Immobilisation of Radioactive Wastes in Glass

CHAPTER 19

Immobilisation of Radioactive Wastes in Glass Contents 19.1 The Vitreous State 19.2 Glasses for Nuclear Waste Immobilisation 19.3 Immobilisation Mechanisms 19.4 Borosilicate Glasses 19.5 Cations in Silicate Glasses 19.6 Degree of Polymerisation 19.7 Role of Boron Oxide 19.8 Role of Intermediates and Modifiers 19.9 Difficult Elements 19.10 Selection Rules for a Nuclear Wasteform Silicate Glass 19.11 Phosphate Glasses 19.12 Glass Composite Materials 19.13 Vitrification Technology 19.14 Development of Vitrification Technologies 19.15 Calcination Processes 19.16 Cold Crucible Melters 19.17 In Situ Vitrification 19.18 Radionuclide Volatility 19.19 Wasteform Acceptance Criteria References Further Reading

319 323 327 328 331 332 334 335 336 338 339 341 342 346 356 360 363 365 366 367 368

19.1 THE VITREOUS STATE Glassy and vitreous are synonyms, vitreous being derived from the Latin for glass. Glass is an amorphous solid below the glass transition temperature (Tg). Glass transition is a second order-like phase transformation occurring at a Tg that depends on heating/cooling rates. A material is amorphous when it has no long-range order, with no regularity in the arrangement of its molecular constituents on a scale larger than a few

An Introduction to Nuclear Waste Immobilisation DOI: https://doi.org/10.1016/B978-0-08-102702-8.00019-4

© 2019 Elsevier Ltd. All rights reserved.

319

320

An Introduction to Nuclear Waste Immobilisation

times the size of these groups. For example, the average distance between silicon atoms in vitreous silica (SiO2) is B3.6 Å, and there is no order between these atoms at distances above B10 Å. A solid is defined as a material having a definite shape and volume that is neither liquid nor gaseous; for example a material with a degree of connectivity between its molecular constituents which ensures that the geometry of its connecting bonds is three-dimensional (3D). Amorphous materials are either glassy (below Tg) or molten (above Tg). Mechanical behaviour of glasses is drastically different from that of melts; that is glasses are brittle whereas melts are ductile, and the rheological difference is manifested in a continuous change of viscosity rather than any step-wise difference. There are, however, step-wise changes at Tg of thermodynamic parameters such as heat capacity and thermal expansion coefficient that reveal a second order-like phase transformation at Tg. Structurally glasses differ from melts although the difference is very difficult to reveal for atomic or molecular distribution. However for the system of bonds the difference becomes obvious
glasses contain broken bonds (termed configurons) only as point defects or small clusters whereas liquids contain extended structures of macroscopically large clusters termed percolation clusters (Ojovan & Lee, 2006). The difference is illustrated by Table 19.1. Glasses can be formed by several methods: melt quenching, physical vapour deposition, solid state reactions via thermochemical or mechanochemical methods, liquid state reactions (e.g. sol
gel techniques) and under the action of high pressure (pressure amorphisation). Irradiation of crystalline solids can also result in formation of amorphous solids. Glasses, however, are most frequently produced by melt cooling below their Tg fast enough to avoid formation of crystalline phases (Varshneya, 2006). Because of this the International Commission on Glass defines glass as a uniform amorphous solid material, usually produced when the viscous molten material cools very rapidly to below its Tg, without sufficient time for a regular crystal lattice to form. Glass-forming materials such as dioxides do not require fast cooling whereas quickly crystallising materials such as metals require very rapid cooling (quenching), for example the early metallic glasses had to be cooled extremely rapidly B106K/s to avoid crystallisation (Fig. 19.1). Below Tg amorphous materials have a 3D bond geometry as found in crystals and hence solid-like behaviour. Above Tg fractal structures are formed by broken bonds due to thermal fluctuations. Although Tg

Table 19.1 Structural features of amorphous materials; configurons (broken bonds) are depicted as small red dots Glass (T , Tg)

Liquid (T . Tg)

Broken bonds (configurons) in the form of point defects The concentration of configurons increases with temperature

Clusters made of configurons (broken bonds) form a percolation cluster that penetrates the whole body of material exactly at Tg Macroscopic clusters made of configurons form large structures growing in size and volume with increase of temperature

There is only a small number of clusters made of configurons

322

An Introduction to Nuclear Waste Immobilisation

Metallic glasses

Critical cooling rate, K/s

106

104 Bulk metallic glasses 102

100 Silicate glasses

10–2

0.3

0.5

0.7

0.9

Trg

Figure 19.1 Critical cooling rates above which melts transform to glasses. Trg 5 Tg/ Tm, where Tm is the melting temperature.

depends on the rate of cooling, it can be roughly assessed from Kauzmann’s relation: 2 Tg  TL 3

(19.1)

where TL is the liquidus temperature at which a phase diagram shows a crystal-free melt. A higher TL provides a higher Tg, however high processing temperatures are not acceptable for an efficient waste immobilisation process due largely to volatilisation of some key radionuclides such as Cs and Ru. A much more exact method to calculate Tg is to consider melting as a percolation via broken covalent bonds (termed configurons) with Tg dependent on quasi-equilibrium thermodynamic parameters of the bonds, for example on enthalpy (Hd) and entropy (Sd) of formation of bonds (Ojovan & Lee, 2006): Hd  Tg 5
Sd 1 R ln½ð1 2 θc Þ=θc 

(19.2)

where R is the universal gas constant and θc is the percolation threshold in a system of broken bonds; for example θc 5 0:15 for vitreous silica. Although actual values of Hd and Sd depend on the cooling rate they can be found from available experimental data on viscosity of glasses and melts.

Immobilisation of Radioactive Wastes in Glass

323

The viscosity of melts and glasses is given by the universal equation (Ojovan, 2012)       B D η 5 A1 T 1 1 A2 exp 1 1 Cexp (19.3) RT RT where T is temperature and coefficients A1, A2, B, C and D are directly related to the parameters of broken bonds
enthalpies (Hd, Hm) and entropies (Sd, Sm) of formation (designation, d) and motion (designation, m). The viscosity of melts at high temperatures is typically characterised by a low activation energy of flow η B exp(QL/RT) with QL 5 B 5 Hm B 80
300 kJ/mol whereas that of glasses is characterised by a high activation energy of flow η B exp(QH/RT) with QH 5 (D 1 B) 5 (Hd 1 Hm) B 400
800 kJ/mol. Although, compared to crystalline materials of the same composition, glasses are metastable materials, their relaxation to a thermodynamically stable crystalline structure is kinetically impeded. The metastability of silicate glasses commonly used by various industries is not a practical concern as most oxide glasses are stable for times much longer than any imaginable timescale of our universe. In practice, there is no stress relaxation at room temperatures so that high permanent internal stress is preserved in glass articles made more than several millennia ago. Relaxation processes are controlled by viscosity with characteristic relaxation time required to attain stabilised parameters (Maxwell’s relaxation time) given by: η τM 5 (19.4) G where G is the shear modulus and η is the viscosity. The higher the viscosity, the longer the relaxation time. Viscosity change is thermally activated and glass-forming amorphous oxides are characterised by high activation energies and very high viscosities under normal conditions. For example, fused silica has an activation energy of viscosity at low temperature of QH 5 759 kJ/mol, shear modulus 31 GPa giving relaxation times at room temperature τ M as long as τ M 5 1098 years, incommensurably longer than even the lifetime of the universe, which is about 14 3 109 years.

19.2 GLASSES FOR NUCLEAR WASTE IMMOBILISATION Natural glass has existed since the beginning of time, formed when certain types of rock melted as a result of high-temperature phenomena such as

324

An Introduction to Nuclear Waste Immobilisation

volcanic eruptions, lightning strikes or the impact of meteorites, and then cooled and solidified rapidly. Natural glasses have been used by man from the earliest times, for which there is archaeological evidence; for example, stone-age man is believed to have used cutting tools made of obsidian, a natural glass of volcanic origin. Some of these glasses have been in the natural environment for about 300 million years with low alteration rates of only tenths of a millimetre per million years. Glass and glazes were manufactured far back in human history. Glazed stone beads have been dated to 12,000 BC, and pure glass or glaze has been dated to 7000 BC. Evidence exists of industrial glass works in Egypt around 1500 BC. A new application of glass was discovered in the 20th century (Lutze, 1988; Vashman et al., 1997; Navrotsky, 1998; Donald, 2010; Pegg & Joseph, 2001; Plodinec, 2000; Sobolev, Ojovan, Scherbatova, & Batyukhnova, 1999), as a nuclear waste immobilising material produced via vitrification technology (Fig. 19.2).

Figure 19.2 (A) A stainless steel canister for vitrified radioactive waste (European design). (B) This volume of simulant borosilicate glass is sufficient to hold all the high level waste arising from an 80-year lifetime’s worth of nuclear-generated electricity for a single person. Picture courtesy of BNFL.

