The Purex process was designed to recover plutonium for military applica­tions or for energy production with fast neutron reactors. In the latter case, recovered uranium was also useful. At present, plutonium is recovered for the fabrication of MOx fuels for PWRs and BWRs. The advocates of MOx fuel use insist that it not only is cost effective since it decreases the needs for enriched uranium, but that it can also help reduce the radiotoxicity

of the wastes. However, while plutonium is indeed the main long-term source of radiotoxicity for UOx spent fuels, minor actinide contributions become dominant in spent MOx fuels. Hence the idea to extract also minor actinides in order to transmute them. A number of processes have been proposed lately with that view.

Actinides are extractable by the Purex process when they are in an even state of oxidation. For example uranium is readily extractable in its VI oxi­dation state as the uranyl ion UVIO2+. Plutonium is also easily extracted as PuIV, but not as PunI, which allows separation of plutonium from uranium. Neptunium is present in nitric acid solutions in states NpV and NpVI. This coexistence of odd and even oxidation states of neptunium allows us to consider its extraction by a rather simple modification of the classical Purex process, as shown in reference [145]. The cases of americium and curium are more difficult since they essentially appear in the odd III oxida­tion state in nitric acid solutions. Active research and development in several countries is pursued in order to find efficient processes, compatible with Purex, for their separation [145]. They test different complex organic molecules with high selectivity for americium and curium. One of the main difficulties is that rare earths are also present in the nitric solution in oxida­tion state III, and are co-extracted with americium and curium. Facing this challenge two strategies have been adopted:

1. Use selective stripping with complexants to separate americium and curium from rare earths. Pertaining to this approach are the so-called Talspeak [146], DIDPA [147] and Truex [148] processes.

2. Use a two-cycle separation process with two different solvents where the first step separates americium, curium and lanthanides from the other fission products, and the second step separates americium and curium from the lanthanides. Examples of such processes are TRPO [149] and Diamex [150].

As examples of the two approaches we discuss a little more precisely the Truex and Diamex processes.

The proposed cycle is the Th-U one. The main objectives are

1. to incinerate transuranics

2. to transmute a number of fission products.

The very high proposed thermal neutron flux reaches 1016/cm2/s. Neutron multiplication is obtained either by 233U fission, or by fission of the actinides one wants to incinerate: plutonium, americium or curium. 233U is obtained via neutron irradiation of a 232Th blanket, followed by online extraction of

ПЛІ O-J-J

Pa which is allowed to decay into U away from the neutron flux. This is made possible by the use of a molten salt (a mix of fluorides) fuel, similar to that which was used in the Oak Ridge pilot reactor. The liquid fuel circulates continuously through the protactinium extraction facility. In order to limit the number of neutron captures in protactinium, the thorium blanket is placed in a neutron flux limited to a few 1014neutrons/cm2/s. The region of maximum thermal flux is where the actinides are incinerated. Indeed, very high fluxes have the following advantages.

• Reduced lifetime of the actinides in the reactor. The lifetime of 239Pu in a thermal-neutron flux of 1016 neutrons/cm2/s is less than two days.

• Improved neutron balance of the incineration process.

• Small inventory of fissile matter in the system. The quantity of plutonium necessary to produce 3 GW in a flux of 1016 neutrons/cm2/s is as small as 8 kg, with a daily burn-out of 3.5 kg.

Fission product transmutation would be optimal in the epithermal flux region since it is in the resonances that the absorption cross-sections are

maximum. Fission products with a capture cross-section of 1 barn would live 3 years in a 1016 neutrons/cm2/s flux. In order to prevent stable fission products becoming radioactive by neutron capture, an online separation of fission products to be transmuted is necessary.

The advantages mentioned are, of course, counter-balanced by the great complexity of the system:

• An accelerator able to accelerate protons to at least 1 GeV, with intensities larger than 100 mA.

• A subcritical assembly using molten salt fuel. Although tested on a small scale at Oak Ridge, this technique has to demonstrate its resistance to very high fluxes. Corrosion problems may be serious, even if the use of hastalloy (a special nickel alloy) seemed to be satisfactory in the Oak Ridge conditions. One should also note that since the fuel itself circulates in the primary heat exchangers, any intervention on these, often delicate, components would be very difficult if not impossible.

• A complex online chemistry for separation of protactinium and fission products and continuous injection of the fuel.

Figures 12.4 and 12.5 show one of the designs which have been pro­posed, together with a diagram of the chemical processing. They show the complexity of the system which could only be implemented in countries with a very advanced nuclear technology.

