Nuclear Energy and Environmental Impact

Abstract

Nuclear energy is attracting revived interest as a potential alternate for electric power generation in the event of increased concerns about global warming. Compared to energy produced by combustion of a carbon atom in coal, fission of a U-235 atom will produce about 10 millions times more energy. However, storage of the nuclear waste is an environmental issue. This chapter has four sections with a major focus on introduction of nuclear power plants and reprocessing of spent nuclear fuels. Different nuclear fuel cycles and nuclear power reactors are introduced in the first section, and the cost–benefits of different energy sources are compared. Fuel burnup and formation of fission products are discussed along with operational impacts and risk analyses in the second section. The third section discusses design of nuclear structural components and various degradation modes. Section four discusses reprocessing issues of nuclear spent fuels. Reprocessing of spent nuclear fuel may be an economically viable option and reduces high-radioactive load in the nuclear waste repositories as well. However, there is a concern about proliferation of weapons-grade plutonium separated during reprocessing. Containment of radionuclides in different waste forms is also discussed in this section.

These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

Introduction to Nuclear Energy

Radioactive decay of heavy metals such as uranium, plutonium, thorium, etc., can be converted into a useful energy form. Radioactivity occurs by emission of charged particles (such as α and β) and electromagnetic waves (γ ray). For heavier nuclei (elements with atomic number>40), more neutrons are required for a stable configuration so that the electrostatic repulsion force between the protons can be overcome (Jevremovic 2005). When the nucleus has too many or too few neutrons, it will be in a nonequilibrium condition. In order to reach a stable configuration, the nucleus undergoes a spontaneous transformation by rearranging its constituent particles. This is accomplished by the emission of an alpha particle, a beta particle (either β− or β+), a neutron, or a proton. Depending on the energy conservation, gamma radiation may or may not be present during the radioactive decay. In brief, when atoms containing nuclei in the nonequilibrium condition try to reach stable condition, the excess energy of the nuclei is emitted as radiation. In this process, the material disintegrates. According to Einstein’s principle (E = mc²), the disintegrated matter is converted into energy. For example, burning of 1 kg of uranium in a nuclear reactor results in conversion of 0.87 g of matter into energy which amounts to (0.8 × 10⁻³ kg) × (3.0 × 10⁸ m/s)² = 7.8 × 10¹³ J. For comparison, combustion of 1 kg of gasoline will release only 5 × 10⁷ J of energy, six orders of magnitude smaller than 1 kg of uranium (Murray 2001). In addition to high specific energy, the nuclear energy has an advantage of not releasing carbon dioxide into the atmosphere. Combustion of 1 kg of gasoline will release about 3.2 kg of carbon dioxide to the environment. An anthracite coal–based power plant will release about 1.2 kg of CO₂ for every KWh electricity generated, whereas the lifetime CO₂ emission of nuclear power plants, considering the electricity used for mining and processing operations from fossil fuel power plants, will be 100–140 g of CO₂ /kWh electricity generated (Storm van Leeuwen and Smith 2005).

The major advantages of nuclear energy are:

• High specific energy

• No CO₂ emission

• Spent fuel can be reprocessed and reused, thus conserving natural resources

• Possibility to produce more nuclear fuel than consumed by using fast breeder reactors

• Lower operating cost in terms of fuel cost compared to fossil fuel power plants

Disadvantages are:

• Large capital cost and longer construction time of power plants

• Long-term storage of nuclear waste which is an issue

• Exposure to radioactivity in case of accidents

• Potential proliferation of weapons-grade fuel during reprocessing

Nuclear power plants attract more safety and environmental concerns from the public than other power plants. This chapter addresses some of the environmental issues associated with nuclear power generation. The first three sections introduce nuclear fuel cycles, nuclear power reactors, and issues on operational safety. Information on nuclear spent fuels reprocessing, waste management, and long-term storage is given in the last section.

