R. K. and Birtcher, R.C. (1992): The amorphization of complex silicates by ion-beam irradiation. J. Mater. R., Vol. 7, p. 3080-3102.

Ewing R. C., Meldrum, A., Wang L. and Wang S. (2000): Radiation-induced amorphization. Chapter 12, Reviews in Mineralogy & Geochemistry, Vol. 39/1, p. 319-361.

Ewing R.C., Runde W. and Albrecht-Schmitt T.E. (2010): Environmental impact of the nuclear fuel cycle: Fate of actinides. MRS Bulletin, Vol. 35, p. 859-866.

Fanghänel Th. and Neck V. (2002): Aquatic chemistry and solubility phenomena of actinide oxides/hydroxides. Pure Appl. Chem., Vol. 74, N°: 10, p. 1895-1907. IUPAC.

Fayek M., Janeczek J. and Ewing R. (1997): Mineral chemistry and oxygen isotope analysis of uraninite, pitchblende and uranium alteration products from Cigar Lake deposit, Saskatchewan, Canada. Applied Geochemistry, Vol. 12, p. 549-656.

Gamsjäger H., Bugajski J., Gajda T., Lemire R. J. and Preis W. (2005): Chemical Thermodynamics of Nickel, OECD NEA DATABANK, Issy-lès-Moulineaux, France.

Garrels R. M. and Christ C. L. (1965): Solutions, Minerals, and Equilibria, Harper & Row, John Weatherhill, Inc.

Gascoyne M. (1992): Geochemistry of the actinides and their daughters. 2nd edn, In: Uranium Series Disequilibrium: Applications to earth, marine, and environmental sciences, Eds. M. Ivanovich & R. S. Harmon, Clarendon Press, Oxford.

Grenthe I., Fuger, J. Konings R. J. M. Lemire R. J. Muller A. B., Nguyen-Trung C. and Wanner H. (1992): Chemical Thermodynamics of Uranium. OECD NEA, Issy-les-Moulineaux, France.

Guillaumont (1966): Contribution à l’étude des espèces ioniques du protactinium en solution aqueuse. Thèse de Doctorat, Université de Paris XI, N° d’ordre 5528.

Guillaumont R., Fanghänel T., Fuger J., Grenthe I., Neck V., Palmer D. A. and Rand M. H. (2003): Update on the Chemical Thermodynamics of Uranium, Neptunium, Plutonium, Americium and Technecium. Vol. 5. OECD NEA DATABANK, Issy-lès-Moulineaux, France.

Handbook of Chemistry and Physics, 73rd Edition by Lide, 1992-93. Copyright CRC Press, Boca Raton; Florida.

Helgeson H. C. (1979): Mass transfer among minerals and hydrothermal solutions. In: Geochemistry of Hydrothermal Ore Deposits. (ed. H. L. Barnes), p. 568-610. John Wiley and Sons, New York.

Helgeson H. C., Murphy W. M, and Aargard P. (1984): Thermodynamic and kinetic constraints on reactions among minerals and aqueous solutions. II. Rate constants, effective surface area, and the hydrolysis of feldspar. Geochimica et Cosmochimica Acta, Vol. 48, p. 2405-2432.

Henrion P. N, Put M. J. and van Gompel M. (1991): The influence of compaction on the diffusion of non-sorbed species in Boom Clay. Radioact. Waste Manage. Nucl. Fuel Cycle, Vol. 16(1), p. 1-14).

148

Hummel W., Anderegg G., Puigdomenech I., Rao L. and Tochiyama O. (2005). Chemical Thermodynamics of Compounds and Complexes of U, Np, Pu, Am, Tc, Se, Ni and Zr with Selected Organic Ligands. OECD NEA DATABANK, Issy-les-Moulineaux, France.

Hummel W., Berner U., Curti E., Pearson F.J. and Thoenen T. (2002): NAGRA/PSI Chemical Thermodynamic DataBase 01/01. NAGRA Technical Report 02-16. Universal Publishers, Parkland, Florida.