Immobilisation of Radioactive Wastes in Glass

325

Michael Faraday described glass as a solution of different substances, one in another; a description still appropriate for characterisation of a multi-component glass. Glasses as solid state solutions are tolerant to compositional changes; their properties change continuously and smoothly (in most cases linearly) with variation of composition. Because of this, vitrification is almost insensitive to slight compositional variations typical of most wastes. The physical and chemical durability of glasses, combined with their high capability to incorporate most elements into their structure make them irreplaceable when highly toxic wastes such as long-lived and highly radioactive wastes need reliable immobilisation for safe long-term storage, transportation and consequent disposal. Waste vitrification is attractive because of the • • • • •

High capability of glass to reliably immobilise a wide range of elements Simple production technology adapted from glass manufacture Small volume of the resulting glassy wasteform High chemical durability of glassy wasteforms in contact with natural waters High tolerance of glasses to radiation damage

Two main glass types are currently used for nuclear waste immobilisation: borosilicate and phosphate (Table 19.2). The exact compositions of nuclear waste glasses are tailored for easy preparation and melting, avoidance of phase separation and uncontrolled crystallisation, and acceptable chemical durability such as leaching resistance. High waste loadings and high chemical durability can be achieved in both borosilicate and aluminophosphate glasses. Moreover, such glasses immobilise well large quantities of actinides; for example, borosilicate glasses can accommodate up to 7.2, mass% of PuO2. Phosphate glasses can accommodate large amounts of aluminium oxides; however, in contrast to borosilicate melts, molten phosphate glasses are highly corrosive to refractory linings, behaviour which has limited their application. Aluminaphosphate glass has been used in Russia to immobilise high level waste (HLW) from nuclear fuel reprocessing since 1987. Recently, borosilicate glasses have been developed to host Hanford high-Al radioactive waste (Savannah River National Laboratory glass, Table 19.2). The most important parameters of nuclear waste glasses are radionuclide normalised leaching rates NRi (g/cm2 day), mechanical strength (MPa), density ρ (g/cm3), thermal expansion coefficient κ (1/K), specific heat Cp

326

An Introduction to Nuclear Waste Immobilisation

Table 19.2 Compositions of nuclear waste glasses Glass/country Oxide, wt.%

SiO2

P2O5

B2O3

Al2O3 CaO MgO Na2O Misc

R7/T7, France DWPF, USA SRNL, USA WVP, UK Pamela, GermanyBelgium Mayak, Russia Radon K-26, Russia P0798, Japan GC-12/9B, China

47.2 49.8 30.5 47.2 52.7



1.1



14.9 8.0 15.2 16.9 13.2

4.4 4.0 25.0 4.8 2.7

4.1 1.0 6.1
4.6


43

52.0



6.6

19.0 3.0



13.7


46.6 46.2





14.2 13.4

5.0 4.2


1.4 0.1 5.3 2.2

3.0
2.5 1.5

10.6 8.7 9.6 8.4 5.9

18.8 27.1 13.5 19.4 18.7

21.2 23.9

7.8 9.8

Waste oxides

28 33 45c 25
38d 30 33a 35b

10.0 20.2 9.1 23.1

DWPF, Defence Waste Processing Facility, Savannah River Site, US; SRNL, Savannah River National Laboratory, US; WVP, Waste Vitrification Plant, Sellafield, UK. a # 10 For fission products and minor actinide oxides. b This glass is designed for sodium-containing LLW and ILW. c This glass has been developed to host Hanford high-Al radioactive waste. d Gribble et al. (2009).

(J/kg K) and thermal conductivity λ (W/m K). Important glass processing parameters are melting temperature Tm, viscosity η (Pa s) and electrical conductivity σ (1/Ω cm) near the melting temperature. Vitrification can be performed efficiently at Tm values below 1200°C so avoiding excess radionuclide volatilisation and maintaining viscosities below 10 Pa s to ensure high throughput and controlled pouring into canisters. Some typical properties of HLW borosilicate and phosphate glasses are shown in Table 19.3. The leaching resistance of nuclear waste glasses is a paramount criterion as it ensures low release rates for radionuclides on any potential contact with water. Vitrified radioactive waste is a chemically durable material which reliably retains active species. Typical normalised leaching rates NR (see Section 14.5) of vitrified wasteforms are below 1025
1026 g/cm2 day. Moreover, as glasses and glass-composite materials (GCM) are highly corrosion resistant, their high nuclide retention is expected to last for many millennia. The excellent durability of vitrified radioactive waste ensures a high degree of environmental protection. Vitrification has been used for nuclear waste immobilisation for more than 50 years in France, Germany, Belgium, Russia, the United Kingdom, Japan and the United States. The total production of all vitrification plants was reported as .22,600 t (Jantzen, 2011a, 2011b).

Immobilisation of Radioactive Wastes in Glass

327

Table 19.3 Typical properties of glasses for nuclear waste immobilisation Glass type

Density, Compressive NR,a 1026 g/cm2day g/cm3 strength, MPa

Borosilicate 2.7

22
54

Phosphate 2.6

9
14

0.3 0.2 1.1 0.4

(Cs) (Sr) (Cs) (Sr)

TEC,b 1/K

Tmax,c K (°C)

8 3 1026

$ 823 (550) . 109

Damaging dose,d Gy

1.5 3 1026 $ 723 (450) . 109

IAEA, International Atomic Energy Agency. a IAEA test protocol for 28th day. b TEC, Thermal expansion coefficient. c Tmax is the maximum allowed temperature of glass representing the limit of its thermal stability. Tmax is defined as the temperature above which the radionuclide NRs increase .102 times. By definition Tmax , Tg. d Irradiation has a small impact on glasses and the damaging dose is the absorbed dose above which the radionuclide NRs increase several times.

19.3 IMMOBILISATION MECHANISMS Vitrification involves melting of waste materials with glass-forming additives so that the final vitreous product incorporates the waste contaminants in its macro- and microstructure. Nuclear waste glasses are not completely homogeneous vitreous materials but contain significant numbers of bubbles, foreign inclusions such as refractory oxides and other immiscible components (Fig. 19.3A). Hazardous waste constituents are immobilised either by direct incorporation into the glass structure or by encapsulation. In the first case, waste constituents are dissolved in the glass melt, some being included into the glass network on cooling while others are confined as modifiers. Encapsulation is applied to elements and compounds which have low solubility in the glass melt and do not fit into the glass structure either as network formers or modifiers (Fig. 19.3B). The solubility limits of elements as oxides in silicate glasses are given in Table 19.4. The ability of species to be incorporated in sodium aluminoborosilicate type glasses depends very much on their oxidation state (Fig. 19.4, after Gin, Jollivet, Tribet, Peuget, & Schuller, 2017). Immiscible constituents which do not mix easily into the molten silicate glass are typically sulphates, chlorides and molybdates as well as noble metals such as Rh and Pd, refractory oxides with high liquidus temperatures such as PuO2, noble metal oxides and spinels. Encapsulation is carried out either by deliberate dispersion of insoluble compounds into

328

An Introduction to Nuclear Waste Immobilisation

Figure 19.3 (A) A secondary electron scanning electron microscopy image of a British Magnox nuclear waste glass. (B) Schematic of waste retention in a glass matrix. Left, incorporation in a relatively homogeneous glass, with some bubbles and inclusions. Right, encapsulation of waste particles in a glass matrix.

Table 19.4 Approximate solubility limits of elements in silicate glasses Element

Solubility limit, wt.%

Al, Si, P, Pb Li, B, Na, Mg, K, Ca, Fe, Zn, Rb, Sr, Cs, Ba, Fr, Ra, U Ti, Cu, F, La, Ce, Pr, Nd, Gd, Th, Bi, Zr Mn, Cr, Co, Ni, Mo C, S, Cl, As, Se, Tc, Sn, Sb, Te H, He, N, Ne, Ar, Br, Kr, Ru, Rh, Pd, Ag, I, Xe, Pt, Au, Hg, Rn

25 15
25 5
15 3
5 1
3 Less than 0.1

the glass melt, immiscible phase separation on cooling or by sintering of glass and waste powders so that the wasteform produced is a GCM (Ojovan & Lee, 2007).

19.4 BOROSILICATE GLASSES Borosilicate glasses are the first choice of material worldwide for immobilising both HLW and low and intermediate level waste (LILW) (Plodinec, 2000). This selection is based on the flexibility of borosilicate glass with regard to waste loading and the ability to incorporate many different kinds of waste elements, coupled with good glass-forming ability, chemical durability, mechanical integrity, and excellent thermal and radiation stability. The major component of borosilicate glasses is SiO2 with relatively high B2O3, CaO, MgO, Na2O and Al2O3 contents and minor amounts

Immobilisation of Radioactive Wastes in Glass

329

Figure 19.4 Solubility of species in sodium aluminoborosilicates as function of oxidation state. Courtesy Sophie Schuller, CEA, France.

of many other oxides. SiO2, B2O3 and Al2O3 are generally network formers because they form strong covalent bonds involving SiO4, AlO4 and BO4 tetrahedra and BO3 triangles. Silicon is the main glass-forming element in a borosilicate waste glass and its basic elements are SiO4 tetrahedra, which comprise bridging or cross-linking and non-bridging atoms of oxygen (NBO). In a silicate glass the SiO4-tetrahedra vertices connect these elements to each other through bridging oxygen atoms so that the network consists of chains of various lengths. The glass network is not regular as in the case of crystalline silica; for example, the bond angle Si
O
Si can range from 120° to 180° while in quartz it is a constant. However, the Si
O bond length remains constant (1.62 Å) as well as the O
Si
O bond angle (109°280 ). Alkali, alkaline earth ions, transition metals and ions of high charge and large size including actinides, cannot readily substitute for Si, B or Al and so are network modifiers entering the gaps in the network structure. They generally have coordination numbers of 6 and higher, form weaker bonds to oxygen than the network formers and act to charge-balance the negatively charged borosilicate or alumina-borosilicate network. This leads to

330

An Introduction to Nuclear Waste Immobilisation

Figure 19.5 (A) Structure of an alkali-silicate glass; (B) a Q2 structural unit which comprises two bridging oxygen atoms and two non-bridging oxygens.

break up of Si
O
Si bonds producing NBO; for example SiO2 ions localised to modifying ions. Fig. 19.5 illustrates the structure of an alkalisilicate glass. The structure is typically described using the Qn designation, where n represents the number of bridging oxygens per tetrahedral unit. Both glasses and melts possess short-range order (SRO) with a typical radius of several Angstroms. SRO structural groups in commercial glasses are usually tetrahedral Si, B, Al, Fe, P surrounded by four oxygen atoms (tetrahedral coordination) or B surrounded by three oxygen atoms (trigonal coordination). Glasses are typically named after their predominant tetrahedral species such as borosilicate glasses which are primarily B and Si. The tetrahedra and trigonal species in the glass link to each other via bridging oxygen bonds. The remaining NBO atoms effectively carry a negative charge and ionically bond positively charged cations such as Na1 or Ca12. The atomic structure of oxide glasses is most exactly represented by Greaves’ modified random network (MRN) model (Fig. 19.6). The MRN has two interlacing disordered sublattices: one is the network region and another consists of regions comprised of large concentrations of atoms which do not enter in the network (e.g. network modifiers). These may form percolating channels at higher concentrations

Immobilisation of Radioactive Wastes in Glass

331

Figure 19.6 Schematic of two percolation channels formed by monovalent alkaline cations (small circles) in silicate glasses.

of network modifiers. The tetrahedra define the network regions, while NBO define depolymerised regions that can form percolation channels. Percolation channels are defined by the NBO atoms at the edges of the highly ordered network regions, which ionically bond to the alkali, alkaline earths or other modifier species in a glass. Moreover, these channels can be revealed as they act as ion-exchange paths for elements that are less well bonded to the NBO. It has been also found that for small length scales the alkali pathways are fractal in structure with Hausdorff dimensionality Df in the range from 1.5 to 2.0 whereas on macroscopic scales the Df rapidly increases to 3D. This structural feature of oxide glasses explains the well-known mixed alkali effect in glasses caused by effective blocking by immobile unlike cations due to low dimensionality (,3) pathways on local length scales.