The RBMK-1000 is a graphite moderated pressure tube type reactor, using slightly enriched (2% 235U) uranium dioxide fuel. It is a boiling light-water reactor, with direct steam feed to the turbines, without an intervening heat-exchanger. Water pumped to the bottom of the fuel channels boils as it progresses up the pressure tubes, producing steam which feeds two 500 MW(e) turbines. The water acts as a coolant and also provides the steam used to drive the turbines. The vertical pressure tubes contain the zirconium-alloy clad uranium dioxide fuel around which the cooling water flows. A specially designed refuelling machine allows fuel bundles to be changed without shutting down the reactor, a design interesting for the extraction of fresh plutonium for military use.

The moderator, whose function is to slow down neutrons to make them more efficient in producing fission in the fuel, is constructed of graphite. A mixture of nitrogen and helium is circulated between the graphite blocks largely to prevent oxidation of the graphite and to improve the transmission of the heat produced by neutron interactions in the graphite, from the moderator to the fuel channel. The core itself is about 7 m high and about 12m in diameter. There are four main coolant circulating pumps, one of which is always on standby. The reactivity or power of the reactor is controlled by raising or lowering 211 control rods, which, when lowered, absorb neutrons and reduce the fission rate. The power output of this reactor is 3200 MWth or 1000 MW(e). Various safety systems, such as an emergency core cooling system and the requirement for an absolute minimal insertion of 30 control rods, were incorporated into the reactor design and operation. Figure II.1 is a sketch of the RBMK reactor.

The most important characteristic of the RBMK reactor is that it possesses a ‘positive void coefficient’. This means that if the power increases or the flow of water decreases, there is increased steam production in the fuel channels, so that the neutrons that would have been absorbed by the denser

water will now produce increased fission in the fuel. However, as the power increases, so does the temperature of the fuel, and this has the effect of reducing the neutron flux (negative fuel coefficient). The net effect of these two opposing characteristics varies with the power level. At the high power level of normal operation, the temperature effect predominates, so that power excursions leading to excessive overheating of the fuel do not occur. However, at a lower power output of less than 20% of the maximum, the positive void coefficient effect is dominant and the reactor becomes unstable and prone to sudden power surges. This was a major factor in the develop­ment of the accident.

Another safety deficiency was the design of the control rods: the boron carbide absorbing section of the rods was preceded by a 4.5 m long pure graphite displacer. The reason for this displacer was to prevent water from replacing boron carbide when the rods were in the high position. Indeed the neutron absorbing character of water would have lessened the effect of the boron carbide. However, the graphite displacer was not long enough to occupy the full length of the channel. Thus, the fall of a control rod initially in its highest position first replaced water by graphite in the lower part of the channel, with a subsequent increase of the reactivity. It is only after the rod is significantly engaged in the core that the absorbant becomes effective. Furthermore, the rods’ insertion was rather slow, taking about 20 s.

The International Panel on Climate Change (IPCC) reviews periodically the evidence for climate change and, depending on the rate of greenhouse gas emissions, makes predictions on future evolutions. Because the future of nuclear energy cannot be discussed without reference to these aspects, we summarize the main conclusions reached by the IPCC as well as by organi­zations such as the World Energy Council (WEC) and the International Institute for Applied Systems Analysis (IIASA) who focus on the prediction of future energy consumption.

Evidence for anthropically induced climate change

Figure 2.1 shows how the concentration of three important greenhouse gases in the atmosphere follows the world population increase.

The correlation between the four curves is striking with an especially fast increase of greenhouse gas concentrations after 1950. Although the greenhouse efficiency of methane and other gases is much larger than that of CO2, its much larger abundance gives it the dominant contribution. Typically, CO2 accounts for 62% of the additional radiative forcing due to anthropic emissions, methane for 20% and other gases for the rest. Water has an amplifying role but is not a driving factor.

Although the increase of greenhouse gas concentrations is unquestion­able, its effects on temperature are more difficult to establish. Figure 2.2 shows a clear rise of the temperature during the last century. However, this increase is not attributable solely to increased concentration of green­house gases. Climatic models indicate that only the temperature increase observed between 1960 and the present can be attributed to greenhouse gas emissions.