IAEA (2005): Thorium fuel cycle — Potential benefits and challenges. AEA-TECDOC-1450. Vienna, Austria.

Janeczek J. and Ewing R. C. (1992): Dissolution and alteration of uraninite under reducing conditions. Journal of Nuclear Materials, Vol. 190, p. 157-173.

Jensen K. A., Janeczek J., Ewing R. C., Stille P., Gauthier-Lafaye F. and Salah S. (2000): Crandallites and coffinite: retardation of nuclear reaction products at the Bangombé natural fission reactor. Material Research Society Proceedings, Vol. 608, p. 525-532.

Johnson J. W., Oelkers E. H. and Helgeson H. C. (1991): SUPCRT92: A software package calculating the standard molal thermodynamic properties of minerals, gases, aqueous species, and reactions from 1 to 5000 bars and 0°C to 1000°C. Earth Sceince Dept., Livermore, CA 94550.

Jörg G., Bühnemann R., Hollas S., Kivel N., Kossert K., Van Winckel S. and v. Gostomski C. L. (2010): Preparation of radiochemically pure 79Se and highly precise determination of ist half-life. Applied Radiation and Isotopes, Vol. 68, p. 2339-2351.

Kim J. I. (1986): Chemical behaviour of Transuranic Elements in Natural Aquatic Systems. In: Handbook on the Physics and Chemistry of the Actinides, Eds. A. J. Freeman and c. Keller, Elsevier Science Publishers B. V., 1986, Chapter 8.

Kim J. I. (1991): Actinide colloid generation in groundwater. Radiochimica Acta, 52/53, p. 71-81.

Kim, S. S., Baik M. H., Kang K. C., Kwon S. H. and Choi J. W. (2008): Solubilities of actinides in a domestic groundwater and a bentonite porewater calculated by using PHREEQC. Journal of Industrial and Engineering Chemistry, Vol. 14, p. 739-746.

Knopp R., Neck V. and Kim J. I. (1999): Solubility, hydrolysis, and colloid formation of plutonium (IV). Radiochim. Acta, Vol. 435, p. 1-8.

Kraus K. A. and Nelson F. (1950): Hydrolytic behaviour of metal ions I. The acid constants of uranium (IV) and plutonium (IV). J. Amer. Chem. Soc., Vol. 72, p. 3901.

Kulik D. A, Berner U. and Curti E. (2004): Modelling chemical equilibrium partitioning with the GEMS-PSI code. PSI Scientific Report, 2003, Vol. IV, Nuclear Energy and safety, Smith B., Gschwend B., Eds. P. 109-122. Paul Scherrer Institut.

Kulik D. A. (2005): Short course with tutorial on Aqueous – Solid Solution Systems involving actinides. Lecture 6: Trace metal incorporation in carbonates and sulphates: an overview (focus on actinides and fission products). Villingen PSI, November 16-18, 2005.

Kulik D. A. and Curti E. (2005): Forward modeling of AqSS equilibria with GEMS-PSI Code. AqSS Short Course, Tutorial 1 (carbonates), Villingen PSI, November 16-18, 2005.

Kulik D. A., Vinogard V. L., Paulsen N. and Winkler B. (2010): (Ca,Sr)CO3 aqueous soolid solution systems: From atomistic simulations to thermodynamic modelling. Physics and Chemistry of the Earth, Vol. 35, p. 217-232.

Langmuir (1997): Chapter 13. In: Aqueous Environmental Geochemistry. Upper Saddle River, NJ: Prentice Hall.

Langmuir D. and Riese A. C. (1985): The thermodynamic properties of radium. Geochimica et Cosmochimica Acta, 49, p. 2423-2432.

149

Lee J. H. and Byrne R. H. (1992): Complexation of trivalent rare earth elements (Ce, Eu, Gd, Tb, Yb) by carbonate ions. Geochimica et Cosmochimica Acta, Vol. 57, p. 295-302.

