During chemical separation, the 232 U follows the uranium product and the 228 Th follows the thorium. The activity Xoe Лад (0 of 228Th that has again built up in the separated uranium during a time t after separation is obtained by applying Eq. (2.14):

XosAWO (8.25)

Л08 ~~ A22

where N22 is the amount of 232U present after separation. For a time scale in years, the 228Th daughters will be in secular equilibrium and the beta activity (XjV)g at time t is just twice the activity given by Eq. (8.25).

During the first few years, decay of 232U is negligible, so that Eq. (8.25) becomes

XosAog (t) = X22A£ (1 — <ГЛ«’) (8.26)

and for a time scale in days, the growth of beta activity due to 212Pb, 212Bi, and 208T1 is given by

where X224 is the decay constant for 224 Ra. Buildup and decay of beta activity and gamma dose as a function of days after separation are illustrated in Fig. 8.13. It is important that the uranium product from thorium irradiation be carried rapidly through the refabrication operations soon after chemical separation.

The high surface activities of the uranium require semiremote refabrication methods, whereby direct bodily contact with the material is avoided, but only distance or light shielding need be used to avoid above-tolerance radiation doses. To illustrate, assume that personnel performing the fabrication operations are separated from the work by an average distance of 1 m and that each person is exposed to 1 kg of separated uranium containing 100 ppm 232 U. From Fig. 8.13, we see that the refabrication must be completed within only 6 days after separation if the workers handling the last-and usually most delicate-stages of fabrication are to receive a typical tolerance dose of no more than 2.5 mr/ht at the end of the fabrication period. Alternatively, for an allowable dose of 2.5 mr/h, averaged over the entire period since separation, the allowable time to complete semiremote fabrication is 11 days.

A longer wait may necessitate remote fabrication, whereby all operations must be carried out behind heavy shields. For example, uranium containing 100 ppm 232 U and aged 35 days since final separation would yield an unshielded dose of 38 mr/h per kilogram. From the data in Fig. 2.4, it is estimated that the fabrication must be carried out behind about 7 cm of lead or 35 cm of concrete if the dose to operators is to be 2.5 mr/h or less.

As the 232 U concentration increases, the allowable time for semiremote fabrication decreases rapidly, and greater shielding thicknesses for remote fabrication are required.

The properties of metallic curium are listed in Table 9.27. Microgram quantities of metallic curium have been prepared by reducing curium trifluoride CmF3 with barium vapor in a vacuum chamber at 1315 to 1375°C, using a tantalum crucible, followed by heating at 1235°C to volatilize excess barium and BaF2 slag [C8]. Larger quantities have been produced by suspending Cm02 in a MgCl2-MgF2 melt and adding magnesium, in the form of a magnesium-zinc alloy, to reduce the curium. Unreacted magnesium and zinc are removed by evaporation [El ].

5.3 Curium Compounds

Curium oxides. The stoichiometric binary oxides of curium are Cm203 and Cm02. The dioxide Cm02 is not stable above about 350°C, changing into nonstoichiometric phases with О/Cm < 2. The sesquioxide Cm203 is stable up to its melting point of 2002°C. A black product near the composition of Cm02 is formed when the sesquioxide Cm2 03 or aqueous compounds such as Cm(OH)3 or Cm2(C204)3 are heated in air or oxygen at 500 to 600°C. The sesquioxide Cm203 is formed by the decomposition of Cm02 at 600 to 700°C in high vacuum, by reduction of Cm02 with hydrogen, or by the decomposition of the dioxide in air at higher temperatures. The stable form of curium sesquioxide at room temperature is the hexagonal lattice.

Curium halides. Like americium, curium forms binary halides in the oxidation states of III and IV. The trifluoride can be precipitated by adding fluoride ions to a solution of Cm(III), drying in vacuum over P205, and reducing in HF gas at 600°C. The tetrafluoride CmF4 is prepared by heating anhydrous CmF3 in elementary fluorine at 400°C. The trichloride CmCl3 is very hygro­scopic [K2]. The tribromide and triiodide are also formed.

The principal types of thorium-based fuel to which the Thorex process has been applied are

1. Th02-U02 fuel with stainless steel cladding

2. Irradiated Th02 with aluminum cladding

3. HTGR, AVR, or THTR fuel, consisting of particles of Th02 or ThC2 and particles of U02 or UCj imbedded in a graphite matrix

Stainless steel-clad fuel. Mechanical decladding such as has been used extensively for U02 fuel is also the preferred method for decladding stainless steel-clad Th02-U02 fuel. Mechanical decladding by shearing fuel bundles was used successfully at the West Valley plant of Nuclear Fuel Services, Inc., for decladding this type of fuel from the first core of the Indian Point 1 reactor. Mechanical decladding could also be used for zircaloy-clad fuel, but with complications in subsequent dissolving because of reactivity of zircaloy with the HNO3-HF reagent used in the dissolver.

Aluminum-alloy-clad fuel. With aluminum-alloy-clad fuel mechanical decladding, followed by selective dissolution of fissile and fertile material in nitric acid (such as was used for fuel clad with stainless steel), is not feasible because the aluminum alloy also is dissolved by nitric acid. Consequently, for aluminum alloy fuel, chemical decladding has been the preferred method. Two methods have been used for chemical removal of aluminum cladding from thoria-based fuel. In the Savannah River 233 U production campaign described by Rathvon et al. [R5], aluminum cladding was dissolved together with the Th02-U02 fuel in a mixture of nitric acid, hydrofluoric acid, and mercuric nitrate, the latter used to accelerate reaction of aluminum with nitric acid by intermediate formation of an aluminum amalgam. In the Hanford 233 U production campaign described by Jackson and Walser [J1 ], aluminum cladding was dissolved in a mixture of sodium hydroxide and sodium nitrate, which left undissolved the Th02-U02 fuel. After removal of most of the resulting solution of sodium aluminate, the solid Th02-U02 was dissolved in a mixture of nitric and hydrofluoric acids.