Nuclear Fuel Cycles

Conversion of nuclear energy can be achieved by fission or fusion reactions. Most of the commercial nuclear power reactors operate based on nuclear fission reaction. The average energy of neutrons used for power generation is about 0.1 eV, which are called thermal neutrons. Neutrons that have energy in the order of 2 MeV are called fast neutrons. Uranium is the most common fissile material used in the nuclear reactors. Naturally mined uranium has 99.24 % U-238, 0.72 % U-235, and 0.0054 % U-234. U-235 is a fissile isotope. Fissile isotopes are the ones that undergo fission reaction upon absorption of slow neutrons (neutrons having energy <0.4 eV). When a neutron is absorbed by U-235, the isotope gains extra energy and transforms to U-236 in an excited state (Murray 2001):

$$ {}_{92}{}^{235}U+{}_0{}^1n\to {}_{92}{}^{235}U* $$

(1)

Since the excited U-236 has a higher mass than U-236 in ground state, the difference in mass is converted into energy of 6.54 MeV. Each fission of U-235 yields about 190 MeV of useful energy. This energy is used for further fission reaction. A continuous and self-sustaining fission chain reaction is required for nuclear power generation. This is achieved by containing a critical mass of uranium, which is about 50 kg of U-235 in the nuclear reactor. For each megawatt-day of thermal energy, 1.3 g of U-235 is consumed (Knief 1992).

The natural isotope U-235 and artificial isotopes such as Pu-239 and U-233 require only slow (thermal) neutrons to induce fission. On the other hand, U-238, which is abundant in nature, can only be activated by fast neutrons of at least 0.9 MeV to initiate fission. If the fission occurs by thermal or slow neutrons, the material is considered as “fissile.” If a material is converted into a “fissile” material by irradiation, then that material is classified as a “fertile” material. For example, fertile materials such as U-238 and Th-232 can be used for generating “fissile” materials such as Pu-239 and U-233, respectively, by the following neutron reactions (Murray 2001):

$$ {}_{92}{}^{238}U + {}_0{}^1n\to {}_{91}{}^{239}U\to {}_{93}{}^{239}Np+{}_{-1}^{\kern0.5em 0}\kern-0.2em e\to {}_{94}{}^{239}Pu+{}_{-1}^{\kern0.5em 0}\kern-0.2em e $$

(2)

$$ {}_{90}{}^{232}Th+{}_0{}^1n\to {}_{90}{}^{233}Th\to {}_{91}{}^{233}Pa+{}_{-1}^{\kern0.5em 0}\kern-0.2em e\to {}_{92}{}^{233}U+{}_{-1}^{\kern0.5em 0}\kern-0.2em e $$

(3)

Nuclear fuel cycle describes different stages of preparation, use, safe storage, and reprocessing of the spent fuel. The fuel cycle starts by mining of fuel ore and ends by reprocessing. If reprocessing is not carried out, then it is called as once-through nuclear cycle. The fuels that can be used in the nuclear power reactors are uranium, plutonium, minor actinides, and thorium. Most of the commercial nuclear reactors use uranium as a fuel with different levels of enrichment depending on the type of reactor. Thorium is also used in some countries at a small scale. This section will focus on only these two fuels.

Uranium Fuel Cycle

Mining and Extraction: Uranium ore is mined in several forms, most notably as uranite (UO₂), pitchblende (UO₃ + U₂O₅), coffinite (U(SiO₄)_1−x(OH)_4x), brannerite (UTi₂O₆), davidite (REE(Y,U)(Ti,Fe)₂₀O₃₈; REEs = rare earth elements), and thucolite (U-containing pyrobitumen) (Dahlkamp 1993). Uranium mining process is similar to mining of other metals such as copper, gold, etc. The mines can be of open-pit or underground type. In some cases, in situ leaching of ore is carried out without deep excavating of the earth. Lower-grade ores are concentrated by using a heap leaching process. The excavated ores are collected as heaps and a leaching agent (mostly low concentration of sulfuric acid) is sprayed on the heaps and drained through the ore collection. In this process, uranium is preferentially leached out of the oxide ore and carried by the draining leaching agent. This solution is further processed.

In situ leaching process is used by some mining companies. In this process, the surface of the earth is not disturbed by drilling or excavating operations. Instead, a leaching agent (mild acid/alkaline solution containing oxidizers such as hydrogen peroxide) is pumped through the porous surface of the ore deposits. The solution passing through the ore will leach out the uranium. The uranium-containing solution is collected for further processing.

