M. SALVATORES, SeniorScientificAdvisor, Nuclear Energy Directorate CEA (France) and Idaho National Laboratory, USA

Abstract Partitioning and Transmutation (P&T) is considered as a way of reducing the burden on geological disposal, required to isolate long-lived isotopes (present in the spent fuel) from the biosphere in stable deep geological formations for long periods of time. In this chapter, the physics of transmutation and the type of reactors where TRU transmutation can be performed together with basic feasibility issues will be discussed. Key issues of any transmutation strategy are the impact on the fuel cycle (specific features required for major fuel cycle installations, doses experience by workers, etc.), the development of appropriate fuels and of chemical separation processes will be indicated. Possible fuel cycle implementation scenarios, depending on specific country policies, and the potential benefits of geological repositories will be discussed.

Key words: transuranics, waste management, radiotoxicity, fast neutron spectrum, geological disposal.

17.1 Introduction

I n commercial water-cooled, reactors (LWRs and CANDUs), plutonium (Pu) and the so-called minor actinides (neptunium (Np), americium (Am), curium (Cm), etc.), collectively called the transuranic (TRU) elements, are formed through neutron capture by nuclides in the fuel. The capture of a neutron by U-238 without a fission reaction results, finally, in the creation of Pu-239 (see Fig. 17.1). Neptunium-237 is produced primarily from neutron capture in U-235 and subsequent nuclear reaction and decay. From Pu-239, neutron captures lead to the creation of heavier isotopes that, coupled with subsequent decay, create key gateway isotopes (e. g., Pu-241, Pu-242, Am-243). Neutron interactions with these result in the formation of higher mass actinides (i. e., Cm, Bk, Cf). Thus, neutron absorptions, not leading to fission, transmute Pu-241 into the non-fissile Pu-242, which in turn is the gateway to Am-243. Neutron capture in Am-243 leads to the production of Cm-244, which may decay to Pu-240 by alpha emission, spontaneous fission or transmute to Cm-245 by neutron capture. In general, depending on the relative rates of these three processes, an actinide will either reach a equilibrium concentration or else increase monotonically with burn-up.

Figure 17.1 summarizes the uranium nuclei transmutation chain under neutron irradiation and the associated Bateman equations, where n is the nuclide j density, oaj is the absorption cross section of isotope j, o1 K is the cross section corresponding to the production of isotope K from isotope j, A is the decay constant for isotope j, AK is the decay constant for the decay of isotope j to isotope K and, finally, Ф is the neutron flux.

The management of spent fuel is a major challenge for all countries where nuclear energy has been developed, regardless of the perspective applied to its future utilization, from further development to progressive phase out.

Most of the enduring hazard from spent fuel stems from only a few chemical elements — plutonium, neptunium, americium, curium (see Table 17.1) — and some long-lived fission products such as iodine and technetium, present in the spent fuel at concentration levels of kilograms per ton. These radioactive by-products, although present at relatively low concentrations in the spent fuel, are a hazard to life forms when released into the environment. As such, their disposal requires isolation from the biosphere in stable deep geological formations for long periods of time.

Partitioning and transmutation (P&T) is a way of reducing the burden on geological disposal. Since plutonium and the minor actinides are mainly responsible for the long-term radiotoxicity (as expressed by its effective dose coefficient, see e. g. Ref. 1), when these nuclides are removed from the waste (partitioning) and fissioned (transmutation), the remaining waste loses most of its long-term radiotoxicity. Moreover, the P&T strategy may also allow a reduction in the amount of heat generated by radioactive waste and this can have a significant impact on the repository size (and therefore cost). The potential benefits and impact of P&T on geological disposal will be discussed in Section 17.5.