Fuel imbedded in graphite. Fuel for the HTGR under development in the United States was described briefly in Sec. 7.3 of Chap. 3. The preferred fuel [B20, R6] consists of hexagonal graphite blocks 79.3 cm high by 35.9 cm across flats, bearing longitudinal holes, some empty for coolant flow, others filled with fuel rods. The fuel rods consist of a graphite matrix in which are imbedded two kinds of microspheres. Thorium-bearing microspheres consist of Th02 with a two-layer “BISO” pyrolytic graphite coating. Uranium-bearing microspheres consist of uranium carbide UC2, with a three-layer “TRISO” coating consisting of two layers of pyrolytic graphite separated by a mechanically strong, combustion-resistant layer of silicon carbide.

After irradiation this fuel is declad by crushing the graphite blocks to free the fuel-bearing particles. The mixture of graphite and fuel particles is then burned with oxygen at 875°C in a fluidized bed [N7, Y1 ]. This removes the graphite matrix and the BISO graphite coating of the Th02 particles. The TRISO-coated particles retain their silicon carbide coating. The denser Th02 residue from the BISO particles is separated from the lighter silicon carbide-coated residue from the TRISO particles by elutriation with C02 gas. The irradiated Th02 from the BISO particles is dissolved in thorex dissolvent, НЫ03-НР-А1(Ы0з)з and reprocessed by the Thorex process.

The TRISO particles are crushed to break their silicon carbide coating The UC2 thus freed is then oxidized to U308, dissolved in nitric acid, and freed from fission products by the Purex process. More details of the two types of fuel particles and their treatment prior to reprocessing is given in Chap. 3, Sec. 7.3.

Other reactors of this type, such as the AVR and THTR reactors in Germany, use somewhat different fuel particles, such as mixed Th02-U02. However, the decladding procedure recommended still involves crushing the fuel, burning the graphite, and converting carbides to oxides.

A disadvantage of burning the graphite moderator is the larger volume of carbon dioxide, radioactive because of 5730-year 14 C, which must be removed completely from combustion

products, probably by absorption as CaC03 with Ca(OH)2 slurry [Cl6], and then stored as long-lived, low-level radioactive waste.

Attempts in the 1960s [B14] to avoid burning the graphite blocks were not successful. Even when the blocks were ground to under 200 mesh, the fuel was not completely leached from the graphite. And hydrocarbon reaction products of the carbides with nitric acid caused poor phase separation in later solvent extraction. Thus, all HTGR reprocessing systems now plan to convert fuel to U02 and Th02 before dissolution.

A reactor design in which a clean mechanical separation could be made of components bearing fuel nuclides and fission products from the bulk of the graphite moderator would obviate the need for burning the moderator and would reduce greatly the amount of C02 to be formed, absorbed, and stored as CaC03.

HLW solidification has to serve two purposes:

1. Immobilization of the waste

2. Long-term fixation of long-lived radionuclides

Immobilization of the waste helps to facilitate and to add safety to all handling and storage operations before the waste is in its final disposal location. Whereas liquid HLW is definitely

not suitable for transportation on an industrial scale, solidified waste may well be when its integrity can be maintained and appropriate shielding is provided. Interim storage of solidified waste presents fewer problems because of significantly reduced container corrosion and relaxed cooling requirements. The nature of the waste solidification product is of particular importance during the period of several decades while the repository is being filled, and hence open. In the long term, the basic requirement of the geologic disposal concept is a repository located deep underground under geologic conditions that ensure permanent isolation of radioactive waste material from the environment, regardless of its particular form. The inherent stability of the solidification product is considered an additional release barrier for radioactivity within the overall safety concept, as part of a system of engineered barriers including the container and, possibly, an overpack and highly absorbing backfill materials.

Product alternatives. The general requirement of a solidification product is stability against destructive influences to which a highly radioactive solid may be exposed, i. e., irradiation stability, thermal stability, mechanical stability, and chemical stability. Although stability of the waste form will not last forever and although it will be even impossible to verify long-lasting stability of the waste form, it is highly desirable as an additional safety barrier in the geologic waste repository concept. Consequently, the solidification product will have to be made as stable as practically achievable.

The more important product alternatives, classified in terms of two principal lines—calcine and glass—are shown in Table 11.6.

Calcines are products obtained by removing the volatile components of the waste, i. e., water and nitrate, at temperatures between 400 and 900° C. The result is a mixture of oxides of fission products, actinides, and corrosion products in particulate form with a specific surface of

0. 1 to 5 m2/g. The plain calcine is not very stable chemically because of its large surface area and the chemical properties of some of the oxides, and it is highly friable. To improve the properties of calcines, advanced forms are developed. One such product is the so-called multibarrier waste form, a composite consisting of calcine particles with inert coatings, such as pyrocarbon, silicon carbide, or aluminum, embedded in a metal matrix. Another advanced cal­cine is the so-called supercalcine. This is essentially a ceramic obtained by adding appropriate chemicals to the HLW to form refractory compounds of fission products and actinides when fired at 1200°C. Supercalcine requires consolidation by embedding in a matrix but does not need to be coated, as the material is supposed to have inherent chemical stability.

Glasses are products obtained by melting the waste oxides together with additives such as Si02, B203) A1203, P2Os, Na20, and CaO. On solidification, the melt forms a glass or a near-glassy solid with good stability. Borosilicate glass is the type of solidification product most

thoroughly studied all over the world. Phosphate glass has long been abandoned in the United States, where it had been studied for the first time. It was still considered in West Germany for quite a while, but has been given up there as well. Only in the Soviet Union is work on phosphate glass kept alive.

Glasses also have certain drawbacks, such as the possibility of devitrification leading to products with less predictable properties. Advanced developments along the glass line are glass ceramics obtained by controlled crystallization of glass to avoid uncontrolled devitrification. Another advanced product is vitromet, i. e., glass or glass ceramic beads embedded in a metal matrix, with extremely high heat conductivity and mechanical strength.

