Project Topics
Get Complete Project Material File(s) Now! »
Geological nuclear waste disposal
Geological disposal was defined in a 1995 Collective Opinion of the Nuclear Energy Agency
(NEA) Radioactive Waste Management Committee entitled “The Environmental and Ethical Basis of Geological Disposal”[1]. A deep geological repository is a nuclear waste repository excavated deep within a stable geologic environment (typically below 300 m or 1000 feet). Nowadays, deep geological disposal has been widely adopted in many countries to be the preferential solution for final disposing of intermediate-level long-lived (IL-LL) and high-level (HL) radioactive waste, at depths between 250 m and 1000 m for mined repositories, or 2000 m to 5000 m for boreholes. On the other hand, intermediate-level short-lived (IL-SL) and low-level (LL) radioactive waste will be disposed in the near-surface, at ground level, or in caverns below ground level (at depths of tens of meters). As shown in Figure 1.1, a multiple barriers principle is applied for the disposal of HL radioactive waste in Sweden.
The International Atomic Energy Agency (IAEA) estimates that up to the early 2000s, low-and intermediate-level radioactive waste worldwide has been generated up to a cumulative volume of 7.3×106 m3, and the produced cumulative volume of HL waste has reached 8.3×105 m3[2]. However, nearly 0% of HL waste is in disposal, with ~80% of LL/IL-SL waste and ~20% of IL waste in disposal. Thus, the disposal of high-level radioactive waste is extremely urgent and a sustainable disposal of low- and intermediate-level radioactive waste should be also supported in the long term.
Multi-barrier systems
A mined repository is the most widely proposed option within the deep geological disposal concept, generally relying on a multi-barrier system to isolate the waste from the biosphere for centuries. This multi-barrier principle creates an overall system robustness that enhances confidence that the waste will be successfully contained. The multiple barriers typically comprises metal (e.g., iron or copper) or concrete canisters/casks containing the vitrified radioactive waste, cement or clay (e.g., bentonite) buffer and/or backfill materials, and repository host rocks (e.g., Callovo-Oxfordian clay[3], Boom clay[4], or Opalinus clay[5], granite[6], welded volcanic tuff rocks[7], and layered salt strata or domes). The choice of waste container materials and design, as well as the buffer/backfill material varies depends on the type of waste to be contained and the nature of the host rock-type available.
In France, about 80% of the electricity is generated from its 58 nuclear power plants, being a world leader in the technology. The French radioactive waste disposal agency, ANDRA, is designing a deep geological repository, Cigéo project, in a clay formation at Bure, in eastern France. This will host vitrified HL waste and long-lived IL waste. With the grey 150-million-year-old Callovo-Oxfordian argillaceous rock as the host rock, the disposal concept in France is to use the reinforced concrete, stainless steel casks, and glass as the engineered barrier system (EBS). The specially formulated Fe-reinforced cement backfill material would provide a long-lasting alkaline environment that contributes to containment of the waste by preventing many radionuclides from dissolving in the groundwater. Similar cement-based schemes have been proposed in Switzerland, Czech Republic, Finland, Sweden, Germany, Spain, UK, and US[8].
Besides, huge amounts of cementitious materials are also used for IL waste disposal in France, with an approximately three-time volume of the waste. In the repositories, reinforced cementitious materials are used for tunnel support and backfill, waste containers and waste matrixes, which are considered as barriers that inhibit the mobility of RNs in case of eventual leakage.
Sorption behavior of radionuclides in potential barriers
The mobility of radionuclides (RN) is a key issue regarding the safety assessment of nuclear waste repositories, governed by the geochemical interaction with each barrier. To a large extent, the extremely low mobility of the most radiotoxic RNs, the actinides, increases the reliability of long-term safety of deep nuclear waste repositories. However, a few RNs, mainly anionic species like I129, Cl36, Se79, and Tc99, would diffuse fast and contribute to the ultimate radioactive exposure risks to the biosphere, according to leakage scenarios[9]. Less mobile RN on barriers will result in a larger distribution coefficient (Kd). Figure 1.3 shows the effect that even a small Kd value would have on 129I fluxes and retention time at the top of the Callovo-Oxfordian GBS[10]. In the spent fuel, 96% of the mass is the remaining uranium: most of the original 238U and a little 235U. Thus, uranium migration behavior in multi-barrier system is also of essence when evaluating nuclear waste disposal options. A detailed scientific understanding of the chemical form of solid phase associated RNs, and of the possible sorption mechanisms at work, constitutes a priority research area to enhance our confidence for repository safety.