19.5 CATIONS IN SILICATE GLASSES The ability of cations to enter into the glass network structure is characterised by their field strength, which is defined as F5

Z a2

(19.5)

332

An Introduction to Nuclear Waste Immobilisation

Table 19.5 Function of cations in a silicate oxide glass structure Element Valence, Z

Ionic distance for oxides, Å

Coordination number

Field strength, 1/Å2

Bond strength, kJ/mol

Function

Si B

4 3 4 5

1.60 1.50

4 3 4 4

1.57 1.63 1.34 2.1

443 498 372 368
464

Network formers: FB1.5
2.0

4 4 3 3 3 3 2 4 4 2 2

1.96 1.96 1.77 1.89 1.88 1.99 1.53

4 6 4 6 4 6 4 6 8 4 6

1.25 1.04 0.96 0.84 0.85 0.76 0.86 0.84 0.77 0.53 0.45

455 304 335
423 224
284

2 2 2 2 2 1 1 1 1

6

0.43 6 8 8 8 6 6 8 12

0.34 0.27 0.33 0.28 0.23 0.19 0.13 0.10

P Ti Al Fe Be Zr Mg Fe Pb Ca Sr Li Na K Cs

1.55

2.28 2.03 2.10

2.74 2.48 2.69 2.10 2.30 2.77

Intermediates: FB0.5
1.0 263 338 255 155 310 151 134 134 151 84 54 42

Network modifiers: FB0.1
0.4

where Z is the valence and a is the ionic radius (Å) in the oxide. Lower field strength ions (e.g. alkalis) are network modifiers, whereas ions with higher field strength (such as Si, P or B) are network formers. Table 19.5 lists cations according to their function in the glass structure.

19.6 DEGREE OF POLYMERISATION The internal structure of glasses made of interconnected structural blocks is characterised by the degree of polymerisation: the higher the degree of polymerisation the more developed its 3D network. The degree of polymerisation of a glass fT is given by the ratio of the number of networkforming cations to the number of oxygen atoms and can be expressed as P γT fT 5 (19.6) γM2 O 1 γMO 1 3γM2 O3 1 2γ MO2 1 5γ M2 O5 1 3γMO3

Immobilisation of Radioactive Wastes in Glass

333

Table 19.6 Structure of silicate glasses at different polymerisation degrees Glass

Structure

Ratio fSi 5 [Si]/[O]

SiO2

Three-dimensional network (silica)

0.500

Me2O  2SiO2 MeO  2SiO2

Two-dimensional network (micas)

0.400

Me2O  SiO2

One-dimensional network (metasilicates): 0.333 chains, trinomial or six-member coils

MeO  SiO2 [Si2O7]2

Discrete double tetrahedra (pyrosilicates)

0.286

2Me2O  SiO2 2MeO  SiO2

Discrete tetrahedra (orthosilicates)

0.250

where γ x is the molar fraction of a given oxide (x) in the glass and γ T is the molar fraction of network-forming oxides. Energetically the most important depolymerisation process can be viewed as an acid-base reaction: 2Mn O 1 SiO2 5 2nMð2=nÞ1 1 SiO42 4

(19.7)

or in terms of elementary Qn units: Q4 1 2O22 5 Q0

(19.8)

where M is the modifying cation and n is its valence state. In practice adding a network modifier such as Na2O progressively depolymerises the network; intermediate stages of polymerisation include Q3 sheets, Q2 chains and Q1 dimers such as Si2O7, along with fully linked SiO4-tetrahedra Q4. Often the polymerisation of silicate glasses is expressed on the basis of the ratio fSi 5 [Si]/[O], where [Si] and [O] are the numbers of silicon and oxygen atoms in the glass. Table 19.6 gives typical values of the degree of polymerisation of silicate glasses. The degree of polymerisation reaches a maximum at fSi 5 0.50 for vitreous silica dropping to 0.25 for modified glasses. Nuclear waste glasses are highly polymerised structures. For example, nuclear waste borosilicate glasses are made up of four groups: 1. Network-forming SiO2 2. Charge balancing network formers MTO2 (T 5 Al, B, Fe13, M 5 Alkali)

334

An Introduction to Nuclear Waste Immobilisation

3. Network modifiers MO, excess alkali not tied to T, as well as alkaline earth oxides 4. Lanthanide and actinide oxides Nuclear waste glasses typically consist of approximately 77.5 mol.% of fully polymerised SiO2 and MTO2, and 22.5 mol.% of partly depolymerised component.

19.7 ROLE OF BORON OXIDE Boron oxide is commonly substituted for silica in nuclear waste glasses. Introduction of B breaks up Q3 units and creates Q2, Q4 and also small amounts of Q1. The presence of B has several positive effects. B at less than 15 wt.% reduces the thermal expansion coefficient and improves chemical durability and resistance to mechanical abrasion. At relatively low temperatures (,500
600°C) boron stabilises the glass structure forming BO4 groups. At high temperatures B follows the general trend toward lower coordination number for cations, which may occur in more than one-coordination geometry. B then assumes trigonal plane coordination with three oxygens and becomes a network modifier. B reduces the viscosity of glasses at high temperature and raises it at low temperature and as a result the glass becomes ‘shorter’, which is beneficial for vitrification. A short glass is one with a sharp dependence of viscosity on temperature. A long glass having lower activation energy has smoother temperature dependence. Borate glasses are dominated by superstructure units, which comprise well-defined arrangements of the basic BO3 triangular and BO4 tetrahedral structural units with no internal degrees of freedom in the form of variable bond or torsion angles; for example boroxol (B3O6) and triborate (B3O7) units. On adding a network modifier (such as an alkali oxide) to boron oxide, the glass properties do not change as a monotonic function of composition but exhibit a maximum or minimum value at a specific composition
the so-called boron oxide anomaly. The addition of a network modifier to vitreous B2O3 initially leads to an increase in the coordination number of the boron atoms from 3 to 4, rather than to the formation of NBO atoms (i.e. the conversion of BO3 triangles into BO4 tetrahedra), and an important structural parameter is the fraction of the boron atoms which are four-coordinated.

Immobilisation of Radioactive Wastes in Glass

335

Boron in a three-coordinated position in a borosilicate glass is a relatively unstable component. Acids easily leach boron along with sodium bonded to it. Boron remains in a three-coordinated position in a sodium-borosilicate glass until the concentration of alkali metal is low according to ΨB 5

γNa2 O 1 , 3 γ B2 O3

(19.9)

where γNa2 O ½0; 1 and γB2 O3 are the molar fractions of sodium and boron oxides in the glass. Two structural networks coexist when ΨB , 1/3 and metastable phase separation (liquation) occurs in such a glass on heating, with formation of microscopic volumes of different composition: of high silica and sodium borate content. By acid etching the sodium borate phase from this glass, porous glass can be produced. BO4 tetrahedra occur for higher concentrations of Na2O when ΨB . 1/3. Ions of Na1 localise near these tetrahedra by forming [(BO4/2)Na1]2 complexes.

19.8 ROLE OF INTERMEDIATES AND MODIFIERS Cations with intermediate ionic strength such as Ti, Al, Be and Zr are intermediate elements. Al creates Q3 at the expense of Q4 and Q2 units in silicate glasses. Al ‘lengthens’ the glass; that is increases its working range, improves mechanical and chemical resistance and reduces the tendency for demixing. Al is a network former at low concentrations and occurs as AlO4 tetrahedra in the glass structure. This structural unit improves the glass stability and hence the chemical durability. Alkalis (e.g. Na) are located near AlO4 tetrahedra and balance their negative charge so that alkalis (Na) are no longer modifiers in the silicate network. Being strongly bonded to AlO4 tetrahedra, these alkali cations are not readily leached compared to alkalis more weakly bonded to NBO. This is true if the content of Al is relatively small, when the ratio γ Na2 O =γ Al2 O3 . 1. Al2O3 in glasses significantly reduces the diffusion coefficient of water, which also results in improved glass durability. Increasing the Al2O3 content has the same effect as decreasing the total alkali oxide content. However, adding too much Al2O3 may be deleterious for the processing efficiency as higher processing temperatures are required. In practice partial replacement of SiO2 by Al2O3 from 3% to 10% is considered to significantly improve the glasses water stability.

336

An Introduction to Nuclear Waste Immobilisation

Table 19.7 Structural states of Al and B in borosilicate glasses Ratio Ψ B

Structural states of B and Al

ΨB . 1 1/3 , Ψ B , 1 0 , Ψ B , 1/3

[AlO4] [AlO4] [AlO4]

[BO4] [BO4]



[BO3] [BO3]

The states of B and Al in borosilicate glasses as a function of Ψ B are given in Table 19.7. Ti is a network former in silicate glasses as the TiO4 tetrahedra form glass-structural units and, like Al, Ti increases the viscosity and stabilises the glass. Ti is unique among cations as it readily takes up four-, five- or sixfold coordination in glasses and crystals. Five-coordinated Ti is the dominant species in glasses rich in TiO2 at concentrations exceeding 16 wt.%. It behaves simultaneously as a network former and network modifier although dominant in the former role. Five-coordinated Ti is likely to bond to both NBO and bridging oxygens, acting as a new Q4 species with one additional NBO. Li is usually added to improve glass properties. Other modifiers, particularly Na, can be present in the waste. Na is a network modifier that tends to increase the number of NBO and apparently increases glass alterability by destabilising its structure. A similar adverse effect on the initial alteration rate is observed for all the alkali metals but over quite different composition ranges. CaO, MgO and ZnO increase the viscosity of alkali silicate glasses at low temperatures (400
600°C) but decrease it at high temperatures (1000
1300°C). The overall effect on viscosity is that a shorter glass results. Ca, Mg and Zn improve glass durability by stabilising the glass structure. Oxygen atoms bound to silicon form the coordination polyhedral cells around Ca, Mg and Zn. However, volatility is increased with addition of CaO as well as Na2O and B2O3, so that a technological compromise must be found, which depends on waste composition.

