Nuclear Reactors

Fission Process

In nuclear reactors energy is released by fission of heavy nuclides into two lighter fission products. The kinetic energy of these nuclei, together around 160 MeV/fission, and the ionizing radiation released in the fission process, together around 30 MeV/fission, are converted to heat and transported from the reactor to the energy conversion system by the primary coolant. In the fission process, on average 2.5 new fission neutrons are released, which can be used to sustain the fission chain reaction in the core. Decay of radioactive fission products both complicates and facilitates the safe operation of a nuclear reactor.

  1. At stationary operation, around 6% of the total energy release originates from fission product decay. Although this fraction diminishes quickly after reactor shut down, it is sufficient to severely damage the fuel in the reactor core in the absence of adequate cooling. This means that measures should be taken to always transfer the fission product decay heat to the environment by safe path ways.
  2. A small number of fission neutrons are produced by decay of neutron-rich fission products, so-called precursors. Only due to this delayed neutron production, the fission chain reaction can be controlled by relatively simple means. This is accomplished by absorbing so many fission neutrons that the chain reaction would die out with production of prompt fission neutrons alone. Then the delayed fission neutrons are needed to sustain the fission chain reaction in the core, and the reactor period, the characteristic time to describe reactor power changes, is determined by the decay of precursors, which on average takes more than 10 seconds.

Characterization of Nuclides

Most reactors use U-235 as a fuel. This is a fissile nuclide, which means that it can be induced to fission after absorbing a neutron of any energy, including one with virtually zero energy. The most common uranium isotope, U-238, present in natural uranium for 99.3%, has a much higher fission barrier and can be induced to fission only after absorption of a fast neutron (E>1 MeV). Such nuclides are called fissionable. Because U-238 produces a fissile plutonium isotope (Pu-239) after capturing a neutron (capture is the process of neutron absorption that does not induce fission), it is also fertile. The process of fissile nuclide production is called conversion, and the amount of fissile nuclides produced compared to the ones consumed is called the conversion ratio. Most reactors in operation today have a conversion ratio around 0.6. A special class of reactors has a conversion ratio larger than unity and is called breeder reactors. The conversion ratio minus one is then called the breeding gain, although much more advanced definitions need to be used for thorough analyses.

Characterization of Reactors

The class of fission reactors operating with low-energy neutrons is called thermal reactors. For fissile nuclides, low energy neutrons have a much larger probability to induce fission than high energy neutrons. Therefore, the fission neutrons, which are released with a mean energy of around 2 MeV, are slowed down by means of scattering events as quickly as possible. To this purpose, a moderator is added to the fuel. The most efficient moderators are light nuclei, like hydrogen, deuterium, beryllium and graphite, as they allow the largest energy transfer per collision and do not capture too many neutrons. About 80% of all reactors in operation today are Light Water Reactors (LWRs), which use ordinary water both as a coolant and as a moderator. Canadian reactor designs employ heavy water as a moderator and either heavy water or light water as a coolant. Both the Russian RBMK design and the Britain Magnox reactors use graphite as a moderator, and light water (RBMK) or CO2 as a coolant. These two types are not being built anymore and the remaining ones will be shut-down in the near future because of safety concerns and bad economics. Modern graphite moderated reactors employ helium as a coolant and a special fuel design, called TRISO coated fuel particles (see later on).

Despite the favorable characteristics of U-235 for thermal-neutron induced fission, the fuel in thermal reactors needs to be enriched in U-235 to be able to achieve a stationary fission chain reaction. LWRs usually operate with a fuel enrichment of 4-5%, which allows them to operate for 12 to 18 months without refueling and without fuel damage due to high burnup (a parameter quantifying the amount of energy generated per unit mass of fuel).