On iron phases existing in canisters and steel reinforcements
Iron phases, including the iron canister (at low pH) and embedded steel (at high pH) in EBS, are expected to retard the migration of RNs through different interactions. Besides, iron corrosion products, such as, magnetite (Fe3O4), hematite (Fe2O3), goethite (FeOOH), and siderite (FeCO3), can also contribute to the RNs immobilization. Vital to the migration behaviors of redox-sensitive radionuclides (RNs)[12], the redox potential (Eh) in EBS will be governed by the corrosion of iron, producing a thin (over a few microns) film based on the dynamic FeII-oxyhydroxides/magnetite/FeIII-oxyhydroxides thin-film “sandwich” structure[13]. Under aerobic conditions, Fe3O4 is the intermediate oxidation product on the steel surface while Fe(OH)2 is the intermediate product under anaerobic conditions[14]. Previous studies have reported that the oxidized RNs species can be immobilized by Fe0 and Fe-oxyhydroxides via reductive precipitation and surface adsorption. For instance, aqueous UVI can be reduced by nanoscale zero-valent iron (NZVI)[15] into UO2. Reductive precipitation can also occur on colloidal hematite with Fe2+ [16] and nanocrystalline magnetite [17, 18]. Besides, FeIII-oxyhydroxides, e.g., goethite, can also remove aqueous UVI by surface complexation[19]. Uranium co-precipitation with iron oxide minerals (e.g., hematite and goethite) indicates that U6+ is incorporated in the Fe oxides as uranate until a point of saturation is reached, not as UO22+ ion, the one with large ionic radius (~1.8 Å)[20]. Furthermore, selective and highly efficient UO22+ adsorption, separating from Ln3+ at pH ~3, is observed on a magnetic nanocomposite, Fe3O4@ZIF-8[21]. Regarding aqueous SeIV, reduced species, e.g., Se0, FeSe, and FeSe2, could be also obtained on these Fe-oxyhydroxides[22]. In addition, the presence of sorbed Fe(II) can enhance Se(IV) sorption on calcite compared to that of Fe-free pure calcite, resulting in half immobilized Se reducing into Se(0). Selenite can be reduced into FeSe by Fe/FeC3 ultrasmall particles according to a previous report[23]. Antimonate (Sb(OH)6-) also belongs to the mobile anions in repositories. In reducing environments, lower oxidation state Sb species, such as SbIII and Sb0, can be formed. For instance, magnetite can reduce SbV into SbIII and the reduction increases with pH values [24]. SbIII was also found present at the surface of Fe-bearing rims [25]. In addition, Sb0 was detected in soil samples from Swiss shooting ranges, probably derived from reduction by Pb0 into PbSb alloy-containing unweathered bullet fragments[26]. Thus, Fe0, with even lower standard reduction potential than Pb0, can also generate Sb0 [27]. Fe-oxyhydroxides[24] also show strong sorption affinity to SbV, resulting in reductive precipitates (e.g., Sb2O3) and structural incorporation[26, 28]. In contrast, molybdate (MoO42-) is much more difficult to be reduced, even by NZVI[29], and non-redox sorption occurred prevailingly on Fe-oxyhydroxides systems[30, 31]. In contrast, ferrihydrite[32], with much higher specific surface area, is proven to be a good sorbent to molybdate[31, 33] via structural incorporation mechanisms during phase transformations. Besides, aqueous selenium oxyanions[34] and Sb(V)[28] can also be removed efficiently during the ferrihydrite-hematite or ferrihydrite-goethite recrystallization. Regarding embedded steel (Fe0) in cement-based medium, the hyperalkaline condition (pH ~13) would heavily influent both steel corrosion processes and the redox potential imposed, and thus sorption behavior of RN anions on the resulting Fe-oxyhydroxides, which are still not well-documented.
On hydrated cement phases existing in backfills and tunnels
Backfill/buffer materials and cementitious materials from the tunnels and alveolis can act as secondary barriers in the case of RN leakage from canisters. Unlike redox-reactive iron oxides, the cementitious structure can add potentially a retardation factor for RNs diffusion via surface adsorption, ion exchange, or co-precipitation reactions. Regarding ordinary Portland cement (OPC), its hydration products include crystalline portlandite, amorphous calcium silicate hydrate phases (CSH, the major component in hydrated cement)[35], and other minor phases – hydrogarnet, hydrotalcite, ettringite[36, 37] and AFm phases[38, 39], gypsum, calcite, metal sulfides from blast furnace slags, and possibly pyrite.