19.9 DIFFICULT ELEMENTS S, Cl and Mo are less glass-compatible elements. Sulphur in the form of Na2SO4, SO3 or SO2 can be incorporated in the borosilicate glass structure as up to 1 wt.% of SO3. Greater than 1% SO3 causes formation of a separate water-soluble (yellow) phase. The solubility of six-valent Mo is

Immobilisation of Radioactive Wastes in Glass

337

Figure 19.7 Yellow phase ‘peanut’ in a British Magnox waste simulant glass. Courtesy Rick Short, ISL, University of Sheffield.

Figure 19.8 Gas bubbles (A) and crystalline phases (B
F) in an as-cast British Magnox waste simulant glass. (B) RuO2, (C) Pd, (D) Cr, Fe, Ni-spinels, (E) lanthaniderich phase and (F) Zr-rich phase. Courtesy P.B. Rose, ISL, University of Sheffield.

also limited. During vitrification both Mo and S separate from the melt forming on its surface a so-called yellow phase (Fig. 19.7). This phase contains alkali sulphates, alkali chromates, alkali molybdates, CaMoO4 and Ba(Sr)CrO4. F can be retained in silicate melts by the addition of Ca to form CaF2, giving an opacified glass. RuO2, if present, is not soluble in the glass melt and is readily encapsulated in the form of separate particles (Fig. 19.8).

338

An Introduction to Nuclear Waste Immobilisation

Lanthanides and actinides in borosilicate glasses are microscopically immiscible, being inclined to nanoscale phase separation although the glasses remain macroscopically homogeneous.

19.10 SELECTION RULES FOR A NUCLEAR WASTEFORM SILICATE GLASS The most durable waste-immobilising glass should be a silica glass containing a minimum amount of waste oxides. However, such a wasteform is not used because of both the high processing temperature required and the low content of waste it can accommodate. In waste vitrification melting should be done in the range # 1100
1250°C since at higher temperatures excessive volatilisation of both radioactive and non-radioactive constituents occurs. Glass formers other than silica have to be added to lower the processing temperature, among them boron oxide. The glassformulation methodology typically defines a range of compositions around a reference formula to allow for slight fluctuations in the composition of the waste feed stream. Fig. 19.9 illustrates schematically the effect of additives on leaching rates; for example addition of Al2O3 improves the

Figure 19.9 Typical effects of oxides on leaching rates of silicate glasses. See http:// glassproperties.com/chemical_durability/.

Immobilisation of Radioactive Wastes in Glass

339

Table 19.8 Important ratios for borosilicate glass formulations Designation Molar ratio

Empirical optimum

A

Si/Al, B

B C D

Oxygen/network formers Network formers/network modifiers Network modifiers/oxygen not required for the network formers to be four-coordinated

Higher or about 1.5 2.2
2.4 About 2 Close to 2

leach resistance of borosilicate glasses whereas higher content of alkalis (Li2O, Na2O, K2O) reduces glass durability. As the cost saving incentive is to increase the waste loading in a wasteform the optimal glassy wasteform compositions are tailored as a compromise between waste loading and final glass durability accounting also for processing parameters on vitrification. Generally, no more than two parts by weight of glassmaking additives are required to convert one part of radioactive waste, as oxide, to a satisfactory glass. Empirical studies have provided guidelines for formulation of acceptable nuclear waste glasses (Table 19.8). When ratio A . 1.5, radionuclide leachability is low although the melting temperature of the glass increases as A increases. Glass formation is best and leachability low when 2.2 # B # 2.4. Excess oxygen is needed to balance the electrical charge of the network modifiers. Density increases as A decreases and B increases. Ratio C should be B2 for good glass formation. As C decreases, the glass network is disrupted and glass forming ability decreases.

19.11 PHOSPHATE GLASSES Phosphate glasses have been intensively studied in Russia, at the Eurochemic Corporation at Mol, Belgium, at Oak Ridge National Laboratory and the University of Missouri-Rolla in the United States. Russia has immobilised HLW from nuclear fuel reprocessing plant RT-1 in the Ural region in alumina-phosphate glass since 1987 (Vashman et al., 1997; Jantzen, 2011a). Molten phosphate glasses are highly corrosive to refractory linings, behaviour which has limited their application. Novel Fe
Pb
phosphate glasses are particularly attractive due to their ability to

340

An Introduction to Nuclear Waste Immobilisation

Figure 19.10 Structural units of phosphate glasses.

Figure 19.11 (A) Glass formation areas in the system Na2O
Al2O3
P2O5, compositions in wt.%. (B) Solubility (content) of Al2O3 as a function of (Na/P) ratio at constant temperatures.

accommodate enhanced amounts of refractory oxides and their high chemical durability. A number of Na
Al
phosphate, Fe
Al
phosphate and zinc phosphate compositions exhibit improved chemical durability. Fe
Pb
phosphate glasses which melt from 800 to 1000°C are not as corrosive as earlier phosphate compositions. Phosphate glass structure is built around PO4 tetrahedral units described using the Qn designation (Fig. 19.10). In a pure P2O5 system, the glass is a 3D network of branching Q3 units with three bridging oxygens and one doubly bonded oxygen per tetrahedral unit. Addition of modifying alkali or alkaline earth cations replaces Q3 units with Q2 units with the cations creating ionic cross-links between the phosphate units. At a P2O5 concentration of approximately 50 mol%, the Q3 units disappear and the structure consists of only Q2 units in the form of linear phosphate chains. Further addition of modifying cations at concentrations greater than 50 mol% begins to convert Q2 units to Q1 units and finally Q0 units. Phosphate glass is particularly attractive for immobilisation of high Al and Na wastes. Fig. 19.11 shows the glass-forming regions of the

Immobilisation of Radioactive Wastes in Glass

341

Table 19.9 Solubility of elements in the phosphate glass Element

Solubility limit, ppm

Ru Rh Pa Ag Te Zr Mo La Ce Nd Sm Fe Cr Ni

20
60 20
60 300
600 $ 2.6 3 103 $ 103 $ 7 3 103 $ 7 3 103 (11
14) 3 103 (12
16) 3 103 (20
24) 3 103 (28
32) 3 103 5 3 103 500
2000 500
2000

Na2O
Al2O3
P2O5 system (Poluektov, Schmidt, Kascheev, & Ojovan, 2017). For low-to-moderate melting temperatures, the optimum range of the Na to P ratio is from 1.0 to 1.3. This ratio can be increased at higher temperatures, and at 1400
1500°C phosphate glasses can be made with up to 40% Al2O3. In contrast to borosilicate glasses phosphate glasses incorporate significantly larger amounts of corrosion products as well as actinide oxides, molybdates and sulphates. Lanthanides and actinides in phosphate glasses tend to complex strongly with phosphate ions. Table 19.9 gives data on solubility of some HLW components in melted phosphate glass at 1000°C. Some properties of alumina-phosphate glass composition (wt.%) 21.2Na2O
3Cs2O
19Al2O3
1.5Fe2O3
0.1Cr2O3
52P2O5
3SrO used to immobilise HLW in Russia are given in Table 19.2 (Section 19.3).

19.12 GLASS COMPOSITE MATERIALS GCMs are used to immobilise glass-immiscible waste components such as sulphates, chlorides, molybdates and refractory materials requiring unacceptably high melting temperatures. GCMs comprise both vitreous and crystalline components. Depending on the intended application, the major component may be a crystalline phase with a vitreous phase acting as a bonding agent, or alternatively, the vitreous phase may be the major component, with particles of a crystalline phase dispersed in the glass matrix. GCMs may be produced by dispersing both melted materials and fine

342

An Introduction to Nuclear Waste Immobilisation

Figure 19.12 Waste encapsulation in GCMs: Microstructure of (A) Synroc-glass with zirconolite crystalline phase; (B) GCM for immobilising yellow phase; (C) GCM for immobilising iodine. GCM, Glass-composite material. Courtesy Dan Perera, ANSTO.

crystalline particles in a glass melt and may be used to immobilise longlived radionuclides (such as actinide species) by incorporating them into the more durable crystalline phases, whereas the short-lived radionuclides may be accommodated in the less durable vitreous phase (Ojovan & Lee, 2007; Ojovan & Lee, 2011). Synroc-glass is an example of a GCM with Synroc crystalline phases in a vitreous matrix (Fig. 19.12A). Synroc-glass has been developed for sodium-rich wastes such as those at the Hanford site in the United States. Crystalline phases such as zirconolite and perovskite are the hosts for actinides, and waste loadings of 50%
70% by weight have been demonstrated in such composites with a high durability. The French have developed a U-Mo GCM to immobilise Mo-rich HLW. Another example is the GCM developed to immobilise sulphur-enriched waste streams in Russia (Fig. 19.12B) containing conventional borosilicate glass vitreous phase with uniformly distributed particles comprising up to 15% by volume of yellow phase. The durability of this GCM is similar to that of conventional wasteform glasses (Table 19.10). GCMs are potential host materials for highly volatile radionuclides such as 129I (Fig. 19.12C). Such GCMs can be produced by sintering an intimate mixture of glass powders and iodinecontaining sorbents, possibly under applied pressure.

19.13 VITRIFICATION TECHNOLOGY Vitrification technology comprises several stages, starting with evaporation of excess water from liquid radioactive waste, followed by batch preparation, calcination, glass melting, and ending with pouring and cooling of

Immobilisation of Radioactive Wastes in Glass

343

Table 19.10 Basic properties of borosilicate glasses and Russian yellow phase glasscomposite material (GCM) Parameter

Borosilicate glasses for high sodium waste

30
35

Viscosity, Pa s, at 1200°C Resistivity, Ω m, at 1200°C Density, g/cm3 Compressive strength, MPa 137 Cs 90 Sr Cr, Mn, Fe, Co, Ni Rare earth elements (REE), Actinides (An) Na B SO422

3.5
5.0 0.03
0.05 2.5
2.7 80
100 1025
1026 1026
1027 B1027
1028 B1028

Leach rate, g/cm day, (28-day International Atomic Energy Agency test)

Waste oxide content, wt.%

1025
1026 ,1028 B1025
1026

GCMs

30
35 1 up to 15 vol.% of yellow phase 3.0
6.0 0.03
0.05 2.4
2.7 50
80 B1025 1026
1027 1027
1028 B1028 1024
1025 # 1028 B1024
1025 with up to 15 vol.% yellow phase

vitrified waste blocks with some small amounts of secondary waste (IAEA, 1979; IAEA, 1992). Thin-film evaporators (Section 16.2.1) are used to evaporate the water. There are two types of nuclear waste glass preparation processes currently used (Fig. 19.13): • •

one-stage vitrification two-stage vitrification

In the one-stage vitrification process (Fig. 19.13A) glass forming additives are mixed with concentrated liquid wastes and so a glass-forming batch is formed (often as a paste). This batch is then fed into the melter where further water evaporation occurs, followed by calcination and glass melting, which both occur directly in the melter. In a two-stage vitrification process (Fig. 19.13B) the waste is calcined (Section 19.15) prior to melting. After calcination the required glass-forming additives (usually as a glass frit) together with the calcine are fed into the melter. Two types of melters are currently used at waste vitrification plants (WVPs): Joule heated ceramic melters (JHCM) and induction-heated

344

An Introduction to Nuclear Waste Immobilisation

Figure 19.13 Schematic of one-stage (A) and two-stage (B) vitrification processes.