The class of reactors in which the neutrons remain at high energy (E>100 keV) is called fast reactors. Because the probability for neutron absorption, and thus for neutron induced fission, in the high-energy region is low, the fissile fuel fraction (either U-235 or fissile plutonium isotopes) should be 15-20%. Due to the larger number of fission neutrons released, a fast reactor can be designed to produce more fissile nuclides than are consumed during operation. This sub class of fast reactors is called fast breeder reactor. It is clear that reactors with a fast neutron spectrum lack a moderator in the core, and that they need a non-moderating coolant like sodium or lead-bismuth, which both have a high atomic mass, or helium (low density). Breeder reactors with a thermal neutron spectrum are only feasible if they use thorium as a fertile material and U-233 as a fissile material. The Molten Salt Breeder Reactor was based on this concept.

Scheme of the Molten Salt Breeder Reactor (MSBR).

Safety of Nuclear Reactors

Reactor safety is based on a defense-in-depth strategy with at least three barriers:

  1. Measures to prevent the occurrence of incidents.
  2. Measures to control incidents.
  3. Measures to reduce the impact of incidents.

As mentioned before, the decay of fission products is both a blessing and a curse for safe reactor operation. Provisions should be made for the fission product decay heat to be transferred to the environment without any damage to the fuel in case of a loss of cooling incident. In present-day reactors, this is mainly accomplished with active means, like multiple emergency core cooling systems with various redundancies built-in. Inherently safe reactors, like the modern graphite-moderated gas-cooled High Temperature Reactor (HTR), rely much more on physics principles like heat conduction and radiation from the core to the surroundings, cooling of the reactor structures by density-driven natural circulation of air, and a high-integrity fuel design that can withstand extremely high temperature. Consequently, such a reactor can easily withstand a loss of cooling incident, making it a normal operation condition instead of an accidental one.

Decay heat
Decay heat as a function of time after shut-down of the reactor.

Decay of precursors releases delayed neutrons that are essential for the safe operation of any nuclear reactor. Although the number of delayed neutrons is very small compared to the promptly-generated fission neutrons (0.7%), the fission chain reaction would die out without them. This is accomplished by absorbing so many neutrons that the delayed neutrons are essential to maintain a stationary fission chain reaction. The reactor is called critical then, and the reactivity of the reactor, a measure for the surplus of neutrons in one generation compared to the previous one, is said to be zero. When the reactivity increases (becomes positive), the state of the reactor becomes supercritical. For reactivity levels smaller than the delayed neutron fraction, the delayed neutrons are still needed to maintain the fission chain reaction and the reactor is called delayed supercritical. The power increases but with a time period of tens of seconds, which is easily manageable by automatic control systems. For higher levels of reactivity (larger than the delayed neutron fraction), the reactor becomes critical with prompt neutrons alone. This situation is called prompt supercritical and has to be avoided under any circumstance as the reactor period is extremely short (on the order of milliseconds) and a power excursion cannot be avoided. Serious incidents that increase the reactivity of the reactor are called Reactivity Induced Accidents (RIA). Inherently safe reactors should be able to withstand any RIA without SCRAM (use of shut-down control rods) or designed such that prompt criticality cannot occur.

Inherently Safe Feedback Mechanisms

In practice, control of a nuclear reactor is enhanced by several negative feedback mechanisms relying on physical phenomena.