The first requirement of P&T strategies is the deployment of spent fuel reprocessing to separate the TRU elements. These are then refabricated into fuel

Nuclide Half-life Energies of primary emissions (MeV) Specific activity Dose coefficients (Sv/Bq) а P (Ci/g) (W/g) (Neutron min 1 mg 1) 237 Np 2.14 x 106y 4.78 7.07 x 10“4 2.07 x 10“5 < 7 x 10“6 1.1 x 10“7 238Np 2.10 d 0.25 1.24 2.61 x 105 1.27 x 103 239Np 2.359 d 0.332 0.427 2.32 x 105 5.86 x 102 8.0 x Ю10 238Pu 87.404 у 5.49 17.2 0.570 155 2.3 x 10“7 239Pu 2.4413 x 104y 5.15 6.13 x 10“2 1.913 x 10“3 1.35 x 10“3 2.5 x 10“7 240 Pu 6580 у 5.16 0.227 7.097 x 10“3 53.7 2.5 x 10“7 241 Pu 14.98 у 4.9 0.02 99.1 4.06 x 10“3 4.7 x 10“7 242 Pu 3.869 x 105 у 4.90 3.82 x 10“3 1.13 x 10“4 95.3 2.4 x 10“7 241 Am 432.7 у 5.48 3.43 0.1145 3.55 x 10“2 2.0 x 10“7 242Am 16.01 h 0.63-0.67 8.11 x 105 2.08 x 103 24!mAm 144 у 5.207 Isomeric 10.3 3.08 x 10“2 1.9 x 10“7 Transition 243Am 7370 у 5.27 0.200 6.42 x 10“3 2.0 x 10“7 242Cm 162.7 d 6.11 3.32 x 103 122 1.21 x 106 1.3 x 10“8 243Cm 32 у 5.79 45.9 1.677 2.0 x 10“7 244Cm 18.099 у 5.81 80.94 2.832 6.87 x 105 1.6 x 10“7 245 Cm 8265 у 5.36 0.177 5.89 x 10“3 3.0 x 10“7 246 Cm 4655 у 5.39 0.312 1.01 x 10“2 5.58 x 105 2.9 x 10“7 247Cm 1.56 х 107 у 4.87 9.28 x 10“5 2.94 x 10“6 2.7 x 10“7 248Cm 3.397 х 105 у 5.05 4.24 x 10“3 5.34 x 10“4 2.58 x 106 1.1 x 10“6 249 Cm 64 m 0.9 1.18 x 107 2.06 x 104 250 Cm 1.74 х 104 у 8.20 x 10“2 ~0.1 6.49 x 108 2.9 x 10“7 252Cf 2.64 у -7 x Ю10

and ‘transmuted’ i. e. fissioned (or ‘burned’) in a neutron flux to produce useful energy. How this is done very much depends on national long-term energy goals (e. g. the phase-out of nuclear power versus an indefinite continuation) but, if deployed in conjunction with both thermal and fast reactors, it is possible to take maximum advantage of the energy available from uranium while, at the same time, reducing the amount and the toxicity of the resulting waste2 3. Of course, key requirements are that this must be done safely and at reasonable cost.

It has been estimated that about 10 million sealed radioactive sources have been produced worldwide and at least 2 million are currently in use in the USA.9 They have many uses including medical treatment and diagnosis, industrial radiography, borehole logging, level gauges and even domestic smoke alarms. When they come to the end of their useful life, many will still constitute a significant potential hazard and, furthermore, may be too long lived and too radioactive for either decay storage or near-surface disposal. Most of the countries that own such sources have no nuclear infrastructure and do not, therefore, possess facilities where the sources can be safely stored, let alone disposed of; many fatalities have been caused by loss of control. In response, the IAEA has developed a Code of Conduct on the Safety and Security of Radioactive Sources. This aims to provide guarantees that, where decay storage is impractical, sources will be returned to the manufacturer at the end of life. For those sources that cannot be returned to the manufacturer (e. g. legacy sources), the IAEA BOSS (borehole disposal of sealed sources) system10 can be applied (see Section 18.5.5).

Financing of any capital project may be achieved by equity, which means selling shares in a project, or by debt, i. e. borrowing money. An important distinction between the two relates to exposure to risk because lenders will often have first call on the assets of a failed project and may be favourably treated for tax purposes. Holders of equity, on the other hand, are completely at risk to project failure. Because of the reduced risk to lenders, the rate of return on borrowed money is lower than that required from equity. Consequently, project sponsors tend to prefer debt to equity. Conversely, would-be lenders will wish to see that a project has equity financing in place as a means of gaining confidence in the viability of the project and of limiting their own exposure.15 As a result, an equal split between debt and equity would not be an unusual outcome.