On portlandite and CSH phase
It can be foreseen that the extremely alkaline environment imposed by portlandite will dramatically decrease the mobility of cationic RNs, due to their enhanced adsorption at high pH. Furthermore, CSH phase, referred to as “cement gel”, is expected to have a high density of sorption sites for both cations and anions due to its low crystallinity and imperfect “structure”[40]. The electrostatic sorption of alkali metal ions (including Cs+) on negatively charged CSH was found to increase inversely in proportion to the Stokes radius of the hydrated ions for cements with various Ca:Si ratios[41]. Tits et al.[42] have shown that the sorption process controlling the retention of Th(IV) and Eu(III) on CSH is fast. Related sorption mechanisms were investigated[43, 44], indicating that the surface complexed or co-precipitated Eu can be attributed to a Eu-substitution for Ca in the framework of the CSH and a high hydroxylated environment. Sorption of anions, such as Cl- and I-, is found to be directly related to the surface charge of CSH phases below millimolal level[45] and Noshita et al.[46] concluded that electrostatic sorption may dominate I- uptake. Besides, CSH materials could be one of the cementitious phases controlling U(VI) uptake in degraded cement pastes[47]. Sorption of UO22+ [48] on hardened cement paste (HCP) has been also studied, showing that precipitation of calcium-uranate predominates.
On AFm phases
Numerous studies have reported RNs sorption behavior on cement hydration products[49]. Through hardened cement paste (HCP), anionic RNs will diffuse much faster than the cations due to the hyperalkaline condition, thus, more intriguing research topic is the sorption behavior of anions. Among the hydration products, AFm phases, hydrated calcium aluminates belonging to layered double hydroxides (LDHs)[38], have been considered as effective phases for the remediation of aquatic environments, to remove anionic contaminants. LDHs refer to a family of lamellar mixed hydroxides with a generic chemical formula [MII6−xMIIIxO6]x+(An-)x/n•mH2O. In the case of the AFm phase (AFm-SO4) in cement, MII, MIII, and An- represent CaII, AlIII, and SO42-, respectively. When concrete is subject to Cl- corrosion (e.g., exposed to sea water, or deicing salt), or when Cl-bearing mixing water, aggregate and admixtures are used to produce it [50, 51], AFm-Cl2 ([Ca4Al2O6]2+•(Cl-)2•mH2O) can also be present in cement hydration products. These micro- and nano-porous phases have a unique structure composed of positively charged main layers and negatively charged interlayers, being tolerant of oxyanions substitution in the crystalline layer and anion exchange with weak-binding intercalated anions. As widespread artificial LDHs around the world, AFm phases can be detected or thermodynamically predicted to occur in bottom ash [52, 53] and Class C coal fly ash (CFA) [54-56], which are two of the main coal combustion products (CCPs) with a global production of over 750 million tonnes per year [57]. AFm phases have strong ability to remove toxic soluble anions via surface adsorption and anion exchange processes. It is found that only the AFm-SO4 phase, competing with ettringite, hydrotalcite, chloride- and carbonate- containing AFm phases, can take up trace levels of iodide. The preferential uptake mechanisms are figured out by co-precipitation experiments, indicating the formation of a solid solution between AFm-I2 and AFm-SO4 for the I-SO4 mixtures (Figure 1.4) [39].
Besides, AFm phase also shows a good affinity for selenium oxyanions, especially for selenate, which is immobilized by anion exchange with intercalate sulfate. On the other hand, sorption mechanisms of selenite are dominated by surface adsorption and Ca co-precipitation, at low and high loadings, respectively [58, 59]. Non-specific interaction of selenium oxyanions with the surface of the cement minerals, and no direct bonding to neighboring atoms can be detected by X-ray absorption spectroscopy[60]. In addition, aqueous arsenic(V) can also be removed efficiently by monosulfate via anion exchange, causing the transformation of monosulfate to ettringite [61].
On ettringite (AFt) phase
Ettringite, with a general formula of [Ca3(Al,Fe)(OH)6 12H2O]X3 H2O where X denotes a double-charged anion (typically sulfate) or two units of a single-charged anion, occurs in natural alkaline environments, associated with other phases like portlandite, gypsum or afwillite. However, a non-natural analogue of ettringite also occurs during the early hydration of Portland cements in the presence of gypsum[62]. Portland and modified Portland cements are widely used in nuclear waste repositories, thus ettringite is assumed to play a key role in the RNs immobilization processes. The unit cell of ettringite consists of a column part ([Ca6[Al(OH)6]2•24H2O]6+) with the inter-column spaces occupied by 3 units of SO42- ions and 2 units of H2O molecules, which hold the columns together through electrostatic forces[63].
