In standard lightwater reactors, only 1 to 2% of the energy potential of natural uranium is used. In fact, essentially only the 235U isotope is fissioned. The dominating 238U isotope (99.28% in natural uranium) contributes only marginally. However, a two-step process can be envisaged first to convert 238U into 239Pu (by neutron capture) and successively to fission 239Pu. In this manner one can make a much more efficient use of natural uranium. The first step of the process outlined above can be envisaged if the neutrons issued by fission (and which have energies around an average value of 2 MeV) are moderated to lower energies as little as possible.
In fact, the properties of nuclear reactions induced by neutrons of energies higher than few kiloelectronvolts ("fast neutrons") limit the neutrons which disappear by parasitic captures in the structures. Moreover, at these energies there is a favorable competition between fission and capture. The neutron balance in the core of this type of reactor (called fast reactor) is positive in the sense that neutrons are available: out of the average number of neutrons emitted per fission (2.4 for 235U, 3 for 239Pu), one neutron is needed to keep the chain reaction and one neutron is available to "breed" new fissile material if parasitic capture and captures in fissile materials are kept as low as possible. A similar "excess" of neutrons is not available in thermal reactors.
The need to keep the neutrons issued by fission as "fast" as possible makes it necessary to avoid the presence of light elements (like hydrogen) in the core, since neutrons loose higher amounts of energy when interacting with such nuclei. As mentioned above, the neutron capture probabilities are smaller for most materials for fast neutrons; as a consequence, the choice of structural materials for fast reactors is mostly related to their performance under irradiation and not to their nuclear properties. In the same energy domain, however, fission probabilities ("cross-sections") are small compared to those associated with thermal reactor neutron energies. Thus, there is the need for a high density of fissile nuclei per core unit volume and hence, for a high fissile material content in the nuclear fuel. Optimal values for large, fast power reactors (1500 MWe) are found when the Pu/U ratio in the fuel is in the range of 0.15 to 0.25. The typical value for the volume of a 1,000 MWe core is about 10 m3 containing approximately 5 tons of Pu. The resulting high heat density needs a specific coolant fluid with very high thermal transfer capability. Liquid metals, and in particular, liquid Na have been unanimously chosen.
The Na thermal conductivity at 500°C, 67 W m−1 °C−1 is a hundred times higher than for water at 40°C. Na stays liquid over a large range of temperatures (~100°C to ~900°C) and this allows high core outlet temperatures, favorable to increased thermodynamic efficiency. Despite these and other advantages of Na, some disadvantages have to be mentioned; in particular, the high chemical affinity for oxygen and water. The opacity of Na is an obstacle for easy in-core inspections, and the fact the Na is solid below 98°C, makes it necessary to keep Na temperature above 150°C in any configuration of the power plant. Since, Na becomes radioactive under irradiation, the formation of the radioactive isotope 22Na (decay half-life ~2.5 years) needs precautions. Finally, Na should be kept well purified to avoid corrosion.
The core of a typical fast reactor is made with fuel assemblies (hexagonal geometry), and can be surrounded by assemblies which contain the "fertile" material (238U). The control of the chain reaction is provided by control rods in the core, which contain an absorber material (such as B4C, natural or enriched in the isotope 10B).
Typical fuel assemblies (e.g., in the case of the French fast reactor SUPER-PHENIX) are made of thin (4 to 5 mm) hexagonal tubes in stainless-steel, which contain a regular lattice of fuel pins (271) with diameter less than I cm. Each pin is made up of fuel pellets of mixed UO2−PuO2 within a stainless-steel cladding, less than 0.5 mm thick. Pins are kept in place by an helicoidal spacer around the pin itself. The height of the fissile column is 100 cm. The Na enters from the bottom of the assembly with T = 395°C and has T = 545°C at the outlet. The maximum heat generation at the center of the core corresponds to a power of 460 KW/liter.
Economy motivates the increase of irradiation time as far as possible. A major constraint is represented by the metallurgical behavior of materials (e.g., swelling). Present experience indicates the possibility of reaching fuel burn-ups of 100,000 MWd/ton (of heavy metal) and higher if appropriate choices are made for fuel-pin cladding materials (e.g., high-Ni SS) and for hexagonal tubes (e.g., ferritic SS).
Fuel types other than mixed oxides can be envisaged; in particular, mixed nitrides which could offer the advantage of higher density and better thermal conductivity. Metal alloys have also been used and are presently being actively studied in the USA, within the frame of a pyrometallurgical technology development for fuel reprocessing and fabrication.
As far as safety characteristics are concerned, temperature increase gives rise to a reduction of core reactivity, partly by nuclear effect (Doppler effect) which acts without any delay. However, the ejection of Na from the center of the core can result in a reactivity increase, and measures are taken to eliminate the risk of void formation (in particular, boiling). The high thermal inertia associated with the large mass of Na, which is kept in normal operation at approximately 300°C below the boiling point, is a further guarantee that boiling of Na, if ever attached, will be reached only after long-enough delay.
The risks related to Na fires and Na-water exothermic reactions are handled with appropriate design features of Na circuits, and several sophisticated detection instruments are provided. Fast reactors of different design, size and fuel type have been built and operated in the world for more than 40 years. The largest plant is SUPER-PHENIX in France, an industrial prototype of 1,200 MWe (see figure). A large amount of relevant experience has been gathered on the operation of this reactor type. Breeding properties have been successfully proved experimentally.
However, the present context of relatively low Uranium scarcity does not favor the industrial deployment of breeders before a few decades. But breeding capability is not the only attractive property of fast reactors. They are well-suited for using plutonium fuel and for burning it efficiently, if needed, to avoid stockpiling. Moreover, their excellent neutron economy indicated above offers possible solutions for the burning of long-lived radioactive wastes.