Figure 19.14 (A) Joule heated ceramic melter used at Nuclear Research Centre, Karlsruhe, Germany for vitrification of highly active waste concentrate. (B) Cold crucible melting (1) and rectangular carbon steel containers (2) used at Moscow SIA ‘Radon’ for vitrification of low and intermediate level waste. Courtesy (A) S. Weisenburger, Forschungszentrum (FZK) Karlsruhe; (B) S.A. Dmitriev.

melters, which can either be hot (induction hot crucible) or cold (e.g. cold crucible melters, CCMs) (Fig. 19.14). Melting of nuclear waste glasses must be performed below 1200°C because of the volatility of the fission products, notably Cs and Ru, so

Immobilisation of Radioactive Wastes in Glass

345

avoiding excess radionuclide volatilisation and maintaining viscosities below 10 Pa s to ensure high throughput and controlled pouring into canisters. A more fluid glass is preferred to minimise blending problems, however higher fluidity is associated with higher carryover of volatile radionuclides (Cs, Ru, Tc). Phase separation on melting occurs in waste streams containing glass-immiscible constituents which can be immobilised as an isolated and phase separated disperse phase (i.e. in a GCM). Two streams come from the melter: 1. The glass melt containing most of the radioactivity 2. The off-gas flow, which contains off-gases and aerosols The melt waste glass is poured into containers (canisters) typically made of stainless steel when immobilising HLW or carbon steel for vitrified LILW. These may or may not be slowly cooled in an annealing furnace to avoid accumulation of mechanical stresses in the glass. When annealing is not used, some glass cracking occurs, resulting in a large surface area being potentially available for attack by water in a repository environment. Table 19.11 estimates the final surface area Sf of unannealed nuclear waste glass in cylindrical containers compared to the surface area of annealed samples, when the minimum final surface area is Sa. Despite the higher final surface areas of unannealed glasses these are sufficiently durable to ensure a suitable degree of radionuclide retention. Hence in many cases annealing is not applied in vitrification facilities. The second stream from the melter goes to the gas purification system, which is usually a complex system that removes from the off-gas not only radionuclide but also chemical contaminants. Operation of this purification system leads to generation of a small amount of secondary waste. For example, the distribution of beta gross activity at the Pamela WVP was (%): .99.88 in waste glass, and the rest in secondary waste; for example ,0.1 in intermediate level waste, ,0.01 in cold waste and ,0.01 in off-gas. Fig. 19.15 shows a typical off-gas purification system for a nuclear WVP. Table 19.11 Increase of surface area of nuclear waste glasses depending on cooling regime Cooling regime

Sf/Sa

Annealing applied Slow cooling Free air cooling Water quenching

1 4 12 22

346

An Introduction to Nuclear Waste Immobilisation

Figure 19.15 Technological flow-sheet diagram of LILW vitrification plant at Moscow SIA ‘Radon’: (1) LILW interim storage tank; (2) LILW transportation vehicle; (3) pump; 4, concentrate collector; (5) rotary film evaporator; (6, 13, 26) pumps; (7) condenser; (8) condensate collector; (9) batch (feed) mixer; (10) glass formers bins; (11) glass formers mixture bin; (12) screw feeder; (14) cold crucible; (15) bag filter; (16) HEPA filter; (17, 19, 23, 27) heat exchangers; (18) scrubber; (20) heater; (21) catalytic reactor for reduction of nitrogen oxides; (22) catalytic reactor for oxidation of ammonia; (24) fan; (25) sorbent bin; (26) pump; (28) ammonia balloon; (29) glass canister; (30) annealing furnace. HEPA filter, High efficiency particulate air filter; LILW, low and intermediate level waste. Courtesy S.V. Stefanovsky, Russian Academy of Sciences.

Containers of vitrified waste are the final product containing the overwhelming part of the waste contaminants. Comparing the volume of the glass blocks produced to the initial waste volume gives the volume reduction factor (VRF), which is one of the most important factors characterising the efficiency of the processing technology. VRF is crucially dependent on the waste composition and for a typical vitrification process is on the order of 5.

19.14 DEVELOPMENT OF VITRIFICATION TECHNOLOGIES Immobilisation of HLW by vitrification has been studied extensively over the last 50 years in Belgium, Canada, China, Denmark, Germany, Italy, India, Japan, Korea, Russia, Slovakia, the United Kingdom and the

Immobilisation of Radioactive Wastes in Glass

347

United States. The potential use of glass to immobilise nuclear waste was initially investigated in the early 1950s in Canada using natural nepheline syenite as the starting material. This was mixed with an acidic solution of the waste material together with lime, and the mixture melted at 1250
1350°C in a fireclay crucible. An active pilot plant was subsequently constructed at Chalk River to demonstrate the feasibility of the vitrification process on a larger scale. Radioactive blocks of glass were produced between 1958 and 1960 although the vitrification programme was terminated in 1960 because no fuel reprocessing was foreseen in Canada at that time. Battelle Pacific Northwest National Laboratory (PNNL) in the United States developed an ‘In-Can-Melter’ technique (Fig. 19.16) utilising a spray calciner directly attached to the disposal canister, which also acts as the glass-manufacturing vessel. Glass frit is added to the calcine as it drops into the container, which is surrounded by a zoned furnace. The calcinefrit mixture is heated to 1000
1100°C forming a molten glass. As the canister is filled, the heating band of the furnace is raised to match the reaction zone. The containers are generally filled to the 90% level, at which time the calcine-frit feed is moved to a second canister already in place. The filled ‘can’ has been removed and replaced by an empty one ready for the next switching operation.

Figure 19.16 In-Can-Melter vitrification process.

348

An Introduction to Nuclear Waste Immobilisation

A fluidised bed calciner in conjunction with the in-can production technique is less developed although PNNL later developed another process based on application of JHCMs. In the United Kingdom also, work started in the 1950s, initially using natural soils as the base material for glass formation. These glasses had to be melted at high temperatures (about 1500°C) to produce homogeneous, bubble-free products. Subsequently, alkali borosilicate glass compositions were developed that could dissolve up to 30% waste oxides and that could be melted at lower temperatures. Between 1958 and 1962 a vitrification pot process was developed at the UK Atomic Energy Authority’s Harwell laboratory called the fixation in glass of active liquors (FINGAL) process. HLW plus glass-forming additives were fed into the final storage container, which was surrounded by a zoned furnace (Fig. 19.17). Heating begins at the lowest portion of the furnace, and three layers form: molten glass at the bottom, covered by calcine, and with a layer of boiling liquid on top. As the glass level is raised, the higher heating elements of the furnace were activated. After the glass and waste had been calcined, melted and homogenised, the pot was removed and replaced with another pot containing waste and frit ready to be calcined and melted. The FINGAL process was later modified and scaled up, becoming known as the highly active residue vitrification experimental studies (HARVEST) process. The HARVEST process included the use of an annular container, with an

Figure 19.17 British rising glass vitrification process fixation in glass of active liquors, FINGAL.

Immobilisation of Radioactive Wastes in Glass

349

additional heater in the annulus, and utilisation of internal fins in the container for heat removal. Simple pot vitrification processes that do not employ a separate calcination stage have been used in France (Pilote Verre: PIVER), Italy and China. Such a process is currently used commercially in India (WIP: Tarapur) for the immobilisation of HLW generated during the reprocessing of spent fuel. France has pioneered development of vitrification processes. Several types of matrices were investigated, including crystalline materials, phosphate and borosilicate glasses. The first laboratory scale unit was named Vulcain. It was commissioned in 1957 and was followed by a first vitrification pilot unit, Gulliver, commissioned in 1964. This early work culminated in the pilot-scale facility PIVER, featuring a single inductionheated pot where the three operations of evaporation of the HLW solutions, calcination of the residue and glass formulation were performed. PIVER operated successfully from 1969 to 1973 to produce 12 t of glass containing 5 3 106 Ci of activity. PIVER resumed operation in 1979 to vitrify HLW from the reprocessing of fast-breeder fuels. It was decommissioned between 1988 and 1990. Although these and other related vitrification processes were successful in pilot plant trials, batch processing considerations led France to choose a continuous melting process for fullscale development. The French process includes a separate calcining stage. Waste is first calcined employing a rotary kiln before being fed under gravity into a metallic pot heated inductively (Fig. 19.18). This is a

Figure 19.18 Two stage vitrification process AVM. AVM, Atelier de Vitrification de Marcoule.

350

An Introduction to Nuclear Waste Immobilisation

continuous process with new glass frit and waste material being supplied to the furnace and molten glass being fed from the furnace via a freeze
thaw valve directly into a separate storage canister. The major advantage of this process is that much higher throughputs can be achieved. The Atelier de Vitrification de Marcoule (AVM) was the first in-line vitrification facility which proved the technology of the French two-step vitrification process which converted HLW solutions to a solid form in a rotary calciner and then vitrified them in an induction-heated metallic melter. The first industrial HLW vitrification plant using AVM technology started operation in 1978 at Marcoule. Similar plants (R7 and T7) have been built at La Hague. HLW vitrification plants R7 and T7 are designed to produce 600 canisters of glass per year. R7 entered active service in June 1989 whereas the T7 facility entered active service in July 1992. HLW vitrification plants R7 and T7 have produced more than 14,000 glass canisters (5573 t) with overall radioactivity about 6430 3 106 Ci to 2008 (Jantzen, 2011a). The alternative continuous process uses JHCM and involves melting the glass in a tank constructed from refractory ceramic blocks by passing an electric current through the molten glass using submerged electrodes. Calcined HLW is mixed with glass frit or glass forming additives and fed into the JHCM (Fig. 19.19). Cold material covering the surface of the

Figure 19.19 Schematic of one-stage vitrification processes in a Joule heated ceramic melter used in the United States, Russia, Germany, Belgium and Japan.