Ettringite is an anion exchanger and it has been shown that partial or full replacement of SO42-(Figure 1.5) is possible in the case of SeO42-/SeO32- [58, 60, 64, 65], MoO42-[66], and Sb(OH)6-[67]. Anionic contaminant removal mechanisms can take place in two different ways: i) adsorption in the interchannels of already formed ettringite structure; ii) structural incorporation into ettringite lattice[68, 69]. Furthermore, ettringite particle surfaces exhibit a net negative charge, so anion substitution with trivalent and divalent cations in the bulk[70] is probably more important than surface adsorption[71]. It is proved that the incorporation mechanism of SeO42- in ettringite depends on coexisting SO42- in aqueous environments. Without SO42-, SeO42–substituted AFm phase was formed as an intermediate [65]. Anion exchange with structural SO42- is the main mechanism for immobilization of SeO42-. In contrast, SeO32- is easily immobilized to form inner-sphere complexes in ettringite[64]. Besides, a consistent observation is that uptake of MoO42- is low or non-existent which can be attributed to its large size in comparison to sulfate[66, 72]. Ettringite showed an anion preference in the order of B(OH)4- >SeO42- > CrO42- > MoO42-[66]. Conclusively, two main factors, size and electronegativity difference of oxyanions compared to sulfate, are thought to be inversely proportional to the extent of solid solution [71].
On hydrogarnet, hydrotalcite, gypsum, and calcite.
Regarding other phases existing in hydrated cement paste, such as hydrogarnet, hydrotalcite, gypsum, and calcite, they exhibit lower sorption potential than other cement phases. Sorption on hydrogarnet (3CaO•Al2O3•6H2O) may occur through ion substitution, e.g., Cr(III) may substitute Al-sites [73]. Besides, hydrogarnet has been identified as a major phase in chromium ore processing residue and has a capacity to host Cr(VI)[74]. The maximum amount of chromate corresponds to a replacement of about one out of every eight hydroxyl tetrahedral per unit cell by a CrO42- tetrahedron. Hydrotalcite (Mg6Al2(CO3)(OH)16•4(H2O)) has a positively charged layer structure and can incorporate di- and trivalent cations. Previous reports also show hydrotalcite-like minerals’ affinity for oxyanions, e.g., SeO32-/SeO42-[75-77], MoO42-[75, 78], and AsO3-, facilitated by the intercalation of guest anions in most cases. As(V) undergoes strong chemical binding with hydrotalcite-Fe(III) unlike Mo(VI) and Se(VI). In addition, As(V) has the strongest affinity for hydrotalcite varying in the order As(V) > Mo(VI) > Se(VI) and selenium oxyanions have heavier partial intercalation than As(V) and Mo(VI)[75]. Generally, an interlayer distance change can be observed once the intercalation occurs[78]. However, high concentrations of soluble carbonate prevent anion exchange and it has been predicted that hydrotalcite is quantitatively less important than ettringite and monosulfate in cement [79]. In hydrated OPC, gypsum and calcite phases are seldom considered as possible sinks for oxyanions in nuclear waste, although they are abundant in HCP. Calcite is highly crystalline and has a much smaller reactive surface area and a high symmetry structure. In the case of co-precipitation, Se(IV) substituting for carbonate sites in the calcite structure is favored[80], leading to the formation of a Ca(SeO3)X(CO3)(1−X) solid solution[81]. Conversely, Se(VI) does not incorporate, but catalyzes the step propagation rate of calcite without changing the growth pattern [82]. The co-precipitation can be viewed as a sequence of adsorption and entrapment events [81]. Oxyanions, like AsO33- and SeO32-, possessing a lone electron pair and a carbonate-like pyramidal trigonal (coordination number is three) shape, are more favored to be sorbed by calcite, than other oxyanions, such as SeO42- and MoO42-[71]. Conversely, SO42- tetrahedron sites in gypsum prefer tetrahedral SeO42- substitution, forming a range of Ca(SO4,SeO4)•2H2O solid solutions [83]. However, solid solution formation with gypsum will be limited for molybdate and antimonite[71].