Where a government seeks to make a financial contribution to the cost of an NPP this, too, may be in the form of a loan guarantee, a loan or the purchase of equity. As for actually raising the money, governments have many ways of doing this. It may be secured against future tax, electricity or other revenues or, when making an agreement to import nuclear technology from another country, as a form of barter using exportable goods such as agricultural products or uranium ore. Funds may also be raised in advance through a surcharge on electricity sales from existing plant. In a few cases it may be available through a sovereign wealth or infrastructure fund.

Similar considerations will apply with respect to private capital except that the financial options will be more limited. Large utilities may pledge revenues from electricity sales from existing plant as well as the new one to repay debts. Use of on-balance sheet financing will help to demonstrate that the utility has confidence in the project and the financial strength to carry it through.

Until the current financial crisis, at least, global markets were more than adequate to meet the costs of NPP construction. The problem is not the absence of capital but the difficulty of persuading fund-holders to release it.15 The underlying issue is that of risk: what is its source, how great is it and who should bear it? This is explored in the next section.

Chemical-based technologies for uranium enrichment have been investigated. These make use of a slight preference for higher oxidation states shown by 235U compounds compared to 238U. A process investigated in France uses solvent extraction to preferentially discriminate between compounds in the IV and VI oxidation states, whilst a process developed in Japan uses solid phase ion exchange to preferentially discriminate between the III and IV oxidations states. Both processes require many stages to achieve appropriate levels of enrichment and neither has been developed to commercial operation. A thermal diffusion process was briefly employed during the Manhattan Project but soon abandoned, while plasma separation methods have been studied in the USA and France.

The fissile material in today’s nuclear power stations is predominantly uranium and plutonium as oxides, pressed and sintered to pellets, which are filled into tubes and arranged in assemblies (Sections 9.2 and 9.4). Other material forms have been tried in the past or are being considered, but none of them is applied on a commercial scale in power reactors.

9.5.1 Oxide pellet design and manufacture

Pellet fabrication for light water reactors starts with uranium hexafluoride (UF6) enriched to the required concentration of U-235 in an enrichment plant (Chapter 7). The UF6 is received in solid form in containers, heated to gaseous form and chemically converted to uranium dioxide (UO2) powder. Several conversion processes are available (Assmann, 1982), e. g.:

• Ammonium diuranate (ADU) wet conversion where UF6 is hydrolised in an aqueous ammonia solution from which ammonium diuranate is precipitated and then calcinated to UO2.

• Ammonium-uranyl carbonate (AUC) wet conversion where uranyl — tricarbonate is calcinated to produce UO2 powder.

• A dry process where UF6 is converted to UO2 powder by mixing it with steam and hydrogen in a kiln.

The dry process is a more environmentally friendly conversion technique than ADU and AUC since the clean waste, HF, can be disposed of more easily than the uranium-contaminated waste from the wet ADU and AUC processes. The dry process is therefore the preferred one today.

Doping agents are mechanically blended to the powder via a pre-blend, which ensures good homogeneity. The pre-blend is a mix of about 5% of the UO2 powder lot and the doping agents. The UO2 powder is further conditioned by adding pore former to obtain the desired pore distribution in the sintered pellet.

The powder is pressed to so-called green pellets with a pressure of 400-500 MPa obtaining about 50% of the theoretical density of UO2. The green pellets are sintered in a continuous, electrically heated production furnace in a reducing hydrogen atmosphere. The sintering takes 4-5 hours at a temperature of 1700-1800 °C. The density is increased to about 95% of the theoretical density or 10.5 g/cm3. Variants of this basic process are used to reduce the time and energy requirements and to influence the microstructure (e. g. grain size) of the ceramic pellets (Harada, 1997). In any case, the final product must have an oxygen-to-uranium ratio very close to 2 since deviations from stoichiometry will influence properties such as the fission gas diffusion coefficient and the thermal conductivity. Excess oxygen will also corrode the cladding from inside.