It is the common principle of supercalcine and glass ceramic to have stable crystalline phases hosting fission products and actinides. Along the same line, a third group of crystalline solidification products, synrock (synthetic rock), has been developed. All crystalline waste forms contain significant quantities of glassy phases remaining from their formation at high temperatures.

Fission-product content. The solidification products may incorporate different fractions of fission-product oxides. This fraction is desired to be high on economic grounds. Costs of handling, packaging, and transportation are considerable and depend on the volume and the number of containers to be handled. On the other hand, the fission-product concentration is limited by chemical reasons or by reason of heat production. Chemical limitations are typical for glasses where either phase separations may occur or the product may not be a glass at all. For borosilicate glasses, 20 to 25 w/o (weight percent) of fission products is about an upper limit. Higher concentration may lead to the segregation of a yellow crystalline phase composed of alkaline and alkaline earth molybdates. This easily soluble phase contains long-lived fission products such as 90 Sr and 137Cs.

No chemical limitations exist for calcines, which may consist of pure waste oxides. In this case, however, the heat production in the solid may impose a limitation depending on the burnup and the fraction of nonradioactive oxides in the waste.

Making a composite means a further reduction of the fission-product concentration in the final product. If small particles of glass or calcine are embedded in a close-packed matrix, the fission-product content is reduced by about one-third.

Recently, there has been a trend to give less priority to a high-fission-product content and to a small product volume. Economic penalties may be compensated by benefits due to more flexibility in handling, interim storage, and disposal of the waste products.

Irradiation stability. Any solidified HLW will be exposed to energetic radiation from radioactive decay of fission products and actinides. Part of the radiation energy is dissipated in elastic collisions with atoms from the solid material, thereby displacing them and causing radiation damage. This may affect macroscopic properties such as mechanical or chemical ones,’ and it may cause storage of energy.

The following types of radiation will occur in the waste:

Gamma radiation (average energy 2 MeV) and beta radiation (1.5 MeV) from fission-product decay

Alpha radiation (6 MeV) from actinide decay and some alpha radiation from (n, a) reactions (e. g., with boron)

Recoils from alpha decay (100 keV)

Fast neutrons from spontaneous fissions and from (а, л) reactions with light elements, and fission recoils

The energy dissipated in elastic collision is at least two orders of magnitude lower for gamma and beta radiation than for the others. The total fast-neutron and fission recoil doses

are by several orders of magnitude lower than for alpha radiations and their recoils. Thus, mainly alpha particles and even more their recoils have to be considered. Their effect can be simulated on a reduced time scale by incorporation of an appropriate amount of Cm203 (mainly 244 Cm) into a synthetic waste solid. The alpha-radiation rate of Cm203 is about 1 X 1014 min-1*g-1. Figure 11.6 shows the total alpha dose per metric ton of heavy metal reprocessed. The first 100 years are considered most significant for radiation damage, as the total alpha dose increases only by a factor of 10 over the next 100,000 years. The 100-year alpha and alpha-recoil dose can be simulated during 1 year with a Cm203 content of 1 w/o.

Among the possible consequences of radiation damage on solidified waste, energy storage has to be considered a potential risk. The temperature of the waste solid would suddenly rise if stored energy were released. A quite thorough experimental study by Roberts et al. [R2] on energy storage in calcines and borosilicate glass comes to the conclusion that there will be hardly more than 50 cal/g stored. With an average heat capacity of 0.2 cal/(°Og), this corresponds to a maximum temperature rise of 250°C, which should be tolerable.

Other possible effects of radiation on the solidified waste are deterioration of mechanical properties and changes of volume due to radiation damage or as a consequence of helium formation from alpha decay.

Furthermore, one may think of radiation influencing the chemical stability of the solid. There is no experimental evidence for any of these effects [H2]. In all test procedures a long-term dose has been simulated in a short time, which will probably rather enhance the effects. The apparent radiation stability should be not too surprising, bearing in mind that an alpha dose of 1018/cm3 over 100 years is a rather modest one compared to the fast-neutron doses to which materials in a nuclear reactor are exposed.

Thermal stability. Heat generation in the solidified waste causes it to be at an elevated temperature for more than 100 years. The specific heat generation in a solid with 20 w/o fission products is shown in Fig. 11.7. A cylindrical waste block being a homogenous heat source will have a radial temperature gradient. Given the heat generation, the temperature difference between the surface (the surface temperature is determined by the storage condi­tions) and the centerline is a function of the heat conductivity of the material. The maximum temperature difference in the waste cylinder is

In this equation q is the homogenous thermal power density (W/m3), d is the diameter of the waste cylinder (m), and к is the thermal conductivity [W/(m,0C)].

Equation (11.1) is essentially a solution of Eq. (11.7) and is based on a few assumptions and simplifications, e. g., no axial heat conduction, constant average heat conductivity and specific heat, constant heat source, steady-state heat transfer, one-dimensional (radial) heat flux, cylindrical geometry in the waste and in the surrounding material, e. g., salt, and no heat source in the salt.

Table 11.7 shows the thermal conductivities of several solidification products. In Fig. 11.8 maximum temperature differences are plotted against the age of the waste assuming 20 w/o fission-product oxides in the solid. For the calculations typical canister diameters of about 250 and 500 mm and typical thermal conductivities of 0.25 (particulate calcine), 1.2 (glass), and 10 (vitromet) W/(m*°C) have been used [E2].

When a 6-year storage period prior to solidification is assumed, glass in canisters 25 cm in diameter and glass-metal composites in canisters 50 cm in diameter exhibit maximum temperature differences well below 100°C. However, glass in a large canisteT and plain calcine even in a small canister will produce centerline temperatures as much as 300°C higher than the surface temperature, thus possibly reaching 500 or 600°C. After about 200 years the temperature gradient will have essentially disappeared.