Table of contents :
Chapter 1. Introduction-Barriers around geologic nuclear waste repositories
1.1. Geological nuclear waste disposal
1.2. Multi-barrier systems
1.3. Sorption behavior of radionuclides in potential barriers
1.3.1. On iron phases existing in canisters and steel reinforcements
1.3.2. On hydrated cement phases existing in backfills and tunnels
1.3.3. On clays
1.3.4. On granitic minerals
1.4. Objectives of the thesis
Chapter 2. Evidence of multiple sorption modes in AFm phases using Mo as structural probe
2.1. Introduction
2.2. Materials and methods
2.2.1. Materials and Chemicals.
2.2.2. Adsorption Experiments.
2.2.3. Aqueous Data Modelling.
2.2.4. In-situ X-ray Diffraction.
2.2.5. Solid Phase Characterization.
2.3. Results and discussion
2.3.1. Wet Chemistry Experiments.
2.3.2. In-situ Time Resolved X-ray Diffraction.
2.3.3. Local Geometry of Adsorbed Molybdate.
2.3.4. Mo Distribution in Reacted LDHs Particles.
2.3.5. Complexation environment around adsorbed Mo.
2.3.6. Estimation of edge sites density.
2.3.7. Environmental implications.
Chapter 3. Selenite uptake by AFm phases: a description of intercalated anion coordination geometries
3.1. Introduction
3.2. Materials and methods
3.2.1. Materials and chemicals.
3.2.2. Adsorption experiments.
3.2.3 Aqueous data modelling.
3.2.4 In-situ time-resolved XRD.
3.2.5 Solid phase characterization.
3.3. Results and discussion
3.3.1 Batch sorption isotherm.
3.3.2. In-situ time resolved XRD.
3.3.3. Sorbed Se coordination.
3.3.4. Sulfur K-edge EXAFS results.
3.3.5. Linear relationship between basal spacing and hydrated anions’ radii.
3.3.6. Structural stability of AFm-SeO3.
3.3.7. Environmental Implications.
Chapter 4. Determination of redox potentials imposed by steel corrosion products in cement-based media
4.1. Introduction
4.2. Materials and methods
4.2.1. Materials and chemicals.
4.2.2. Synthesis and characterization of Fe-bearing phases.
4.2.3. Batch sorption experiments.
4.2.4. XANES spectroscopy.
4.2.5. PDF spectroscopy.
4.3. Results and discussion
4.3.1. Aqueous phase results.
4.3.2. Surface Se and Sb species.
4.3.3. Surface U and Mo species.
4.3.4. PDF analysis of reacted NZVI.
4.3.5. “In-situ” experimental Eh values.
4.3.6. Environmental implications.
Chapter 5. RNs (i.e., U, Se, Mo, and Sb) sorption behavior on hydrated Fe-bearing CEM-V/A cement
5.1. Introduction
5.2. Materials and methods
5.2.1. Materials and chemicals.
5.2.2. Preparation of powder samples of hydrated Fe-bearing cement products.
5.2.3. Synthesis of each separated CEM-V/A cement hydration product.
5.2.4. Kinetics experiments.
5.2.5. Batch sorption experiments.
5.2.6. Preparation of RNs-containing Fe0-bearing cement cores.
5.2.7. Polishing of RNs-containing cement cores and Micro-probe analysis.
5.2.8. Sulfur K-edge, Selenium K-edge, and Molybdenum K-edge XANES-EXAFS spectroscopy.
5.2.9. PDF analysis.
5.3. Results and discussion
5.3.1. XRD characterization and dissolution kinetics of Fe-bearing hydrated cement.
5.3.2. U(VI) sorption on hydrated CEM-V/A cement particles.
5.3.3. Se(IV) sorption on hydrated CEM-V/A cement particles.
5.3.4. Mo(VI) sorption on hydrated CEM-V/A cement particles.
5.3.5. Sb(V) sorption on hydrated CEM-V/A cement particles.
5.3.6. Microprobe mapping on Fe0/cement interface in presence of RNs.
5.4. Conclusions
Chapter 6. The influence of surface impurities on the reactivity of pyrite toward aqueous U(VI)
6.1. Introduction
6.2. Materials and methods
6.2.1 Materials
6.2.2 Batch sorption experiments
6.2.3 Spectroscopic and Microscopic Analyses
6.3. Results and discussion
6.3.1 Pyrite Characterization
6.3.2 Aqueous phase analysis for Type I pyrite suspension
6.3.3 Aqueous phase analysis for Type II pyrite suspension
6.3.4. XPS Surface features
6.3.5. XAS Characterization on uranium speciation
6.3.6. SEM surface morphology
6.3.7. Influencing factors on pyrite reactivity.
6.4. Conclusions