Lemire R.J. (1984): An assessment of the thermodynamic behavior of neptunium in water and model groundwater from 25 to 150°C. AECL-7817, Atomic Energy of Canada Limited, Pinawa, Manitoba, Canada, 53 pages.

Lemire R. J., Fuger J., Nitsche H., Potter P., Rand M. H., Rydberg J., Spahiu K., Sullivan J. C., Ullman W. J., Vitorge P., and Wanner H. (2001): ChemicalTthermodynamics of Neptunium and Plutonium. OECD NEA, Issy-les- Moulineaux, France.

Lierse C., Treiber W. and Kim J. I. (1985): Hydrolysis reactions of Neptunium(V). Radiochim. Acta, Vol. 38, p. 27f.

Liu, D. J., Bruggeman C. and Maes N. (in prep.): Influence of organic matter on the solubility of Th in Boom Clay geochemical conditions.

Liu D. J., Bruggeman C. and Maes N. (2008): The influence of natural organic matter on the speciation and solubility of Eu in Boom Clay proewater. Radiochimica Acta, 96, 9-11, p. 711-720.

Lothenbach B., Ochs M. and Hager D. (2000): Thermodynamic data for the solubility of tin(IV) in aqueous cementitious environments. Radiochim. Acta, Vol. 88, p. 521-526.

Lumpkin G. R., Smith K. L., Blackford M. G. Gieré R. and Williams C.T. (1998): The crystalline-amorphous transformation in natural zirconolite: evidence for long-term annealing. Mater. Res. Soc. Symp. Proc., 506, p. 215- 222.

Maes A., Bruggeman C., Geraedts K. and Vancluysen J. (2003): Quantification of the interaction of Tc with wissolved Boom Clay humic substances. Environmental Science and Technology. Vol. 37 (4), p. 747-753.

Maes A., Geraedts K., Bruggeman C., Vancluysen J., Rossberg A. and Hennig C. (2004): Evidence for the interaction of technetium colloids with humic substances by X-ray absorption spectroscopy. Environmental Science and Technology. Vol. 38 (7), p. 2044-2051.

Maes N., Wang L., Delécaut G., Beauwens T., Van Geet M., Put M., Weetjens E., Marivoet J., van der Lee J., Warwick P., Hall A., Walker G., Maes A., Bruggeman C., Bennett D., Hicks T., Higgo J., Galson D. (2004): Migration Case Study: Transport of radionuclides in a reducing Clay Sediment (TRANCOM-II) - Final Scientific and Technical Report of the EC TRANCOM-II project, BLG-988, SCK-CEN, Mol, Belgium.

Maes N., Bruggeman C., Govaerts J., Martens E., Salah S. and Van Gompel M. (2011): A consistent phenomenological model for natural organic matter linked migration of Tc(IV), Cm(III), Np(IV), Pu(III/IV) and pa (V° in the Boom Clay. Physica and Chemistry of the Earth, Vol. 36, p. 1590-1599.

Marivoet J., Volckaert G., Labat S., De Cannière P., Dierckx A., Kursten B., Lemmens K., Lolivier P., Mallants D., Sneyers A., Valcke E., Wang L. and Wemaere I. (1999): Values for the near-field and clay parameters used in the performance assessment of the geological disposal of radioactive waste in the Boom Clay formation at the Mol site. SCK-CEN Report, Vol. 1 & 2, R-3444, Mol, Belgium.

Mercier F., Moulin V., Guittet M. J., Barre N. Toulhoat N., Gautier-Soyer M. and Toulhoat P. (2000): Applications of NAA, PIXE and XPS for the quatification and characterization of the humic substances/iodine association. Radiochimica Acta, Vol. 88 (9-11), p. 779-785.

Moon H. (1989): Equilibrium ultrafiltration of hydrolysed thorium (IV) solutions, Bull. Korean Chem. Soc., Vol. 10, p. 270-277.

Neck V. and Kim J. I. (2001): Solubility and hydrolysis of tetravalent actinides. Radiochim. Acta, Vol. 89, p. 1-16.