One limitation of the maximum temperature in a solidifled-waste block is given by the need to maintain the immobilization of the waste. Table 11.8 shows softening and melting temperatures of various products. For high-melting materials volatilization of individual fission products below the melting temperature has to be considered.

A long-term effect promoted by high temperature is devitrification of glass, converting it

into a thermodynamically more stable form. This effect is supported by the presence of a great number of components and of impurities that may act as crystallization nuclei. Both conditions are present in a waste glass. The crystallization process in a multicomponent borosilicate glass with simulated fission-product oxides has been quite thoroughly investigated [H4]. It was found that partial crystallization may occur within days at temperatures above 600°C depending on the composition. Fission products are selectively enriched in certain crystalline phases. The remaining glass phase still containing fission products is enriched in boron oxide. The devitrified product may therefore be less leach-resistant than the original glass.

With full radioactivity, phosphate glasses showed strong devitrification at 500°C with deterioration of leach resistance. Under the same conditions borosilicate glasses did not devitrify within 7 months [B3]. Recent investigations show that extensive additional tests under hydrothermal conditions are required to simulate underground storage conditions.

Figure 11.8 Maximum temperature difference in a cylinder of solidified waste for different diameters and thermal conductivities (70 liters/MT of heavy metals; 30,000 MWd/MT, 150 days aged prior to reprocessing).

Ceramic-type products such as supercalcine and glass ceramics have been subject to a crystallization process and are therefore thermodynamically more stable. Consequently, as long-term structural changes are less likely with those solidification products, they may maintain their original properties at a higher temperature than glasses.

The temperature gradient will give rise to thermal stresses in any monolithic material, which in turn may cause cracking. The stress at any point of a glass sample depends on the difference in temperature between this point and the average. At the surface of a cylinder, where cracking is most likely, the longitudinal and the circumferential stresses are given by [Kl]

Fn

°=J^(Ta-Ts) (H.2)

with a = stress (kg/cm2), E — modulus of elasticity (kg/cm2), a = linear expansion coefficient (1/°C), ц. = Poisson’s ratio, Ta = average temperature (°С), and Ts = surface temperature (°С). The maximum difference between surface and average temperature beyond which stress cracking is to be expected may be roughly estimated as a function of the expansion coefficient by inserting E = 7X 10s kg/cm2, ц. = 0.20, and a fracture stress of 1000 kg/cm2, which are reasonable for an average soft glass. Then this temperature difference limit is

Аги™, <* З X 10-4 — (11.2a)

a

Table 11.9 shows expansion coefficients of certain solidification products and the corresponding temperature-difference limit. However, even if the stationary temperature differ-

ence could be kept below these limits, there may be severe stresses due to fast cooling of the glass.

Thermal expansion has to be looked on under one more aspect. Different expansion coefficients of glass and canister material may cause stress in the canister wall. It has been observed that this may significantly promote corrosion of the canister.

Mechanical stability. When a block of solidified waste is crushed by mechanical impact, fragments of various size will be formed. Two consequences are to be considered: (1) The fraction of radioactivity leached in a certain period of time will be increased in proportion to the increase in surface area; (2) the formation of very small particles in the order of 100 Aim and less may enable radioactive material to be spread by air. Even larger particles may be carried by water.

Nonmonolithic calcines do not have any mechanical stability, metal matrix products are the most stable ones. Glass ceramics are more stable than glasses.

Chemical stability. The only chemical attack on the solidification products deserving serious consideration is leaching by water or brine, if such exist in the repository. Chemical interaction between solid rock salt or other geologic material and any of the solidification products under consideration will not be significant unless the temperature rises above the melting point of the salt or the solidification product. Diffusion of fission products into the salt at reasonable temperatures is not a significant safety concern either.

As leaching is the most likely mechanism by which radionuclides from the waste may be remobilized, the leaching behavior is the most thoroughly studied property of solidification products. As a consequence, there is a wide spectrum of procedures and results. The samples are powders, grains, or small blocks. The leaching procedures are characterized by different temperatures, different leach liquors-e. g., pure water, seawater, saturated NaCl solutions-and different renewal schemes for the liquor. The latter is important for the result, as the average leach rate is found the smaller the longer the liquor remains in contact with the sample. With a soxhlet-type apparatus continuous renewal may be achieved. In the majority of the experiments an integral leaching rate R [g/(cm2-day)] is determined in order to compare the leach resistance of different solidification products. In addition, a number of experiments have been designed to study the time dependence of the leaching process. They provide the information necessary to calculate the total radioactivity release upon accidental contact with water over a long period of time. To have a sound basis for extrapolation to the leach behavior of a large glass cylinder over a time period relevant for long-term disposal considerations, those experiments have to employ a suitable leach technique, e. g., leaching of fine powders. In general, the average leach rate observed decreases with increasing duration of the leach experiment. Most experimental leach rates are obtained from leach processes lasting for days or weeks.

Leach rates measured on borosilicate glasses at room temperature are in the range of 10"6 to 10"4 g/(cm2 ’day), largely depending on experimental conditions. Almost identical results were obtained in some simultaneous tests with water and NaCl solution. A careful study of the spectrum of data and procedures leads to the conclusion that 10~s g/(cm2 • day) is a reasonable weighted average suitable to characterize the leach resistance of borosilicate glasses in water at room temperature and atmospheric pressure [B7, Т1]. Leach rates of phosphate glasses are within the range of borosilicate glasses. There is experimental evidence that devitrification in­creases the leaching rates, more in the case of phosphate glass than in the case of borosilicate glass.

For borosilicate glass ceramics, leaching rates of the order of 1СГ5 g/(cm2,day) and greater have been found. Remarkably, the controlled crystallization process used in making these ceramics does not increase the leachability relative to that of the parent glass, in contast to the increased leachability occurring on spontaneous devitrification.