Fuel pellets have precise dimensions. The nominal pellet diameter, which in most designs is in the range 8-10 mm, must be met within a tight tolerance of ±12 pm in order to obtain an accurately known gap between the pellets and the cladding. However, due to friction forces between the powder and the die, the green pellets have density variations leading to uneven sintering shrinkage. The pellets are therefore ground to a cylinder with precise dimensions. In addition, a chamfer is ground to the end faces to reduce chipping, which is a cause of fuel failure. An end face dishing is formed to accommodate the larger thermal expansion in the centre of a fuel pellet. The overall shape of the final product is shown in Fig. 9.10.

In the past, moisture in pellets caused cladding corrosion from inside, hydrogen embrittlement and fuel failure. Before insertion into the cladding tube, the pellets are therefore dried in a vacuum at 120-150 °C to reduce absorbed water on the surface to less than 10 ppm.

The most distinguishing feature of the CANDU reactor is its ability to use natural uranium fuel, although the CANDU reactor would also be efficient at utilizing other fuels and fuel cycles. The natural uranium fuel cycle is simple and inexpensive, involving uranium mining, milling, refining, conversion to UO2 and fabrication of CANDU fuel bundles. For many countries, this avoids the expense and infrastructure associated with enrichment technology and contributes to self­reliance in nuclear fuel supply. As was discussed previously, the fuel bundle itself is small, simple and inexpensive. This results in the fuel cycle cost for a CANDU natural uranium fuelled reactor being roughly half that of an LWR (in $/kWh). The impact on CANDU fuel cycle cost of lower burnup is more than offset by the avoided costs of enrichment and conversion and by lower fuel fabrication costs, relative to LWR fuel (NEA/OECD, 1994). The natural uranium fuel cycle is also readily localized and all countries that have CANDU reactors also fabricate their own CANDU fuel.

Because of the focus on high neutron economy in the CANDU reactor and fuel design, the fuel utilization (thermal energy derived from the mined uranium) is up to 50% higher than for an LWR (Boczar et al, 1996). A contributory factor is the high conversion ratio in the CANDU reactor: the amount of plutonium created per unit of energy produced is about double that of a typical LWR. By the time the fuel has reached its discharge burnup, over 50% of the energy is being produced by the fissioning of plutonium.

The mass of used CANDU fuel produced per unit of electricity generated is about five times greater than for a typical PWR (Boczar et al, 1996). This is offset, however, by its lower decay heat in terms of the size and cost of used fuel storage or disposal facilities. Indeed, the cost of geological disposal of used fuel is more aligned with the total energy produced from the fuel, rather than its burnup. Hence, the cost of permanent disposal of used CANDU fuel is similar to the cost of LWR fuel disposal, per unit of electricity produced (Allan and Dormuth, 1999). Moreover, the low burnup of CANDU fuel results in a very small quantity of higher-mass actinides (such as americium and curium) per unit energy produced by the fuel, a benefit in terms of radiotoxicity of the used fuel.

It should also be noted that the low burnup of natural uranium fuel avoids some of the fuel performance challenges that arise at higher burnup, and this contributes to the excellent performance of CANDU fuel, with very few fuel failures. Occasional fuel failures can be detected and removed on-power with low economic penalty.

The resource efficiency of the CANDU reactor means that there is little residual fissile material remaining in the used fuel, making its recovery and recycling unattractive at the present time. The recoverable fissile content of used CANDU natural uranium fuel is about one-fifth of that in used LWR fuel. U-235 is at the level of enrichment plant tails (around 0.2% U-235 in total uranium) compared to around 0.9% in used LWR fuel, so there is no economic incentive to recycle the U-235 since there are several hundred thousand tonnes of depleted uranium in the form of enrichment plant tails readily available. The concentration of plutonium in used CANDU NU fuel is less than 40% of that in used LWR fuel: around 3.7 g Pu/kg HE compared with ~10 g Pu/kg HE for used LWR fuel, depending on the burnup. The fissile content of the plutonium in both used fuels is around 70%. Hence, the cost ($/kg HE) of recovering the plutonium would be much higher than for used LWR fuel. Nonetheless, the total quantity of plutonium in used CANDU natural uranium fuel for a given amount of electricity generated is about double that of LWR fuel, and the plutonium from used CANDU natural uranium fuel in Canada is potentially a valuable future resource if it could be economically recycled (Boczar et al., 2010).