The treatment of finite systems requires that boundary conditions be defined. In this simplified discussion we consider the case of a homogeneous medium surrounded by a vacuum. At the boundary of the medium, there is only an outgoing one-sided flux J+ while J_ = 0. This means that the current J = J+_ J_ > 0 and thus that, since grad(‘) > 0, ‘ decreases from the inner to the outer region. Extrapolating ‘ linearly in the vacuum region, where the diffusion equation is not valid, the extrapolated value should vanish at some distance dextra. By comparison with exact calculations one finds that ‘ is a good approximation of the true solution for dextra ~ 2D. This is usually very small compared with the multiplying medium size so that a simple, but sufficient, approximation of ‘, at least for qualitative discussions, is obtained by requiring it to vanish at the boundaries of the medium. To illustrate this, we solve the diffusion equation for a semi-infinite homogeneous reactor bounded by two parallel planes [55].

The purpose of hybrid reactors is to produce energy as well as a neutron excess which could be used for nuclear waste transmutation. As a con­sequence, it is important to evaluate to what extent the energy produced by fission in the multiplying medium exceeds the energy of the primary particle beam. As an example of expected values of k, we discuss the
experimental measurement of the energy produced in a subcritical system as was done in the FEAT experiment at CERN [122].

After these very simple examples, we are able to ‘build’ a more realistic reactor. As a starting point, we want to study a light water reactor (critical) with a lead reflector, loaded with UO2 fuel. Suppose that the water volume is about 4 times more than the fuel volume. The core is a cylinder with a diameter equal to its height. The core radius is 1 m, the reflector thickness is 50 cm and the iron tank containing the two is 5 cm thick.

Homogeneous core

In a first step, suppose that we consider a homogeneous core of (H2O + UO2) with 1.35% (compared to 238U) 235U enrichment.

The condition VH2O/ VUO2 = 4 implies a fuel density of pfuel = 2.8 g/cm3 (taking UO2 « 10 g/cm3) and the atomic composition of the fuel is 6 moles of H2O for 1 mole of UO2. Now, we are able to write the MCNP input file:

Homogeneous core c

c Exterior c

1 0 1:-2:3 imp:n=0 c

c Iron tank c

2 1 -7.87 -1 2 -3 (4:-5:6) imp:n=1 c

c Lead reflector c

3 2 -10.34 -4 5 -6 (10:-11:12) imp:n=1 c

c Core c

4 3 -2.8 -10 11 -12 imp:n=1

c tank/reflector surfaces

1 cz 155

2 pz -155

3 pz 155

4 cz 150

5 pz -150

6 pz 150

c reflector/core surfaces

10 cz 100

11 pz -100

12 pz 100

c Material

ml 26000.55c 1 $ Iron of the tank

m2 82000.50c 1 $ Lead of the

reflector

m3 92235.60c 0.0135 92238.60c 0.987 & $ 235U (1.35%) + 238U 1001.60c 12. 8016.60c 8. $ 6 H2O + 1 O2 (of the

fuel)

sdef pos 000 erg 2.5 kcode 1000 1 10 150 totnu

The keff of the reactor with that geometry is keff = 1.001 ± 0.001.

The highest-power cyclotron system presently in operation is the cyclotron ensemble of the Paul Scherrer Institute (PSI) at Zurich, Switzerland.

The PSI cyclotrons. The characteristics of the PSI cyclotrons are given in Table 6.3. From the table we obtain a 27% RF-to-beam efficiency. The line-to-RF efficiency amounts to 66%, while the total line-to-beam efficiency reaches 18%. Compared with LAMPF, one notes a better efficiency of the RF power units, but a lower RF-to-beam transfer efficiency.

Let 72 and 73 be the attenuation factors in zones 2 and 3. In zone 3 the flux should remain finite at infinity. It must have the form

u3(r) = a3 e-73r. (10.5)

At radius R2, one writes the continuity of neutron flux and current, and obtains a relation between the coefficients a2 and b2 which appear in

and

u2(r) = a2 e-‘2r(1 +C e2’l(r-Rl)). (10.10)

Note that, as expected, b2 cancels out for large values of R2 and, also, if the properties of medium 3 are the same as those of medium 2.

The value of a2 is obtained from the value of the neutron current at

r = Rp.

Defining

one obtains

Preliminary design studies should allow selection of the most promising concept, which could be adapted to the different nuclear policies of the European Union members. In a first step, the XADS is viewed as an actinide incinerator. In this context, the fast spectrum is clearly the most appropriate. Two cooling system concepts are considered: the lead-bismuth eutectic (LBE) and helium. Both concepts have advantages in terms of safety and potentialities.