Calcines are well known to be readily leachable in water. Hot pressed supercalcine shows a leach resistance similar to that of the other chemically stable products.

In general, it may be concluded that a leach rate in the range of 10“6 g/(cm2 *day) is probably something like a lower limit unless a substantially different solidification technology is employed. Less leachable products may be high-temperature glasses or ceramics or very sophisticated composites.

It should be mentioned that leach data presently available have been obtained under standard laboratory conditions. Leaching experiments modeling conditions that may be experienced in a specific type of repository, including solutes in the leach liquor and elevated temperature and pressure, have been initiated in a number of laboratories.

Extrapolation of long-term leaching. An important task in evaluating the risk of radioactivity release by water leaching is the mathematical description of the process. All efforts to do this on the basis of a physical understanding of the process have not led very far. This leaves us with empirical approaches where much of the physics of the process is packed into coefficients and exponents obtained by fitting experimental leach curves.

It is reasonable to assume two limiting cases of leaching kinetics, dissolution of the waste form and diffusion of radioactive species from the waste form, representing upper and lower limits of the fraction leached within a given time period. The corresponding types of equation are (F = fractional release)

Fi=fl, f (11.3)

and F2 = a2/t (11.4)

Geometric changes of the sample on leaching have not been taken into account. Equation

(11.4) is the well-known type of equation approximating a solution of Fick’s second law for a semiinfinite sample valid for fractional releases up to about 25 percent.

These are oversimplifications very likely indicating upper and lower limits of radioactivity release. Essentially the complex process will be composed of both dissolution and diffusion.

This can be described by equations of the type

F3 = a3/t + a33t (11.5)

This type of equation may work well for relatively short periods. For long-term leaching, however, it assumes that corrosion becomes more and more dominating and eventually controls the process. Therefore an entirely empirical approach with the following type of equation appears to be the most appropriate:

Ft, = a^tx (1 > x > 0.5) (11.6)

There are examples where test runs can be best fitted with jc = |. None of these simple equations takes into account that a piece of solidified waste will become smaller on leaching and so will the surface area. Although this simplification is on the safe side, the effect will be small for a full-size glass cylinder within the time period of interest.

Figure 11.9 shows the results of sample calculations of cumulated fractional releases from a 1.65-m-Iong and 0.234-m-diameter waste cylinder over a period of 106 years taking into account decay of radioactivity. They are obtained by fitting Eqs. (11.3), (11.4), and (11.6) to experimental data of sodium-leaching from a 144-day column leach experiment with a powdered borosilicate glass (1236 cm2/g specific surface), recalculation to the specific surface of the glass block and coupling with the ORIGEN program for LWR uranium waste from

30,0 MWd/MT burnup fuel [Е1] . The total fraction of sodium leached in the experiment was about 30 percent. The constants a„ in Eqs. (11.3), (11.4), and (11.6) include a factor a’„ from fitting experimental data and a geometric conversion factor. They are

аі = e’jScy, = 4.45 X КГ5 (yr1) a2 = a, = 2.53 X 1CTS (yr’1/2)

^sample

aA = dA -^У1 . = 3.24 X КГ5 (yr-*)

^sample

Scyi = 6.53 X 1СГ2 cm2/g (specific surface of waste cylinder)

Sample = 1.24 X 103 cm2/g (specific surface of leach sample)

x = 0.666

The curves indicate the fraction of the radioactivity present in the waste at the time of reprocessing that will be in solution if the glass block has come into contact with water 10, 65, or 300 years after reprocessing. An increase with time means that the leaching process is faster than the overall decay of radioactivity and vice versa. Thus, the curves represent the fraction of the initial radioactivity available for release into the environment at any time when the geologic barrier may fail. For comparison, the top curve of Fig. 11.9 shows the total fraction of the initial radioactivity available at any time. The plot thereby demonstrates that the solidification products presently envisaged for final disposal of highly radioactive waste may in fact provide an effective release barrier.

Figure 11.9 is an example of how to extrapolate leaching data. It should be noted that this extrapolation is based on experimental data not necessarily representative of conditions expected in a geologic repository. In general such extrapolations have considerable uncertainty. They are based on the assumption that the basic properties of the glass are not significantly altered in a period of thousands of years.

Product evaluation. Table 11.10 lists the stability criteria of solidification products. The stability categories are of different relevance. Categories 3 and 4 are directly relevant for the release-barrier function of the solidification product. Radioactivity release from the final disposal site may occur by leaching or, less likely, via spreading particulate matter by air or water. Mechanical effects may also alter the leachable surface of the waste form. Categories 1 and 2 are not directly relevant for the release of radioactivity but for maintaining the original properties of the solidification product, which are to prevent radioactivity from being released.

The radiation dose to be expected is moderate and smeared out over a long period of time. Serious effects on stability are unlikely, and the experimental data presently available do not indicate such effects. Properties relating to thermal stability have been characterized in some detail. They are considered important for maintaining the product stability over intermediate periods of time.

In the final evaluation, product technology has to be considered as well. When the point is to select a product for near-future demand, technology will have a high weight. It will be better to have the waste solidified in an acceptable form even if not in the very best one. One has to keep in mind that the amount of highly radioactive waste to be solidified over the next two decades will be a small fraction of the waste that nuclear energy will produce altogether. In the long run, however, stability criteria should clearly dominate.

It should be noted that the canister may add considerably to the durability of a high-level radioactive waste package. Often it may be easier to achieve a high degree of sophistication with a nonradioactive canister material than with a highly radioactive solidification product.

Figure 11.10 is a semiquantitative approach to a rating of solidification products according to stability and technological simplicity [E2].

Policy considerations. Calcines as such do not seem suitable as final solidification products. They do not have any favorable properties with regard to mechanical and chemical stability, which are the categories of primary relevance for the release of radioactivity.