Nguyen-Trung C., Cuney M. and Poty B. (1995): Etude experimentale de la solubilité du β-Nb2O5 crystallisé dans des solutions aqueuses (0 ≤ pH ≤ 14) à force ionique elevée (0.5 M ≤ I ≤ 6 M) à 25 et 50°C, sous 0.1 MPa. Andra Technical Report C. RP.0CRE.95.001.

Nordstrom D. K. and Munoz J. L. (1994): Geochemical Thermodynamics, Blackwell Scientific Publication, Oxford.

150

OECD (2003): SAFIR-2: Belgian R&D programme on the deep disposal of high-level and long-lived radioactive waste. An International Peer Review. Radioactive Waste Management. 77 pages.

Oelkers E. H. and Schott J. (2009): Thermodynamics and Kinetics of Water-Rock Interaction. Reviews in Mineralogy & Geochemistry. Vol. 70. Mineralogical Society of America, Geochemical Society, 569 pages.

Östhols E., Bruno J. and Grenthe I. (1994): On the influence of carbonate on mineral dissolution: III. The solubility of microcrsystalline ThO2 in CO2-H2O media. Geochim. Cosmochim. Acta, Vol., 58, p. 613-623.

Olin A., Nolang B., Ohman L., Osadchii E. and Rosen E. (2005): Chemical Thermodynamics of Selenium, OECD NEA, Issy-les-Moulineaux, France.

Olofsson U. and Allard B. (1983): Complexes of actinides with naturally occurring substances. Literature survey. SKBF 83-09.

ONDRAF/NIRAS (2001): Aperçu technique du rapport SAFIR 2: Safety Assessment and feasibility Interim Report 2. NIROND 2001-05 F.

ONDRAF/NIRAS (2006): Overview of SFC-1 quality measures and activities - V2 May 2007, NIROND note 2006- 1103, Working Document v. 1.4, June 2006.

ONDRAF/NIRAS (2011): Plan Déchets pour la gestion à long terme des déchets radioactifs conditionnées de haute activité et/ou de longue durée de vie et aperçue de questions connexes Rapport NIRONS 2011-02 F.

Pearson F. J., Berner U. and Hummel W. (1992): Nagra Thermochemical Data Base II. Supplemental Data 05/92. Nagra Technical Report NTB 91-18, Nagra, Wettingen, Switzerland, 284 pages.

Peiffert C., Nguyen-Trung C. and Landais P. (1997): Etude experimentale de la solubilité des oxydes de Nb(V) crystallisé et amorphe dans des solutions aqueuses. Andra Technical Report C.RP.0CRE.97.003.

Penneman R.A. and Keenan T. K. (1960): The radiochemistry of americium and curium. University of California, Los Alamos, California.

Plummer L. N. and Busenberg E. (1987): Thermodynamics of aragonite –strontianite solid solutions: results from stoichiometric solubility at 25 and 76°C. Geochim. et Cosmochim. Acta, Vol. 51, p. 1393-1411.

Plummer L. N., Busenberg E., Glynn P. D. and Blum A. E. (1992): Dissolution of aragonite-strontianite solid solutions in nonstoichiometric Sr(HCO3)2-Ca(HCO3)2-CO2-H2O solutions. Geochim. et Cosmochim. Acta, Vol. 56, p. 3045-3072.

Rand M., Fuger J., Grenthe I., Neck V. and Rai D. (2009): Chemical Thermodynamics of Thorium, OECD NEA, Issy-les-Moulineaux, France.

Rard J. A., Rand M. H. Anderegg G. and Wanner H. (1999): Chemical Thermodynamics of Technecium. OECD NEA, Issy-les-Moulineaux, France.

Richard L. and Gaona X. (2011): Thermodynamic properties of organic iodine compounds. Geochim. et Cosmochim. Acta, Vol. 75, p. 7304-7350.

Riglet Martial Ch. and Capdevila H. (1999): Solubilité des actinides(IV) en milieu hydroxo-carbonate: Bibliographie sur le Thorium et l’Uranium. CEA N.T. SESD 99-29.