Immobilisation of Radioactive Wastes in Glass

351

molten glass (the so-called cold cap) reduces evaporation losses. A pilot vitrification plant based on JHCM was commissioned at PNNL in 1984. In 1985 the joint German
Belgian vitrification plant Pamela at Mol (Belgium) was commissioned. This plant uses a continuous vitrification process in which HLW slurry together with glass frit is fed into a JHCM equipped with two drainage systems. One drain is employed to produce glass blocks, while the second drain is used to feed a glass bead production unit. This produces glass beads with a diameter of 50 mm. The beads are subsequently dispersed in a lead alloy matrix to produce a product called vitromet. The Pamela melter used four Inconel 690 electrodes and operated at power levels up to 100 kW with MoSi2 elements enabling start up. Pamela produced 64 t of waste glass in 411 canisters in 1985
86. Pamela revealed that noble metals (e.g. Pd, Rh and Ru) gradually accumulate at the bottom of the furnace and lead to increased power consumption due to the higher electrical conductivity of this layer. It is now recognised that a sloped furnace bottom is desirable, thus enabling the viscous noble metal rich layer to be purged more easily from the furnace. In 1987 Russia commissioned its first JHCM-based HLW vitrification plant EP-500 at Mayak Chemical Combine (MCC) in the Ural region. This plant had a capacity of 500 L/h in terms of initial HLW. About 1000 m3 of HLW was vitrified during the next 18 months, producing 160 t of waste glass with overall radioactivity about 4 3 106 Ci. EP-500 melter is heated by a number of water-cooled molybdenum electrodes located at the melter bottom. Electrical current to the electrodes was supplied through water-cooled stainless steel tubes and the maximum power requirement of the melter was around 1.5 MW. In 1988 EP-500 was shut down due to a current supply failure and a second EP-500/1-p melter was commissioned. It operated from 1991 to 1997 processing 11,500 m3 of HLW with total activity of 282 3 106 Ci. A third melter EP-500/3 operated from 2001 to 2006, processing 8000 m3 of HLW with total activity 175 3 106 Ci. The fourth melter EP500/4 operated from 2006 to 2010, processing 8100 m3 of HLW with total activity 182 3 106 Ci. This was followed by fifth melter EP500/5 operating since December 2016. Fig. 19.20 shows a schematic of the MCC HLW vitrification plant and EP-500 melter. In 1991 the United Kingdom started its HLW vitrification at Windscale, now Sellafield. The advantages offered by the AVM process, coupled with its successful full-scale operation in France, led the United Kingdom to choose this process, in preference to the HARVEST

352

An Introduction to Nuclear Waste Immobilisation

Figure 19.20 Schematic of one-stage vitrification plant at MCC, Russia. (A) Vitrification plant flow sheet. (B) EP-500 melter design. (1) drainage (pouring unit); (2) molybdenum electrodes; (3) overflow window; (4) feeders; (5) arch; (6) off-gas duct; (7) water-cooled electric power supply; (8) refractory blocks; (9) fireclay refractory blocks. MCC, Mayak Chemical Combine.

353

Immobilisation of Radioactive Wastes in Glass

method, for its commercial waste vitrification programme. Currently, the HLW vitrification plant at Sellafield WVP has three production lines producing 475
600 containers of vitrified waste per annum and up to 2018 the Sellafield WVP has produced over 6000 containers of vitrified waste. The process at WVP at Sellafield is the Atelier de Vitrification de la Hague (AVH) process, which differs from AVM in that an elliptical- rather than cylindrical-shaped melter is used (Fig. 19.21). In 1995 Japan inaugurated its industrial HLW vitrification plant in Tokai. It employs a JHCM with a steeply sloped floor. Use of a threeelectrode assembly allows effective heating and stirring of the melt, presumably resulting in a more homogeneous product. The total waste glass produced by the Japan HLW vitrification plant to date is less than 100 t. In the mid-1990s two large-scale HLW vitrification plants based on JHCM became operational in the United States. In March 1996 the Savannah River Defence Waste Processing Facility (DWPF) became the first plant in the United States to vitrify HLW. The DWPF to December 2015 produced 6890 t of HLW glass containing 56.4 3 106 Ci. The DWPF JHCM melter (Fig. 19.22) has a glass surface area of 2.6 m2 and

(A)

(B)

Constant volume feeders (CVF’s)

Calcine

Glass

Air Sparge Vitrification Cell

Highly active liquor (HAL) is fed into the process via a constant Volume Feeder (CVF)

1150ºC Zone 2

Calciner

Zone 2 Melter Back Pressure Level

Zone 3

Zone 3

Zone 4

Zone 4

Zone 5

1050ºC Cloche

Zone 5

Melter assembly

Off-gas system

Pour Nozzle

Glass Plug

Drain Nozzle

Melter turntable Segregation dome

Elevating tables

Lid placing machine Pour cell

Container carousel

Figure 19.21 Schematic diagram of WVP at Sellafield, UK. (A) WVP flow sheet; (B) hot-walled induction melter for high level waste vitrification. WVP, Waste vitrification plant. Courtesy Nick Gribble, National Nuclear Laboratory, UK.

354

An Introduction to Nuclear Waste Immobilisation

Thermowell Closed circuit TV

Off-gas Stainless steel shell

Mullfrax 202 refractory brick Lid heaters

34′′ 8′ 1/2′′

K3 refractory brick

Feed tubes 10′′ (25 cm)

Electrodes (4)

10′′ (25 cm) 12′′ (30 cm)

Zimul refractory brick

Drain valve

Vacuum

Canister

Figure 19.22 Schematic diagram of the Joule heated ceramic melter operating at the Defence Waste Processing Facility at the Savannah River Site, US. Courtesy J. Marra, SRNL.

operates with a reducing atmosphere aiming to minimise foaming, suppress carryover of 99Tc, 129I and 104Ru, and limit corrosion of refractories lining the melter. It was supplied in 2011 with a set of Ar bubblers to increase throughput. The West Valley vitrification plant West Valley Demonstration Project (WVDP) has completed its campaign. It vitrified 2500 m3 of liquid HLW to produce 275 glass canisters between 1996 and 2002. The WVDP rectangular melter had a sloped floor and pour spout offset aiming to prevent accumulation of noble metal and refractory deposits. High Cr2O3 fuse-cast refractories were used with high-nickel alloy electrodes to provide heating. The melt surface area was 2.2 m2 and operating temperature was 1150°C. The melter used an air lift pour operating batch-wise and a vacuum lift from the top of the melter as an emergency drain (Fig. 19.23).

Off-gas cleaner

Passive cooled feed nozzle Melter viewing system

Off-gas film cooler To SBS Airlift Overflow heaters

Thermocouples

Melter electrodes Canister access port Melter assembly

North cell wall

Filing canister

Load cell Turntable frame

Turntable assembly

Figure 19.23 Schematic of West Valley Demonstration Project melter with sloped floor. Courtesy U.S. Department of Energy.

356

An Introduction to Nuclear Waste Immobilisation

Rest of LAW off-gas system LAW off-gas 1000 SCFM

LAW glass forming chemicals LAW waste from pre-treatment facility 9000 gal batch

Submerged bed scrubber (SBS)

LAW melter feed ~3 GPM

Condensate back to pre-treatment LAW melter ~20 MT inventory 15 MT/d

LAW melter feed prep vessel 3000 gal batch

LAW melter feed vessel 3000 gal batch

(A)

LAW canister 2.5 Canisters per day

(B)

Figure 19.24 Schematic of low level waste vitrification facility at Hanford, USA (A), and an artistic view of of one of its melters (B). Courtesy U.S. Department of Energy.

The more recent HLW vitrification plant at Hanford, USA has moved into the construction stage (see Fig. 1.1B). Vitrification is also used for immobilisation of LILW. The vitrification facility for the low level radioactive waste (LLW) at Hanford, USA is under construction and is expected to produce over 200,000 m3 of glass. The plant has two melting units (Fig. 19.24), each of them with a capacity of 15 t per day. Operation of this plant is envisaged to start in 2019. In Germany (Karlsruhe) the vitrification plant is a compact-sized JHCM with a sloped bottom avoiding sedimentation of noble metals upon melting (Fig. 19.14). Its vitrification campaign completed in 2011, processing B70 m3 of HLW with a gross activity of 24 3 106 Ci at the former German pilot reprocessing plant that had operated for 20 years terminating in 1990. Table 19.12 summarises current operational data on radioactive waste vitrification facilities.

19.15 CALCINATION PROCESSES Two-stage vitrification processes utilise preliminary calcination of the waste. The oldest of the calcination techniques is pot calcination in which the entire procedure takes place in a ‘pot’. The liquid waste is added batch-wise to the container, which is enclosed in a zoned furnace. As the water boils off and oxides form, a crust of calcine forms on the sides and bottom. The process is continued until the container is about 90% filled. The batch nature of this process limits its use for any large-scale application, and quality control is more difficult than in other methods.