Coating and matrixing the calcine particles may overcome the stability drawbacks but will replace them by severe technological difficulties. Calcine may only be discussed as a nonfmal solidification product for interim storage provided that the safety concept does not require inherent stability of the product. In this case it may be an advantage that the calcine can be converted into a final product with some advanced technology possibly available at a later time.

Glass is among the solidification products with the highest mechanical, chemical, and irradiation stabilities presently known. As far as thermal stability of glass is concerned, the possibility of devitrification and the not very high softening temperature are disadvantages causing some uncertainty about its long-term performance. Glass solidification technology is in an advanced state of development and, in fact, is already available on a technical scale. Considering properties and the state of technology, glasses are presently the first choice for HLW solidification.

A ceramic product, thermodynamically more stable than a glass, will reduce the uncertainty about the state of solidified waste after some time of storage or disposal. Among alternative ceramic products, glass ceramic probably requires only slight modifications of the borosilicate glass process, whereas the supercalcine process will be quite different from ordinary calcination processes. It is therefore not unlikely that glass ceramics offer a chance to achieve with reasonable effort a ceramic product with at least the same chemical and mechanical stability as the parent glass. The technological penalty for any ceramic product with tailored crystalline host phases is a reduced flexibility toward the chemical composition of the waste.

The German/Eurochemic metal-matrix process PAMELA, originally developed for phos­phate glass particles, is also suitable for borosilicate glass and for glass ceramic. It will provide a
product of excellent impact stability and extremely high thermal conductivity. The latter may be useful if less aging time prior to solidification or cylinders of larger diameters are desired. Drawbacks are the lower overall melting point of the product due to the metal, uncertainties about the long-term chemical stability of the metal, and the relatively complex technology.

Compared to well-characterized solidification products carefully designed for final storage, unreprocessed spent fuel elements are a less well-defined waste form. Fuel elements are designed for operation in nuclear reactors rather than for final storage. Very little is known about their stability in final storage. However, they have at least two disadvantages: Spent fuel contains radioactive gases at rather high pressure, and it has been damaged by radiation to quite an extent. There is no doubt that spent fuel that is to be disposed of needs some processing before final storage, such as additional canning, to make it suitable for disposal.

Complexing agents may be added to an extraction system either to increase or to decrease the distribution coefficient of a metallic component between aqueous solution and kerosene. For example, Zr(N03)4 is not extracted at all from aqueous solution by kerosene. However, when TBP is added to the kerosene, a compound of probable composition [H4] Zr(N03)4’2TBP is formed that is readily extracted. As an example of a complexing agent that reduces extraction, the effect of adding fluoride ion to the above system may be considered. The more stable, inextractable complex ion ZrF62 " is then formed according to the reaction

Zr(N03)4-2TBP + 6F“ *= ZrF62′ + 4N03" + 2TBP

and in the competition between the fluoride ion and TBP for the zirconium, the distribution coefficient of the zirconium is reduced.

Uranium hydride UH3 is made by reacting uranium metal with hydrogen at temperatures above 250°C and pressure above the dissociation pressure, pHj :

4480

log. o Ph, (Torr) = — + 9.20 (5.3)

A solid solution of uranium and zirconium hydrides is used as fuel in TRIGA reactors. Uranium hydride is often pyrophoric and must be handled with care. It has been used to prepare finely divided uranium metal by reacting massive metal with hydrogen, then crushing the brittle hydride and heating it in vacuum to drive off hydrogen.

1.10 Uranium Halides

Table 5.9 lists uranium halides together with some of their more significant properties.

UF4 is an important intermediate in the production of UF6 and uranium metal. It is made by reacting U02 with an excess of HF vapor,

U02 + 4HF — UF4 + 2H20

as described in more detail in Sec. 9.5. Dissolved in a low-melting eutectic of ZrF4, BeF2, and 7LiF, 235 UF4 was used as fuel in the Molten Salt Reactor Experiment [НЗ].

UF6 is the only compound of uranium volatile at room temperature. It is used as working fluid in the gaseous diffusion, gas centrifuge, and aerodynamic processes for uranium enrichment discussed in Chap. 14. Its principal physical and chemical properties are summarized in Sec. 4.7.

UCU, which boils without decomposition at 791 °С, was used as feed material for the Y-12 electromagnetic uranium enrichment plant. It is hygroscopic and hydrolyzes in moist air.

The other uranium halides have not had significant practical uses.

1.2 Uses

Because the thorium atom density is higher in thorium metal than in any thorium compound, metal is the preferred form of thorium where the highest nuclear reactivity or highest density is wanted. One likely nuclear application is in a sodium-cooled fast reactor where thorium would capture a neutron and be converted to 233 U.

1.3 Phases

The phases of thorium metal and their transition temperatures are listed in Table 6.4. Equations for the vapor pressure of thorium metal are [11 ]

1.4 Density and Thermal Expansion

The theoretical density of thorium at 25°C, from x-ray crystal measurements, is 11.72 g/cm3. The density of cast metal is between 11.5 and 11.6 g/cm3.

The mean thermal expansion coefficient is given in Table 6.5. Because thorium crystallizes in the cubic system, it expands equally in all directions and is not subject to as much distortion on thermal cycling as uranium. For this reason, and because the a-0 transition temperature in thorium is much higher than in uranium, thorium metal reactor fuel has much better limensional stability than uranium metal.

Source’. International Atomic Energy Agency, “Thorium: Physicochemical Properites of Its Compounds and Alloys,” Atomic Energy Rev., Special Issue No. 5, 1975.

Chlorination of zircon has been the process mainly used in the United States because it produces ZrCLt, which is used in the Kroll process for making zirconium metal (Sec. 8.3), and because ZrCU was the feed material for the first process developed for separating hafnium from zirconium, using thiocyanate extraction (Sec. 7.3).