During heap leaching and in situ leaching processes, the surrounding groundwater is continuously monitored for any contamination. Even after shutdown of the mining operation, the monitoring continues. Recent international mining laws require that the mining companies set aside required amount of funds for reclamation of the environment in the neighborhood of the mining operation. This fund will be in place for the required environmental remediation even if the company goes out of business.

The uranium-rich ore from mill, U₃O₈, often called as yellow cake though the actual color is khaki, is purified further either by a solvent extraction or by an ion exchange process. The U₃O₈ is dissolved in nitric acid to form uranyl nitrate (UO₂(NO₃)₂), which is then treated with ammonia to produce ammonium diuranate ((NH₄)₂U₂O₇). This compound is then reduced in a hydrogen atmosphere to form uranite (UO₂). It should be noted that this UO₂ is not the final form of fuel that can be used in the reactor. It requires enrichment of U-235 isotope to more than 3.5 % from 0.7 %. In order to accomplish the enrichment process efficiently, the material needs to be in the form of a gas. Therefore, UO₂ should be first converted to uranium hexafluoride gas (UF₆), get enriched, and converted back to UO₂. In order to form UF₆, UO₂ is treated with HF, which results in UF₄. Further fluorination of UF₄ renders UF₆. Now the UF₆ is taken for enrichment process (Morss et al. 2006).

Enrichment of Uranium

U-235 is a fissile material that utilizes thermal neutrons, which is only 0.7 % in the natural uranium source. Stable operation of a critical reactor requires fissile nuclei. Therefore, the ratio between fissile (U-235) and nonfissile (U-238) should be high in a reactor for stable power generation. Typically, the U-235 content in a fuel should be more than 3.5 %. This can be achieved by the following enrichment methods (Krass et al. 1983):

• Gaseous diffusion

• Gas centrifuge

• Jet nozzle/aerodynamic separation

• Electromagnetic separation

Gaseous Diffusion

Separation of U-235 isotope by gaseous diffusion process occurs because of the difference in velocity of the gas molecules. The rate of diffusion of a gas through an ideal porous medium is inversely proportional to the square root of the molecular weight of the gas. Therefore, when UF₆ is passed through a porous tube, lighter U-235 isotopes will escape the porous container faster than their U-238 counterparts and thus be collected separately. Figure 1 schematically illustrates the working principle of a gas diffusion separator. In this arrangement, compressed UF₆ gas is contained in a large vessel under pressure and allowed to pass through a porous channel that acts as a diffusion barrier layer. The porous layer is an essential component of the diffusion separator because the efficiency of the U-235 enrichment process in the UF₆ stream is determined by the ability of the porous layer permeating the gas molecules. The barrier layer has to withstand the working pressure and be stable in the corrosive UF₆ atmosphere, and the pores should be tiny, in the order of 30–300 times the diameter of a single U atom, around 10 nm, and uniformly distributed in the order of billions per square centimeter area. The barrier layer thickness is around 5–6 mm. The material of construction is a classified information because enrichment of U-235 to more than 80 % becomes a weapons-grade material. Plasma-sprayed Zr, Ta, and Mo coatings of required thickness and porosity can be used as diffusion layer. Silver–zinc alloy pipes after selective etching with HCl also could serve as a diffusion barrier (Makhijani et al. 2004).

Fig. 1

Uranium enrichment by gaseous diffusion process

Since the velocity of diffusion varies only by 0.4 % between U-238 and U-235, multiple stages of gas passage are required to achieve the required enrichment for commercial reactors. In the diffusion cascade, the outlet of the depleted stream is bifurcated by redirecting 50 % of the depleted stream back to the previous stage and the remaining 50 % to the inlet of the next stage. This way one may require 4000 diffusion stages to obtain 99 % ²³⁵UF₆.

Gas Centrifuge

In this process, UF₆ is passed through rapidly spinning cylinders in series. The centrifugal action drives the heavier ²³⁸UF₆ molecules to the wall of the rotating container. Lighter ²³⁵UF₆ molecules are retained closer to the center of the container. Circulation of the gas from bottom to top helps separate the heavier and lighter molecules and pass to next separation stages. Figure 2 schematically illustrates the gas centrifuge (Olander 1978).