  1. A raise of the fuel temperature increases the neutron absorption rate. This so-called nuclear Doppler effect originates from the fact that heavy nuclides exhibit a strong resonance behavior for neutron absorption in the energy range between the fast region where fission neutrons are 'born' and the thermal region where they induce new fissions. The width of the absorption resonances is very small and the probability for resonance absorption very high. When the thermal motion of the fuel nuclides becomes stronger, part of them will move towards the incoming neutron, while another part moves away. As a result, neutrons with higher energy as well as neutrons with lower energy will also be absorbed in the resonance. Effectively the resonance for neutron absorption widens and flattens such that the total area under the resonance remains constant (the total number of fuel nuclides does not change; only the velocity distribution due to thermal motion does). Fortunately, the increase of neutron absorption in the resonance wings exceeds the decrease at the center energy, as in that region the probability for neutron absorption remains close to 100%. As a consequence, any fuel with a substantial amount of fissionable actinides like U-238 and Th-232, will exhibit a strongly negative Doppler feedback effect.
  2. A raise of the moderator temperature reduces the moderation power. For thermal reactors with a fluid moderator, an increase of the temperature leads to a lower density and a smaller scatter probability. Hence the moderating power reduces, the neutron spectrum becomes harder and the neutron balance deteriorates. As a consequence the fission chain reaction dies out. Of course, this effect is much stronger in fluid moderators than solid ones, and does not hold for strongly absorbing moderators. In the latter case, a reduction of the moderator density will decrease absorption so much that the net reactivity effect becomes positive. Usually moderators absorbing neutrons that much are not used in nuclear reactors.
  3. An increase of the fuel temperature leads to an expansion of the core and an increase of neutron leakage. In a fast reactor core, the density of fissile nuclei must be high to maintain a critical reactor. When the fuel expands, the fuel density decreases and neutrons have a larger probability to escape from the core without interaction. Although this might be against intuition (for a thermal reactor, a larger core usually means less leakage), it is one of the dominant feedback effects in fast reactors.

All these feedback mechanisms do not rely on active systems or operator action and cannot be halted. Therefore they are considered to be inherently safe. They usually facilitate control of a nuclear reactor in case of load-following operation. If more electricity output is needed, the secondary cooling system will cool down leading eventually to a reduction of the primary coolant temperature at the core inlet. Due to cooling of the moderator and/or the fuel, the reactivity of the reactor core increases and more power is produced such that the average core temperature (roughly the average of the inlet and outlet temperatures) restores to its preset value. Through this mechanism, the core power automatically adapts within small margins.

Fuel Burnup and Reactivity Control

During reactor operation, fissile nuclides are consumed and fission products are produced. The net effect is a reduction of the reactivity of the fuel, leading to a sub-critical core and a cool-down and eventually shut-down of the reactor. Therefore, in most reactors, more fuel is loaded at the Beginning of Life (BOL) and the excess fission neutrons are captured by control rods containing a strong neutron absorber like boron. In Pressurized Water Reactors (PWRs), boron might even be dissolved in the coolant, although only to a limited extend in order not to deteriorate the negative moderator feedback coefficient.

The introduction of excess reactivity at BOL is a potential risk that needs extra safety measures. Control rod drive mechanisms can fail and control rods could, at least theoretically, be ejected from the core leading to a sudden strong increase of the reactivity. Therefore, the reactivity value of each control rod should preferably be less than the delayed neutron fraction. To avoid such accident initiators, some reactors compensate the over reactivity at BOL by admixing absorbing nuclides in the fuel or by adding burnable absorber pins to the fuel assemblies. The burnable poison distribution needs to be designed accurately to avoid some left-over at the end of the irradiation; otherwise the fuel life time would be unnecessarily shortened leading to an economic penalty. Modern pebble-bed HTRs continuously add fresh fuel and unload spent fuel to compensate for fuel burnup.

During reactor operation, not only fissile nuclides are consumed; they are also produced by neutron capture in fertile nuclides like U-238. Consequently fuel lifetime and fuel burnup usually is much higher than anticipated from the initial fuel enrichment. In breeder reactors, more fuel is produced than consumed, and a much higher fuel burnup could be achieved. Actually the first reactor that produced usable amounts of electricity, the Experimental Breeder Reactor (EBR-1) in Idaho, already achieved breeding.

Fuel Design

Most thermal reactors in operation today employ low enriched uranium (4-5%) as a fuel in the form of sintered UO2 pellets. For a PWR, these pellets have a diameter of 8 mm and a height of approximately 1 cm. They are stacked in a metal cladding, and assembled together in a grid of 17x17 fuel pins. A few hundred of these fuel assemblies make up the reactor core. Control rods can be inserted from the top to control the reactivity during reactor operation. A Boiling Water Reactor (BWR) works similarly, except that the fuel pins are more robust to withstand the boiling processes in the core, and that the control of the reactor is done by neutron-absorbing blades that are inserted between the fuel assemblies bottom-up. Besides control rods or blades, reactivity is partly controlled by boron-acid in the moderator (PWR) or by burnable poison pins (BWR).