Table 19.12 Operational data of vitrification programmes Facility

Waste type

Melting process

Operational period

Performance

R7/T7, La Hague, France

HLW

IHC

1989/1992

AVM, Marcoule, France R7, La Hague, France

HLW HLW

IHC CCM

1978
2012 2010


WVP, Sellafield, UK DWPF, Savannah River, USA WVDP, West Valley, USA EP-500, Mayak, Russia CCM, Mayak, Russia Pamela, Mol, Belgium Karlsruhe, Germany Tokai, Japan

HLW HLW

IHC JHCM

1990
1996


6555 tonnes in 16885 canisters, 262  106 TBq to 2012 1357 tonnes in 3306 canisters, 22  106 TBq to 2012 GCM: U-Mo glass 76 tonnes in 190 canisters to 2012 2200 tonnes in 5615 canisters, 33  106 TBq to 2012 6300 tonnes in 3591 canisters, 1.8  106 TBq to 2012

HLW HLW HLW HLW HLW HLW

JHCM JHCM CCM JHCM JHCM JHCM

1996
2002 1987
Pilot plant 1985
1991 2009
2010 1995


Radon, Russia Radon, Russia Radon, Russia Bohunice, Slovakia WIP, Trombay, India

LILW LILW ILW HLW HLW

JHCM CCM SSV IHC IHPTM

AVS, Tarapur, India WIP, Kalpakkam, India WTP, Hanford, USA VPC, SEPEC Site, China Taejon, Korea

HLW HLW LLW HLW LILW

IHPTM JHCM JHCM JHCM CCM

1987
1998 1999 2001
2002 1997
1985
2002, 2002
2006
Testing 1998
Testing Testing

570 tonnes in 570 canisters, 0.9  106 TBq B6200 tonnes, 643 106 Ci 18 kg/h by phosphate glass 500 tonnes in 2201 canisters, 0.5  106 TBq 55 tonnes in 140 canisters, 0.8  106 TBq 70 tonnes in 241 canisters (110 L), 0.4  106 Ci to 2007 10 tonnes . 30 tonnes 10 kg/h, incinerator ash 1.53 m3 in 211 canisters 18 tonnes, 110  103 Ci to 2012 10 tonnes in 100 canisters, 0.15  106 TBq B1000
2000 tonnes

IHC, Induction, hot crucible; CCM, Cold crucible induction melter; JHCM, Joule heated ceramic melter; SSV, Self-sustaining vitrification; IHPTM, Induction heated pot type melter.

358

An Introduction to Nuclear Waste Immobilisation

Figure 19.25 Schematic of four main types of calcination processes. (A) spray, (B) fluidised bed, (C) stirred bed, (D) rotary kiln.

In spray calcination (Fig. 19.25A) the waste is introduced into the top of an externally heated reaction chamber through a spray orifice along with a jet of atomising air. Water is driven off of the falling droplets and the waste is largely converted to oxides and is collected in the form of a fine spheroidal powder (#1 mm). The reaction furnace is operated to produce a 100°C wall temperature, although the calcine temperature itself is typically in the 350
550°C range. The technique is able to handle waste of almost any concentration. Fluidised bed calcination (Fig. 19.25B) also accepts wastes at almost all concentrations. In this process the wastes are kept in suspension by air jets from below and heated internally to 500
600°C. Evaporation occurs from the surfaces of the original bed particles and results in a product consisting of granular bed material and powdered calcine, both of which are

Immobilisation of Radioactive Wastes in Glass

359

continuously removed from the reactor. Heating of the bed was originally done by means of an exterior furnace, but this resulted in high losses of Ru and Cs. Heating accomplished by kerosene combustion in the bed resolves this problem. The stirred bed reactor at Mol in Belgium is a variation of the fluidised bed approach (Fig. 19.25C), the materials being stirred as well as fluidised by air jets. The stirring approach allows better control of particle size in the finished product. This is particularly important for high-alumina wastes. Addition of phosphate to the waste feed produces metal phosphates and a substantial amount of aluminium phosphate that acts as a secondary containment. Leachability characteristics are considerably improved, and the product is termed super-calcine. The rotary kiln calciner (Fig. 19.25D) has been largely developed in France. The equipment consists of an externally heated (500
600°C) rotating cylinder tilted at a slight angle from the horizontal so that the waste introduced at the upper end is dried and almost completely denitrated before it exits at the lower end. A loose bar in the barrel keeps the calcine free-flowing and prevents wall deposits from building up. Calcines are typically fine powders and thus relatively dispersible, this being a serious consideration in the event of a transportation accident. The untreated materials are also easily soluble in water. These two characteristics have largely eliminated calcines from consideration as a final disposal form. They can be highly reactive chemically and excellent raw materials for further processing. A high sodium nitrate content, typical of many wastes, creates problems in all the calcination processes. This salt has a melting temperature of 307°C and, upon melting, forms a viscous, sticky mass that resists further decomposition. The addition of finely divided metallic iron to the wastes helps with this problem. Powdered silica has been added to the feed stream to enrich calcines in silicates, which are more readily knocked loose by vibrating hammers acting on the outside of the walls. Processes have also been developed where much of the nitrate in the wastes is destroyed by pre-treatment with formaldehyde or formic acid. Most calcines as originally produced will still contain traces to substantial amounts of undecomposed nitrate salts as well as small amounts of residual water. Calcines produced by the fluidised bed process typically have high levels of either alumina or zirconia. Table 19.13 gives some calcination properties which are obviously highly dependent on the compositions of the original wastes, which can vary considerably. Most current WVPs use rotary kiln calcination.

360

An Introduction to Nuclear Waste Immobilisation

Table 19.13 Properties of calcines Property

Calciner type

3

Bulk density, g/cm Highest operating temperature, °C Thermal conductivity, W/m K Specific area, m2/g Porosity, %

Pot

Fluidised bed

Spray

Rotary kiln

1.1
1.4 420

2.0
2.4 500
600

1.0
2.4 700

1.0
13 500

0.35
1.0

0.2
0.3

0.2

0.2

0.1
5.0 40
85

0.1
5.0 45
80

10
20 30
75

0.1
5.0 70
80

19.16 COLD CRUCIBLE MELTERS Recent developments in vitrification include the melting technologies such as plasma and in-container vitrification (see Section 20.6) and application of Cold Crucible Melters (CCM) to immobilisation of wastes, which are difficult to melt, as well as production of GCM and mineral-like immobilising materials. In a CCM the glass melt is heated by induction currents. The melt is preserved inside a cooled volume of glass-batch material (termed the skull; CCM is also known as skull melting) that isolates the high temperature melt from the water-cooled stainless steel or copper tubes which make up the cold crucible walls (Fig. 19.26A). CCMs are used either to melt waste with glass forming additives (Russia) or calcined waste with glass frit (France). A colder surface layer or cold cap of dried batch is formed in the one-stage vitrification CCM as forms in a JHCM, diminishing losses of aerosols and volatilised radionuclides from the glass. The construction elements of a CCM are transparent to the electromagnetic field produced by the induction coils, allowing induction currents to be generated inside the material contained in the crucible. The cooling water in the tubes is at a temperature less than 100°C so that molten glass in contact with them can be kept at B200°C. Even during melting the tubes remain cold and a protective layer made of glass batch material forms between the melt glass and tubes insulating them from the melt. This protective skull (Fig. 19.26B) means that refractory liners required in other melters are not needed. Since active refractory materials are difficult to immobilise (e.g. they cannot be melted easily) this is a distinct advantage. CCMs are well suited for obtaining high throughput in a small volume. CCMs supplied with stirrers (Fig. 19.26A) have enhanced throughputs

Immobilisation of Radioactive Wastes in Glass

361

Figure 19.26 (A) Schematic of Russian stainless steel CCM. The stirrer in this CCM enables production of GCM. (B) A sample of glass from the CCM: the white layer is a part of the glass forming batch (skull) that remains unmelted. GCM, Glass-composite material; CCM, cold crucible melters.

( . 20%) and enable production of GCM to immobilise wastes which are difficult to melt. Low crucible weight facilitates easy dismantling and small amounts of secondary waste compared to conventional melters. Neither glass nor metal adheres to the cooled wall, which is therefore never subject to strong contamination. CCMs have the advantages of extended operational lifetime, flexibility to glass formulations, larger operation temperature range and high glass output at significantly smaller sizes. Since the CCMs are protected from the glass melt by a cold glass batch layer, glass can be produced at higher temperatures, which increases the glass formulation range. In addition, because the melt comes in contact only with solid material of the same composition as glass batch, the final product has a high degree of purity, which is particularly important for meeting product specification. Waste vitrification using CCM may be either one- or two stage; that is with preliminary calcination or with calcination occurring in the melter. One disadvantage of CCMs is their high specific energy consumption. Table 19.14 illustrates the main process parameters of a CCM compared to a JHCM. The first LILW vitrification plant based on CCM was put into operation in 1998
99 in Russia at Scientific and Industrial Association

362

An Introduction to Nuclear Waste Immobilisation

Table 19.14 Process parameters of Joule heated ceramic melter and cold crucible melter Parameter

Melt capacity, kg/h Specific melt capacity, kg/ m2 h Melting ratio, (kW h)/kg Operating temperature, °C Corroding in melt Gross weight of loaded melter, kg Operating mode

Melter type Joule heated ceramic melter

Cold crucible melter

25 40
120

25 170

2.5
3.2 ,1300 Refractory and electrodes .1000

4.4
6.4 ,3000 No ,200

Continuous

Continuous or batch

“Radon” (SIA ‘Radon’). This plant with a melt capacity of 75 kg per hour is currently operated as a prototype plant for the vitrification of operational radioactive waste from nuclear power plants (NPPs). A pilot HLW vitrification plant with CCM is currently under test at spent nuclear fuel reprocessing plant RT-1 in the Ural region, Russia. It started test operations in 1994 aiming to immobilise partitioned radionuclides in glasses, GCM, Synroc-type and other mineral-like materials. France has also intensively investigated application of CCMs for both HLW and LILW immobilisation including joint projects with Russia and the Czech Republic. Several CCM test platforms have been built at the Marcoule pilot facilities. These programmes were initially focused on the treatment of HLW solutions from light water reactor fuel, producing simulated R7/T7 glass. The feeding systems allow simultaneous controlled feeding of solids (frit, powders) and liquids (surrogate solutions, sludges). A CCM was retrofitted in 2011 to the La Hague vitrification line after 20 years of hot walled operation. This has enabled vitrification of some difficult U/Mo-containing wastes which required corrosive glass compositions needing higher melting temperatures than available in hot walled melters. Fig. 19.27 shows a GCM for immobilisation of such a U/Mo-enriched HLW made via CCM. South Korea also plans to operate a WVP based on CCM to vitrify LILW from its NPPs. Pilot plant tests have been successfully completed

Immobilisation of Radioactive Wastes in Glass

363

Figure 19.27 GCM made of a silicate host glass encapsulating Mo-enriched phases to immobilise U-Mo high level waste. GCM, Glass-composite material. Courtesy E. Vernaz, CEA.

for various waste types. CCM melting was also intensively studied for Hanford site wastes in the United States.