Robinson R.A. and R.H. Stokes (1968): Electrolyte Solutions. Butterworths, London.

Robit-Pointeau V. (2006): Etude des phases néoformées lors de la dissolution du combustible nucléaire en condition de stockage géologique: influence des ions silicate. Thèse CEA Saclay, Direction de l’Energie Nucléaire Département de Physico-Chimie Services d’Etudes du Comportement des Radionucléides. CEA-R-6119. 208 pages.

151

Runde W., Meinrath G. and Kim J. I. (1992): A study of solid-liquid phase equilibria of trivalent lanthanide and actinide ions in carbonate systems. Radiochim. Acta, Vol. 58/59, p. 93-100.

Runde W. (2000a): The chemical interactions of actinides in the environment. Los Alamos Science, N° 26, p. 392- 411.

Runde W., Van Pelt C. and Allen P. G. (2000 b): Spectroscopic characterization of trivalent f-element (Eu, Am) carbontaes. Journal of Alloys and Compounds, Vol. 303-304, p. 182-190.

Ryan J. L. and Rai D. (1987): Thorium(IV) hydrous oxide solubility. Inorg. Chem., Vol. 26, p. 4140-4142.

Sannen L., De Raedt C., Moons F. and Yao Y. (2002): Helium content and induced swelling of neutron irradiated beryllium. Fusion Engineering and Design, Vol. 29, p. 470-474.

Schindler P.W. (1967). Heterogeneous equilibria involving oxides, hydroxides, carbonates and hydroxide carbonates. In: Equilibrium Concepts in Natural Water Systems, Adv. Chem. Ser., No. 67, American Chemical Society, Washington DC, p. 196.

Scibetta M., Pellettieri A. and Sannen L. (2007): Experimental determination of creep properties of beryllium irradiated to relevant fusion power reactor doses. Proceedings of the 12th International Conference on Fusion reactor materials (ICFRM-12), Santa Barbara, U.S.A, 4-9 Dec. 2005.

Silva R. J., Bidoglio G., Rand M. H,. Robouch P. B., Wanner H. and Puigdomenech I. (1995): Chemical Thermodynamics of Americium. Vol. 2, OECD NEA DATABANK, Issy-les-Moulineaux, France.

Silva R. J. and Nitsche H. (1995): Actinide environmental chemistry. Radiochimica Acta, Vol 70/71, p. 377-396.

Spahiu K. and Bruno J. (1995): A selected thermodynamic database for REE to be used in HLNW performance assessment exercises. SKB Technical Report, TR-95-35. 80 pages.

Steefel C. I. and Van Cappellen P. (1990): A new kinetic approach to modeling water-rock interaction: The role of nucleation, precursors, and Ostwald ripening. Geochim. et Cosmochim. Acta, Vol. 54, p. 2657-2677.

Stevenson F. J. (1982): In: Humus Chemistry: Genesis, composition, reactions. John Wiley and Sons. New York.

Stumm W. (1992): Chemistry of the Solid-Water Interface. John Wiley and Sons , New York.

Stumm W. and Morgan J. J. (1996): Aquatic Chemistry, 3rd ed., John Wiley and Sons .

Sverjenski D. A. (1984): Prediction of Gibbs free energies of calcite-type carbonates and the equilibrium distribution of trace elements between carbonates and aqueous solutions. Geochim. et Cosmochim. Acta, Vol. 48, p. 1127-1134.

Thorstenson D. C. and Plummer L. N. (1977): Equilibrium criteria for two-component solids reacting with a fixed composition in aqueous phase, Am. J. Sci., 277, p. 1203-1223.

Triay I., Simmons A., Levy S., Nelson S., Nuttal H., Robinson B., Steinkampf W. and Vianni B. (1995): Progress report on colloid facilitated transport at Yucca Mountain. Los Alamos National Laboratory Report LA 12779-MS, Los Alamos.