The Pb-Bi can be used both as the spallation target and the core coolant. When used in the core reflector, it allows optimization of long-lived fission product transmutation (TARC effect [57]). Furthermore, the liquid metal can provide a passive way to help residual heat extraction by natural con­vection. As compared with liquid lead, the eutectic decreases the operating temperature, which allows operation of the system at lower temperature, where the problems of structural material corrosion are simplified. Such a demonstrator would not be completely representative of an industrial reactor, concerning the power density in the fuel. Furthermore, bismuth is a very expensive metal and the presence of 209Bi leads to a significant produc­tion of 210Po during irradiation, a highly radiotoxic nucleus. The LBE concept appears to be a good way to test and develop the concept of a lead-cooled core, even if it would be difficult to consider industrial LBE reactors on a large scale.

The gas concept minimizes neutron slowing-down and facilitates reach­ing very fast neutron spectra, which optimize minor actinide incineration. The inspection of the fuel elements during operation is made much easier with gas compared with liquid metal, which requires coolant draining to detect any structural deterioration. The main problem of a gas cooled concept concerns residual heat extraction in an accident situation. Further­more, most of the fuel considered for use in a gas cooled reactor contains carbon (SiC, graphite, etc.), which forbids the presence of oxygen, and thus the presence of air, in the gas. The gas cooled XADS would operate at low temperature, with classical oxide fuels. This configuration would not be representative of a full-scale gas cooled reactor, which would operate at high temperature, with innovative fuels. However, the gas cooled XADS would be a first step in the development of such reactors.

In this context, different cores are studied: a small LBE core (about 20- 40MWth), a larger (80MWth) LBE cooled concept, and a gas cooled core (100MWth). In the gas cooled system, two types of fuel are considered: a standard cladded pin fuel element and a pebble-bed core concept. The spallation target may be a liquid heavy metal (LBE) which could be separated or not from the accelerator by a window. For the gas cooled concept only, a solid spallation target (tungsten) is being studied.

As can be seen, a large range of options is examined, more or less innovative, in order not to be stopped by a technological impossibility, and at the same time encouraging innovative options which could find an application beyond the ADSR concept.

The fuel considered would be, in a first phase, a MOx fuel with high plutonium concentration (around 20-25%). This type of fuel is relatively standard and would allow a precise study of the subcriticality characteristics. In a second step, different innovative fuels can be considered, such as highly enriched minor actinide fuels or the pebble-bed concept. The XADS appears to be an ideal tool to test new types of fuel in favourable safety conditions, as long as it operates far enough from criticality.

One of the goals of this project is to develop a common integral safety approach on both lead and gas cooled cores. A specific analysis of the subcriticality characteristics has to be performed, in order to validate the well-known advantages of the ADSR. In this respect, present experiments, for example MUSE [134] (Cadarache, France), should provide essential

GAS COOLED ADS DEMO

REACTOR VESSEL

PROTON BEAM

Pb-B. INLET

j_-r —► Pb-B. OUTLET

REACTOR VESSEL

REMOVABLE HEAD

CHECK VALVE

COOLANT

INLET

SHUTDOWN

COOLING SYSTEM (2.)

REACTOR THIMBLE

REACTOR VESSEL

CORE FEEDING PLENUM

CORE SUPPORT PLATE

Figure 13.3. Sketch of the gas-cooled ADSR demonstrator proposed by Framatome ANP.

results, in particular concerning the control of the different multiplication factors.

Once the subcriticality specificity is fully understood, it is clear that such a demonstrator could be an ideal tool to test different aspects of the safety of lead or gas cooled fast reactors, whether subcritical or not.

Figure 13.3 shows a first design of a gas cooled (He) core proposed by Framatome ANP. The power of such a reactor is around 100 thermal MW.

Concerning the accelerator, different aspects will play a role in the determination of the chosen type, such as reliability, availability, stability and reproducibility of the power control. Linac and cyclotron types of accelerator will be investigated. The purpose is to provide a proton beam in the range 600 MeV to 1 GeV. For a thermal power of the whole system of 100 thermal MW a beam intensity of a few mA is needed, depending on the subcriticality level which will be chosen. On one hand, the cyclotron concept requires a limited space but is limited in energy or intensity. On the other hand, the Linac concept is not limited in power specification, but requires a huge and expensive installation (several hundred metres for the proton beam). The investment budget will have a determining influence on the choices made.

The beam transport line could reach the spallation target horizontally or vertically. In the first design proposed by Framatome ANP, the beam arrives vertically above the spallation target. This kind of design implies a con­gestion of structural and handling elements above the core: Pb-Bi circulation (spallation target), fuel element handling structures, beam line (and par­ticularly the magnet). In this sense, a beam hitting the spallation target horizontally, associated with vertical element management and horizontal Pb-Bi circulation (on the other side), could simplify the congestion problems. Obviously, the safety aspects of this kind of design must be precisely studied.