Only a few fast reactors have been in operation till now. They all employ(ed) fuel pins composed of low enriched uranium (15-20%) or MOx fuel (about 20% plutonium oxide admixed through uranium oxide made of natural or depleted uranium) cooled by some liquid heavy metal. Reactors have been designed with metal fuel and with nitride fuel as well, both having the advantage of having a higher density and better heat conduction. Advanced Gas Cooled Fast Reactors (GCFR) are being designed using a ceramic fuel like uranium/plutonium carbides admixed with SiC. These carbides can withstand extremely high temperatures.

Modern HTRs employ TRISO coated fuel particles, which consists of a UO2 fuel kernel made of low enriched uranium (about 10%) covered by a porous carbon buffer layer, a pyrocarbon layer, a SiC layer and another pyrocarbon layer. All together the diameter of the particle is less then 1 mm. The porous layer accommodates the gaseous fission products, while the three other layers maintain integrity up to 1600 °C. Reactors using this fuel can be made inherently safe, if decay heat transfer to the environment is sufficient to keep fuel temperatures mild. In a thermal HTR, heat conduction to the environment is ensured by the graphite moderator and the maximum fuel temperature during a loss-of-cooling incident remains well below 1600 °C. This is the main safety feature of HTRs. In GCFRs, TRISO coated fuel particles can be used as well, although they would have to be cooled directly [Wilfred]. Inherently safe decay heat removal is not guaranteed then.

HTR fuel design
Fuel design of modern High Temperature Reactors.

The Molten Salt Reactor (MSR) employs a fuel dissolved in the coolant itself, which consist of a fluoride salt (LiF/BeF or NaF/ZrF mixtures). These salt mixtures have a very high boiling point (>2000 °C), but also a relatively high melting point (400-500 °C), which requires electrical heating for the entire primary and secondary circuit. This offers clear advantages like excellent heat transfer from the fission process to the coolant, and strong inherently safe feedback mechanisms, because expansion of the coolant automatically reduces the fuel inventory. Furthermore, fission products can be easily removed from the salt and fuel can be injected into the primary system during reactor operation. Fluoride salts have attractive properties for heat transfer and removal. Recently, they have been proposed for coolants in graphite moderated HTRs with TRISO coated fuel particles.

Generations of Reactors

Besides a functional categorization of reactors, they can also be classified in order of appearance. The first-builds are classified Generation 1 and span the period from the 1950's till the mid sixties. Then Generation II appeared up to the mid-eighties of the last century. The bulk of reactors in operation today belong to this class, although the safety level of these reactors in many cases are upgraded to Generation III level. The latter generation of reactors makes use of more passive safety features. For BWRs, for example, generation II uses a few (typically only two) externally placed primary coolant pumps, while in generation III the primary circulation is maintained by tens of smaller pumps positioned internal the reactor vessel. In Generation III+ BWRs the primary coolant circulation is driven by density differences between the coolant in the riser (the column above the core) and in the down comer. No primary pumps are needed then. The Economic Simplified BWR (ESBWR) operates according to this principle and is scheduled for licensing in 2010. Some generation III+ PWRs are already licensed and being built.

Generations of reactors
Evolution of nuclear reactors from Generation I to Generation IV.

Since 2000, research is being carried out on six reactor types scheduled for licensing between 2020 and 2040. These are called Generation IV reactors, and range from a graphite-moderated gas-cooled HTR with Very high outlet temperature (VHTR), the Super-Critical Water Reactor (SCWR) operating with super-critical light water as a coolant and moderator, and several fast reactors designed with liquid heavy metal or helium as a coolant. The most advanced is the MSR, employing a molten salt as a coolant with the fuel dissolved in it. The research for these Generation-IV reactors is coordinated world-wide by the Generation-IV International Forum (GIF). Since the start, major research contributions have been made in the fields of coolants, fuels, and reactor concepts.