Trubert D., Le Naour C. and Jaussaud C. (2002): Hydrolysis of Protactiunium(V). I. Equilibrium Constants at 25°C: + + - A Solvant Extraction Study with TTA in the Aqueous System Pa(V)/H2O/H /Na /ClO4 .

Trubert D., Le Naour C., Jaussaud C. and Mrad O. (2003) : Hydrolysis of Protactiunium(V). III. Determination of Standard Thermodynamic Data. Journal of Solution Chemistry, Vol. 32, No. 6, p. 505-517.

Valcke E., Sneyers A. and Van Iseghem P. (2003): Chemical degradation of cellulose based waste. Final Report. SCK•CEN-R-3785. 03/Eva/P-76.

152

Van der Lee J., Ledoux E., de Marsily G., De Cayeux M., Van de Weerd H., Fraters B., Dodds J., Rodier E., Sardin M. and Hernandez, A. (1994). A bibliographical review of colloid transport through the geosphere. European Commission Report EUR 15481, Luxembourg.

Van Loon L. R. and Glaus M. A. (1998): Experimental and theoretical studies on alkaline degradation of cellulose and its impact on the sorption of radionuclides. PSI Bericht, Nr. 98-07, 136 pages.

Wang L., Dierckx A. and De Cannière P. (2000): Speciation and solubility of radionuclides in Boom Clay. SCK•CEN Report R-3508.

Wang L., Marivoet J., De Cannière P., Sillen X., Moors H. Aertsens M., Put M. and Dierckx A. (2001): Data collection form: The link between migration studies and performance assessment – Belgian Case. The Use of Thermodynamic Databases in Performance Assessment. Workshop Proceedings Barcelona, Spain, 29-30 May 2001.

Wang L., Martens E., Jacques D., De Cannière P., Berry J. and Mallants D. (2009) : Review off sorption values for the cementitious near-field of a near surface radioactive waste disposal facility. Project near surface disposal of category A waste at Dessel. NIRAS-MP5-03 DATA-LT(NF). NIROND-TR 2008-23 E.

Wang, L., Salah, S., and De Soete, H. (2011): MOLDATA: A thermochemical data base for phenomenological and safety assessment studies for disposal of radioactive waste in Belgium – Data compilation strategy. Draft, v. 1.0. External Report SCK•CEN-ER-121.

Wang L. (2011): Solubility of radionuclides in the Supercontainer near-field. Transferability document: use of solubility determined on cAt project in B&C waste disposal assessment. External Report. SCK•CEN-ER-239.

Weetjens E., Marivoet J. and Govaerts J. (2012): Preparatory Safety Assessment. Conceptual model description of the reference case. First Full Draft. External report, SCK•CEN-ER-215.

Wersin P. and Schwyn, B. (2004): Project Opalinus Clay. Integrated Approach for the Development of Geochemical Databases used for Safety Assessment. NAGRA Technical Report 03-06.

Wickham S. M., Crawford M. B. and Bennett D. G. (2005): Belgian Supercontainer Design for HLW and Spent Fuel Disposal. Evaluation of the Reference Design. GALSON Report 0460-5, Version 1.0.

Wood S. A. (1990): The aqueous geochemistry of the rare earth elements and yttrium. 1. Review of available low temperature data for inorganic complexes and the inorganic REE speciation of natural waters. Chem. Geol., Vol. 82, p. 159-186.

Woodhead J. A., Rossman G. R. and Silver L. T. (1991): The metamictization of zircon. Radiation dose-dependent structural characteristics. American Mineralogist, Vol. 76, p. 74-82.

Yui M., Azuma J. and Shibata M. (1999): JNC Thermodynamic Database for Performance Assessment of High-level Radioactive Waste Disposal System. Tokai Works, Japan Nuclear Cycle Development Institute. JNC TN8400, 99- 070.

Zhang J.-W. and Nancollas G. H. (1990): Mechanisms of growth and dissolution of sparingly soluble salts. Chapter 9. In: Mineral-Water Interface Geochemistry. Reviews in Mineralogy. Eds. Hochella M. F. Jr. and White A. F., p. 365- 393.

153