In the early zircon chlorination plants such as used by W. J. Kroll [КЗ] at Albany, Oregon, zircon was first converted to zirconium carbide by reaction with graphite in a graphite-lined arc furnace at 1800°C:

ZrSi04 + 4C -» ZrC + SiO + 3CO

The silicon monoxide, volatile at 1800°C, distilled off. The ZrC was then converted to ZrCU by chlorination at 500°C:

ZrC + 2C12 ZrCU + C

In the newer plants, such as the Wah Chang plant, a mixture of zircon and carbon is chlorinated at 1200°C, to produce ZrCLt in a single step:

ZrSi04 + 4C + 4C12 -»■ ZrCU + SiCU + 4CO

This process has the advantages of operating at lower temperature and of converting silicon to a useful by-product (SiCU) instead of to a troublesome airborne contaminant (SiO). The reaction is endothermic and requires good thermal contact between carbon and zircon. A fluid-bed reactor is used, and energy is provided either chemically by addition of silicon carbide or physically by electrical resistance heating of the bed.

The principal steps in this direct chlorination process for converting zircon to ZrCU are shown in Fig. 7.4. Gases from the chlorinating furnace are cooled to around 100°C to condense crude solid ZrCU and Fed3, then cooled further to condense SiCU, TiCU, and A1C13. The crude ZrCU is purified by sublimation with hydrogen in a stainless steel retort. Hydrogen reduces volatile FeCl3 to nonvolatile FeCl2, which remains in the retort with Zr02 and other nonvolatile impurities. This process removes most of the metals associated with zirconium in zircon except hafnium.

The tendency of positive ions to form stable complexes with anions, eg., ThF3*, ThF22+, PuN033+, PuF62′ in aqueous solution, is analogous to the tendency toward hydrolytic behavior.

In an aqueous solution of uranyl nitrate there are present not only hydrated uranyl ions U022+, but also a series of nitrate complexes U02N03+, U02 (N03 )2, and U02 (N03 )3“. At room temperature there is rapid equilibrium between these complexed species. The formation of the high-nitrate complexes is promoted by the presence of nitrate ions; thus U02N03+ may pre­dominate in solutions 4 M in nitrate concentrations, and U02(N03)2 in more concentrated nitrate solutions [F3].

The relative tendencies toward hydrolysis mentioned in Sec. 1.2 above apply generally to complex formation, except for the reversal of the order of M3* and M02 2+, acording to [K2]

M02* < M022+ < M3+ < M4*

For example, the tripositive ions such as La3* or Pu3+ show relatively little tendency to complex, but stable complexes are formed with the tetrapositive ions M4+ and with hydrated ions of hexapositive metals M022+. The stability of the complex depends also on the properties of the complexing anion; strongest complexes usually occur with anions of weak acids, or small atomic radius, and with large negative charge. The tendency of anions to form complexes increases approximately in the order

CKV < Cl’ < no3- < S<V < PO43- < Г

The presence of strongly complexing anions in solution, such as F~, inhibits hydrolysis.

Cations that complex easily generally form stable complexes with oxygenated organic compounds, such as diethyl ether, methyl isobutyl ketone, and tributyl phosphate (TBP). The purification of uranium by solvent extraction of hexavalent uranium from nitrate solutions, with TBP forming и02(Ж)з)2 ’2TBP, was described in Chap. 5. These metals are extracted most easily from aqueous solutions free of the more highly complexing anions, such as F ", P043", or SO42~.

From the above discussion it follows that tetravalent and hexavalent thorium, uranium, and plutonium can be separated from the trivalent rare-earth fission products by taking advantage of differences in complexing properties. More highly charged cation fission products, such as tetravalent cerium and the fifth-period transition elements zirconium, niobium, molybdenum, technetium, and ruthenium, complex more easily than the trivalent rare-earths and are more difficult to separate from uranium and plutonium by processes involving complex formation.

The tendency toward hydrolysis of some of these elements can be used to advantage in separation processes. For example, in the Redox process for separating uranium and plutonium from fission products, the aqueous feed to the separation plant is made acid-deficient to promote hydrolysis of zirconium to a less extractable species, probably a colloidal hydrate [B5].

Separation by ion exchange also involves tendency toward hydrolytic behavior and complex formation. A cationic resin in the acid form will exchange hydrogen ions for и-valent ions, Mn+, in aqueous solution according to

nHR^resin) M^soln) Ч — MR„ (resin) "i* flff(sOln) (9*6)

where R represents the insoluble resin group.

For an anionic resin in the hydroxyl form the “adsorption” of и-valent anions A"" from aqueous solution is given by

hROH(resin) + Ajsoin) RnA(resin) + n(OH)(‘sotn) (9.7)

For the rare-earths and actinides the relative tendencies toward adsorption on the exchange resin are roughly in the same order as the relative tendencies toward hydrolysis and complex formation, as described above. For example, for the trivalent lanthanides the atomic radius increases with atomic number, and the smaller-radii elements like lanthanum can be adsorbed preferentially from the trivalent elements of high atomic number in this series. The same applies to the separability of trivalent actinide elements, where trivalent elements of higher atomic number adsorb less easily. The adsorbed cations are easily removed from the resin by adjusting the solution acidity and/or eluting with a solution containing a strongly complexing anion, such as citric acid.

Tetravalent plutonium can be adsorbed preferentially from most fission products on cation-exchange resins. Eluting with a solution more acidic than that from which the plutonium was adsorbed removes plutonium from the resin because the higher hydrogen ion concentration displaces the equilibrium of reaction (9.6) and the higher anion concentration, such as N03", tends to complex plutonium in solution. This provides a means of concentrating plutonium in solution as well as decontamination. Separation can occur with elution by proper choice of the complexing nature of the eluent. For example, uranium can be removed preferentially from a cation-exchange resin on which Pu4* and U02 2+ have been sorbed by eluting with dilute sulfuric acid solution.