The VHTR is a thermal neutron system originally designed to operate a once — through fuel cycle with low-enriched uranium fuel and very high fuel burn-up. However, the system’s flexibility will allow it to adopt closed fuel cycles using thorium fuel and ‘burner’ cores that can efficiently transmute Pu (Abram and Ion, 2008; Generation IV International Forum, 2009). It will be possible to use VHTR with Pu fuel and for MA incineration or transmutation, due to the high burn-up capabilities of the coated particle fuel, though the build-up of even — numbered Pu isotopes is an issue that will need addressing. These features can also be used in symbiosis with other reactor types to reduce MA content and decay heat, which effect repository design (see 13.4.3). The deep-burn potential of VHTR avoids multi-recycling of spent fuel. It is especially attractive if it can be shown that ultra-high burn-up coated particles are still able to maintain their barrier function under disposal conditions (Brossard et al., 2009). VHTR could also potentially use Th as a fuel, as shown by the experience of HTR (Mazzini et al., 2009). Using a Th-based fuel cycle in a HTR with Pu as a driver would (see Fig. 13.8) increase the efficiency of TRU fission and achieve higher fuel burn-ups. Another advantage is that Th is about three times more abundant than U.

13.8 Mass of actinides at EOC starting from 1 g of Pu (Cerullo et al., 2005).

The VHTR core can be constructed from one of two basic designs (see Fig. 13.6):

1 the prismatic block type

2 the pebble bed type

From core configurations (deterministic vs. stochastic) and refuelling schemes (batch-wise vs. continuous) points of view, the pebble-bed and prismatic fuel design are quite different (Lomonaco, 2003). Just to give an example, the pebble-bed core
configuration needs specific design tools, as underlined in Bomboni et al. (2009a, 2010, 2012).

Anyway these designs have a number of common features. These include a UO2 kernel surrounded by successive layers of porous graphite, dense pyrocarbon (PyC), silicon carbide (SiC) then pyrocarbon (PyC). This could be enhanced through the use of a UCO kernel or ZrC coating. These coatings have the potential to provide improved burn-up capability, minimized fission product release and increased resistance to core heat-up accidents, even above 1600 °C, which is considered the maximum operating temperature for TRISO fuel (Brossard et al., 2009) (see Fig. 13.9). Coating of fuel particles could be achieved by chemical vapour deposition.

Empirical formulations exist for HTR fuel but little is known about how different process parameters, e. g. gas composition and temperature, would affect the properties and the performance of the resulting fuel (Abram and Ion, 2008). There is also a need to undertake more research on manufacture, characterization and irradiation performance. Irradiation tests are necessary for the fabrication process, fuel design and fission product transport, as well as for post-irradiation and safety testing. Fuel performance must also be assessed for both normal operating and accident scenarios. A key requirement of the fuel is its ability to retain fission products in the fuel particles under a range of accident scenarios with temperatures up to 1600 °C. Although very good irradiation performance has been demonstrated under HTR conditions, the behaviour of

13.9 I nfluence of temperature on TRISO fuel failure fraction. Brossard et al., 2009.

the coated particles under irradiation is not fully understood (Abram and Ion, 2008).