19.17 IN SITU VITRIFICATION In situ vitrification technology has been designed to melt contaminated soils containing sludge and other waste materials and immobilise the contaminants including both stable and radioactive nuclides on site (e.g. in situ) in a vitreous material. The technology uses Joule heating to electrically melt the contaminated soil using an array of several electrodes at temperatures ranging typically between 1500 and 2000°C. The in situ vitrification unit typically consists of electrical power supply system, offgas collection hood with graphite electrodes and off-gas treatment system (Fig. 19.28). The electrical power supply system provides two-phase alternating current at the appropriate voltage and amperage to the graphite electrodes used in the melting process. The off-gas dome-shaped hood supports the electrodes and collects emissions from the vitrification zone. A low vacuum is maintained in the off-gas hood during operation to contain off-gases which are purified by the off-gas treatment system. It typically

364

An Introduction to Nuclear Waste Immobilisation

Back up off-gas system Off-gas treatment Electrode

Power supply

Off-gas hood

Glycol cooler

Quench scrub

Dewater heat filter

Thermal oxidizer (optional) Activated carbon (optional)

Recycling of secondary waste

Clean backfill

Absorption bed (cobbles under gravel under soil) Vitrified material

Contaminated soil zone

Planar melt

Figure 19.28 Schematic of in situ vitrification process. Courtesy IAEA.

consists of a quencher, scrubber, demister, heater, particulate filter, activated carbon adsorber, blower and optional thermal oxidation unit. The quencher lowers the off-gas temperature and the scrubber removes acid gases and large particulates. The off-gas is then dewatered and reheated to prevent wetting of the particulate filters after which it is filtered to remove fine particulates and then polished to remove trace organics using either an activated carbon adsorber or a thermal oxidation unit. Melting of contaminated soil can use starter paths injected through a series of closely spaced holes bored or driven alongside the zone to be treated; for example in planes on either side of the waste or above and below the waste and the two melts are allowed to converge. Since planar melting is typically completed below the surface, this approach enables deeper zones to be treated without higher power requirements. The unmelted soil above the treatment zone acts as a thermal insulator, conserving energy at the melt depth and keeping surface equipment significantly cooler. On operating the melt pool gradually grows deeper and wider until the desired volume has been formed. The molten mass then solidifies into a vitreous monolith material similar to a GCM durable enough to immobilise remnant contaminants. Individual melts can be up to 7 m deep and 15 m in diameter; for example large volumes can be vitrified using adjacent melting to form a massive continuous monolith. A bulk vitrification process is used at Hanford in which liquid waste is mixed with controlled-composition soil in a disposable smelter. Electrodes

Immobilisation of Radioactive Wastes in Glass

365

are inserted to vitrify the mixture and when cooled, the smelter, its contents and the embedded electrodes will be buried as LLW. An in situ vitrification process was attempted in the clean-up of heavily contaminated soil at a nuclear weapons test site at Maralinga in Australia in the late 1990s, but this was abandoned after an explosion arising from lack of care in planning and implementation.

19.18 RADIONUCLIDE VOLATILITY Many waste radionuclides as well as inactive elements may be lost to varying degrees during the melting process. Fig. 19.29 shows typical volatilisation behaviour of radionuclides as a function of temperature. The elements of most concern are ruthenium and caesium because of the 103,106Ru and 134,137Cs radionuclides in the waste. Caesium is not volatilised up to about 400°C. Caesium nitrate is decomposed at 414°C with formation of caesium oxide, caesium peroxide and metallic caesium, which are highly volatile. To suppress caesium volatility acidic oxides are added that form stable silicates or phosphates. Caesium volatility in glass melts is effectively suppressed by boron and titanium oxides. Both radioactive and inactive ruthenium is produced in fission of nuclear fuel. Waste Ru, if volatilised, may subsequently condense and generate plugging problems. As much as 50% of the Ru in spent oxide

Figure 19.29 Volatilisation of nuclides with temperature.

366

An Introduction to Nuclear Waste Immobilisation

fuel elements may remain in insoluble form in the reprocessing plant dissolver. This refractory material is in the form of colloidal RuO2 particles. The fraction that does go into solution is in the form of nitrosyl complexes. These nitrosyl complexes are thought to be converted in the hightemperature processes to higher Ru oxides such as RuO4, generally assumed to be responsible for the Ru losses, although it has been demonstrated that RuO2 also volatilises at 1100°C and above. Sugar is often added in calcination or melting to suppress Ru volatilisation. Technetium is also easily volatilised. Tc can be in the waste in two oxidation states: IV and VII. Tc is volatilised even during the evaporation stage when removing excess water from HLW that contains acidic solutions of Tc(VII) as pertechnetate (TcO4)2. Tc-IV (TcO2), while not as volatile as Tc-VII, also has a high vapour pressure at glass melting temperatures. Other waste radionuclides are not so volatile. However, losses of radionuclides are also caused by mechanical and aerosol carryover from the melter. Batch composition largely governs the total carryover. For example, the presence of chlorides enhances losses of corrosion products whereas reducing agents such as sugar diminish radionuclide volatilisation. Liquation (phase separation) also influences total carryover of radionuclides. Some refractory particles from the HLW solutions remain during the melting process in non-dissolved form being incorporated in the glass matrix as inclusions. This has a beneficial effect on suppression of radionuclide volatilisation. Typical and acceptable losses of the volatile radionuclides from melters are 1%
4%.

19.19 WASTEFORM ACCEPTANCE CRITERIA Vitrification of nuclear wastes ensures maximum reduction of the potential for radionuclide release into the environment. Average diminishing factors Kwf for the vitrified aqueous radioactive waste (see Section 3.6) are of the order of 104
105 sufficient to provide safe long-term storage of HLW in above-ground facilities as well as safe transportation and final disposal. For LILW vitrification enables simplified near-surface disposal facilities with better radionuclide retention than cements and bitumen. Although vitrification is a complex technique it can be competitive due to the high VRF values and excellent retention capabilities of the wasteform. Acceptance criteria for vitrified wastes specify the minimum requirements to be met by the wasteform. Compliance with these criteria

Immobilisation of Radioactive Wastes in Glass

367

Table 19.15 Minimum Waste Acceptance Criteria requirements of vitrified Russian high level waste Parameter

Chemical durability (NR, g/cm2 day) 137 Cs 90 Sr Pu Homogeneity of the glass bulk: structure; chemical composition, wt.% Thermal durability (absence of changes in structure and chemical composition), °C Radiation stability: to β and γ radiation, Gy to α radiation, decay/g Compressive strength, MPa Flexural strength, MPa Elastic modulus, GPa Thermal expansion coefficient, 1/K Thermal conductivity from 20 to 500°C, W/(m K) Content of fissile materials, wt.%

Limits

, 1 3 1025 , 1 3 1026 , 1 3 1027 Homogeneous 6 10 .550 .108 .1018
1019 .9 .41 .54 ,9 3 1026 1
2 ,2

ensures safe operation of storage and disposal facilities. Table 19.15 illustrates the acceptance criteria for vitrified HLW in Russia according to the state standard GOST P 50926-96.

REFERENCES Donald, I. W. (2010). Waste immobilisation in glass and ceramic based hosts. Chichester: Wiley. Gin, S., Jollivet, P., Tribet, M., Peuget, S., & Schuller, S. (2017). Radionuclides containment in nuclear glasses: An overview. Radiochimica Acta, 105(11), 927
959. IAEA. (1979). Characteristics of solidified high-level waste products. TRS-187. Vienna: IAEA. IAEA. (1992). Design and operation of high level vitrification and storage facilities. TRS-339. Vienna: IAEA. Jantzen, C. M. (2011a). Historical development of glass and ceramic waste forms for high level radioactive waste. In M. Ojovan (Ed.), Handbook of advanced radioactive waste conditioning technologies (pp. 159
172). Cambridge: Woodhead. Jantzen, C. M. (2011b). Development of glass matrices for HLW radioactive wastes. In M. Ojovan (Ed.), Handbook of advanced radioactive waste conditioning technologies (pp. 230
292). Cambridge: Woodhead. Lutze, W. (1988). Silicate glasses. In W. Lutze, & R. Ewing (Eds.), Radioactive waste forms for the future (pp. 1
160). Amsterdam: Elsevier Science Publishers B.V. Navrotsky, A. (1998). Thermochemistry of crystalline and amorphous phases related to radioactive waste. In P. A. Sterne, et al. (Eds.), Actinides in the environment. (pp. 267
297). The Netherlands: Kluwer Academic Publishers.

368

An Introduction to Nuclear Waste Immobilisation

Ojovan, M., & Lee, W. E. (2011). Glassy wasteforms for nuclear waste immobilisation, Met and Mats. Trans. 42A, 837
851. Ojovan, M. (2012). Viscous flow and the viscosity of melts and glasses. Physics and Chemistry of Glasses, 53(4), 143
150. Ojovan, M. I., & Lee, W. E. (2006). Topologically disordered systems at the glass transition. J. Phys.: Condensed Matter, 18, 11507
11520. Ojovan, M. I., & Lee, W. E. (2007). New developments in glassy nuclear wasteforms. New York: Nova, 131p. Pegg, I. L., & Joseph, I. (2001). Vitrification. In C. Ho Oh (Ed.), Hazardous and radioactive waste treatment technologies handbook. Boca Raton, FL: CRC Press, 4.2.1-27 (2001). Plodinec, M. J. (2000). Borosilicate glasses for nuclear waste immobilisation. Glass Technology, 41, 186
192. Poluektov, P. P., Schmidt, O. V., Kascheev, V. A., & Ojovan, M. I. (2017). Modelling aqueous corrosion of nuclear waste phosphate glass. Journal of Nuclear Materials, 484, 357
366. Sobolev, I. A., Ojovan, M. I., Scherbatova, T. D., & Batyukhnova, O. G. (1999). Glasses for radioactive waste. Moscow: Energoatomizdat. Varshneya, A. K. (2006). Fundamentals of inorganic glasses. Sheffield: Society of Glass Technology, 682 p. Vashman, A. A., Demine, A. V., Krylova, N. V., Kushnikov, V. V., Matyunin, Yu. I., Poluektov, P. P., . . . Teterin, E. G. (1997). Phosphate glasses with radioactive waste. Moscow: CNIIatominform, 172 p.

FURTHER READING Gribble, N. R., Short, R., Turner, E., & Riley, A. D. (2009). The impact of increased waste loading on vitrified HLW quality and durability. Materials Research Society Symposia Proceedings, 1193, 283
290.