The negatively charged complexes that result with many of the fission products and the actinides in solutions of high concentrations of complexing anion are easily adsorbed on anion-exchange resins. Such adsorption occurs readily with tetravalent and hexavalent uranium, with tetravalent plutonium, and with the fission products that complex easily, such as zirconium, niobium, and ruthenium. Because the extent to which negative complexes are formed for a given metal in solution depends on the concentration of the complexing anion, control of the anion concentration provides a sensitive means of controlling the exchange equilibria for a given metal and for affecting partition between two or more metals in solution. Extensive studies of the adsorption of a wide variety of elements by means of anion exchange have been reported by Kraus and Nelson [K4].

In the 1950s and early 1960s, when metal fuel was still being considered for power reactors, many pyrometallurgical processes were proposed and some were carried through pilot-plant demonstration. Subsequently, when it was realized that the susceptibility of metal fuel to radiation damage limited its usefulness in power reactors, these processes became of less practical importance.

Table 10.1 lists the principal types of pyrometallurgical processes on which experimental work has been conducted. These have been grouped into physical separation processes, in which no chemical reactions take place, and chemical separation processes.

Physical separations Volatilization or distillation Fractional crystallization Extraction with liquid metals Chemical separations Liquid extraction with fused salts From molten fuel From liquid-metal solution of fuel Electrolysis through fused salts Partial oxidation

Volatilization. Many fission-product elements, including krypton, xenon, iodine, cesium (normal boiling point 705°C), strontium (1380°C), barium (1500°C), the rare earths (3200°C), and plutonium (3235°C), are more volatile than uranium (3813°C). Cubicciotti [C17], McKenzie [M5], and Motta [M8], in laboratory experiments, showed that around 99 percent of these more volatile elements could be separated from uranium by vacuum distillation at 1700°C. Because of the high temperature and severe materials problems, volatilization has not been used as a primary separation process, but does contribute to removal of the most volatile fission products in conventional reprocessing. In fractional crystallization or extraction with liquid metals, distillation is used to separate uranium and plutonium from more volatile solvent metals.

Fractional crystallization. Volatile metals with much lower boiling points than uranium, such as magnesium (1103°C), zinc (906°C), and cadmium (767°C), have been extensively studied as solvents for separating constituents of irradiated metal fuel by fractional crystallization, followed by evaporation of the solvent metal from the separated fractions. For example, in liquid magnesium, the solubility of plutonium or thorium is high, but uranium is very low. A process of this type was developed at Argonne National Laboratory [P6] for concentrating plutonium in the uranium metal blanket of a breeder reactor from 1 percent to 40 percent.

Liquid extraction with metals. By operating at temperatures above the melting point of uranium (1133°C), liquid metals partially miscible with uranium, such as magnesium [B3], silver [V3], (m. p. 960°C), and calcium [M4] (m. p. 1482°C), have been used to separate uranium, plutonium, and some fission products. These processes have not proved attractive. The vapor pressure of magnesium at the uranium melting point is too high. The high boiling point of silver (2212°C) makes its separation from uranium difficult. Calcium is so reactive that no suitable container material is available. Another disadvantage of liquid extraction is that metals less extractable than uranium remain in the uranium metal phase.

Liquid extraction with fused salts. Liquid extraction of metal fuels is made more flexible by use of fused salt extractants. The distribution coefficient of elements into the salt phase can be increased by adding to the salt a compound more readily reduced than the corresponding compound of the metal to be extracted. For example, addition of ZnCl2 to MgCl2 increases the distribution coefficient of uranium into the salt phase because the reaction

2U + 3ZnCl2 2UC13 + 3Zn is more strongly displaced to the right than

2U + 3MgCl2 -*■ 2UC13 + 3Mg

Distribution coefficients may be further modified and operating temperatures reduced by dissolving uranium fuel in a low-melting metal such as bismuth or zinc. Separation of uranium from fission products by liquid extraction between molten bismuth and fused chlorides was extensively studied at Brookhaven National Laboratory [D5] in connection with the liquid — metal fuel reactor (LMFR), which used a dilute solution of 235 U in bismuth as fuel. Extraction of fission products from molten plutonium by fused chlorides was studied at Los Alamos [L2] in connection with the LAMPRE reactor.

Workers at Argonne [BIO] extended chloride extraction to higher-melting uranium alloys such as the 20 percent Pu-80 percent U alloy proposed for breeder-reactor cores. In a modified process [A9], this alloy was dissolved in molten zinc and contacted first with a low-melting LiCl-NaCl-MgCl2 salt phase containing sufficient ZnClj to transfer uranium, plutonium, and the more reactive fission products to the salt phase. The salt was then contacted with a cadmium-zinc alloy containing sufficient magnesium to return uranium and plutonium, but not the fission products, to the metal phase, from which cadmium, zinc, and magnesium were finally distilled.

Oak Ridge National Laboratory [R9] has studied liquid extraction with bismuth containing controlled amounts of lithium as a process for removing UF3 and PaF4 from LiF, BeF2, ThF4, and fission-product fluorides in fuel from the molten-salt reactor.

Electrolysis through fused salts. Electrodeposition of metal fuel through a fused-salt electrolyte to separate uranium and plutonium from fission products was studied at Knolls Atomic Power Laboratory [N6].

Partial oxidation. Spent fuel from the core of the EBR-II reactor was reprocessed [H7] by melting the fuel in a Zr02-lined crucible at 1400°C for from 1 to 3 h, after which the

remaining metal was poured into Vycor glass molds for refabrication into fuel pins. Volatile fission products were vaporized from the fuel, and rare earths, strontium, and other metals more reactive than uranium, plus 5 to 10 percent of the uranium, were oxidized by the Zr02 liner and remained in the crucible. This process, called melt refining by Argonne, operated from September 1964 through February 1969 and recycled 2300 kg of irradiated metal whose burnup ranged from 1.0 to 1.3 percent. Most of the fission-product metals less reactive than uranium, such as zirconium, niobium, molybdenum, ruthenium, rhodium, and palladium, remained in the metal. Known collectively as fissium, these were found to improve the stability of the recycled ahoy to radiation.