There are a number of possible methods for dealing with spent fuel:

• direct disposal of coated particles and graphite moderator

• separation of coated particles and moderator, with separate treatment of both fractions

• separation of kernels from coatings and reprocessing kernels for recycling in VHTR systems (or other reactors)

Research on long-term repository/direct disposal is currently focused on the potential of SiC coatings. It is believed that the coating may act as a miniature containment vessel to retain fission products during the repository post-closure period but more information are needed about the long-term integrity of these layers (Brossard et al., 2009). The main focus of research in reprocessing fuel is particle kernel dissolution (Brossard et al., 2009). Other research has concentrated on technologies to separate the highly active graphite fractions from those of low activity and to evaluate the feasibility of reusing the graphite (Brossard et al., 2009).

Many computer programs used for fuel design and licensing calculations are fuel or reactor vendor owned codes, which are not available for general use. These are not considered further here. The remaining, more generally available, codes are discussed further below.

Lattice codes commonly used for LWR applications include CASMO and WIMS. For fast reactor applications, ECCO is most often utilised (at least in Europe). The corresponding whole core neutronics codes are SIMULATE, PANTHER and ERANOS, respectively. Of the Monte Carlo codes, MCNP is the most widely used.

With respect to LWR applications: system thermal-hydraulic codes commonly used include RELAP and RETRAN; the most commonly used core thermal- hydraulics code is VIPRE. With respect to fast reactor applications, the choice is more varied — a selection of core thermal-hydraulics codes used for liquid-metal cooled fast reactor applications can be found in the proceedings of an IAEA technical committee meeting held in Obninsk, Russia, in 1998 (IAEA, 2000b).

Fuel performance codes commonly used for LWR applications include TRANSURANUS, ENIGMA, FEMAXI and FRAPCON (steady-state)/ FRAPTRAN (transient). A fuller list of codes used for water-cooled reactor applications can be compiled from the codes utilised in IAEA’s FUMEX-III co-ordinated research project (Killeen et al., 2009). For fast reactor applications, TRAFIC is the standard code used in the UK; other codes are utilised elsewhere.

Metallic elements at oxidation levels IV and VI are more easily extracted by TBP than levels I, II, III or V. So, uranium and plutonium existing at valences VI and IV respectively are significantly extracted while other actinides (americium, curium at valence III, neptunium at valence V) and most of the fission products (FPs) are poorly extracted.

As a result, the following organic complexes are preferentially solubilized in the organic phase:

UO2 (NO3)2.2TBP and Pu2 (NO3)4.2TBP

which enables the uranium and plutonium to be separated from the fission products.

TBP also has the ability, however, to complex some metallic cations that are ionically bound to nitrate anions:

M”+(aqueous phase) + mNO3- (aqueous) + n TBP ^ M(NO3)m. nTBP (organic phase)

Scrubbing the TBP solvent with nitric acid (fission product scrubbing) can improve the separation of uranium and plutonium by re-extracting into the aqueous phase those fission products that were dragged along by the solvent, particularly the zirconium and ruthenium. A second scrubbing removes the technetium, which is troublesome for uranium/plutonium partition.

The objective here is compatible both with the use of Pu (as a resource) in LWRs (for a limited period of time, assuming that the deployment of fast reactors will be delayed), and with the use of Pu (or Pu+Np) in FRs, while MAs are handled in dedicated reactors (ADSs, low CR critical fast reactors, etc.).

To implement this scenario, the so-called ‘double strata’ strategy can be envisaged: MA fuels (most probably with the addition of some Pu) should be transmuted in dedicated reactors (ADSs or low CR critical FRs) while the bulk of Pu (or Pu+Np) should be multi-recycled in MOX-LWRs (e. g. Ref. 26) or in FRs. The scheme for this scenario is represented in Fig. 17.5.

The main objective of this scenario is to keep the management of MAs in a separate cycle of a smaller size, independent from the fuel cycle associated with energy production, where Pu (or Pu+Np) is multi-recycled. This is an important point, since the reprocessing and the fabrication of the fuel for the dedicated transmuters is associated with very high decay heat and neutron sources, as indicated in Section 17.3.3, which can require costly measures (e. g. remote handling, long cooling times, increased shielding, etc.) to allow feasibility.

The expected reduction of radiotoxicity is close to that expected in scenario (a) above, if the chemical separation performance (e. g. losses during reprocessing) is similar in the two scenarios.