Fig. 2

Schematic of gas centrifuge for enrichment of U-235

The cylinders are heavy in order to impart very high kinetic energy to the gas molecules and rotated at very high speed, in the order of 100,000 revolutions per minute. The linear velocity may approach the speed of sound in the material of construction. Therefore, the centrifuges are very sturdy. Nevertheless, this process requires 40–50 times less energy than the gaseous diffusion process to achieve the same level of enrichment. The diffusion process will consume about 4–5 % of the energy that the enriched fuel will generate in its cycle time. Furthermore, amount of waste heat generated during UF₆ compression for the diffusion process is much higher. This results in utilization of a significant amount of coolants such as Freon-12, etc., in order to cool the gas in the intermediate compression stages. Gas centrifuge process is more advantageous than gas diffusion process considering aforementioned issues.

Electromagnetic Isotope Separation

This process can be used as a stand-alone method for enriching uranium from the feedstock of natural uranium or in tandem with the gaseous diffusion process for producing highly enriched uranium from the low-enriched stream. It should be noted that commercial reactors do not require highly enriched uranium. The electromagnetic separator works on the principle that the radius of the trajectory of a particle traveling in a magnetic field is determined by its mass. The heavier the particle, the larger will be the radius, provided that the particles have the same charge and travel at the same speed. In this process, UCl₄ plasma is generated by heating a solid UCl₄ material and bombarding it with high-energy electrons. This process ionizes uranium. The uranium ions are then accelerated through a strong magnetic field. When these ions complete a half circle, the lighter U-235 ions are nearer to the inside wall of the semicircle channel, and heavier depleted U-238 ions are separated at the outer wall of the circular channel. This process requires a large amount of energy to maintain a high magnetic field. However, the rate of U-235 separation is lower than the other processes (Knief 1992).

Jet Nozzle/Aerodynamic Separation

In this process, UF₆ is pressurized with helium or hydrogen gas and sent through a bank of small circular pipes. The curved pipelines ensure that the heavier molecules stay at the outer surface and lighter U-235 stay at the inner surface of the circular path. This process is again energy intensive because compressing lighter gases such as He and H requires more energy. The compression process will be carried out at multiple stages with intercoolers in order to increase the thermal efficiency of the system. Therefore, the entire process results in generation of a significant amount of waste heat and utilization of a large amount of coolants/refrigerants, and both these outcomes lead to environmental concerns (Olander 1978).

Conversion of UF₆ to UO₂

The enriched UF₆ gas stream with 3.5 % U-235 is taken for fuel rod preparation. The first step is conversion of UF₆ back to UO₂. Oxide fuel is preferred because of its high melting point that shows resistance for meltdown. Metallic fuels (in the form of U or its alloys such as U–Mo, U–Si, U-Zr, etc.) have been used in some experimental fast breeder reactors. Most of the commercial reactors use oxide fuels such as UO₂ or mixed oxide fuels such as a blend of Pu-239 and depleted uranium (Galkin et al. 1982).

UF₆ can be converted to UO₂ using three different chemical routes such as:

• 1. UF₆ is reacted with hydrogen and steam to form UO_2.

• 2. UF₆ is sent through water and thus hydrolyzed. NH₄OH is added to this hydrolyzed solution to precipitate ammonium diuranate. This precipitate is then reduced in hydrogen at 820 °C to form UO_2.

• 3. In this method, a gaseous mixture of CO₂ and NH₃ is hydrolyzed in water to form ammonium uranyl carbonate. This precipitate is then treated with steam and hydrogen at 500–600 °C to form UO₂.

The UO₂ is then purified by several washing cycles, dried, and mixed with an organic binder to press as cylindrical pellets. The compacted pellets are sintered at high temperature (1400–1700 °C). The solid UO₂ pellets are machined by fine grinding operation to final dimensions which are typically about 1 cm in diameter and 1.5 cm in length depending on the type of reactor (boiling water or pressurized water).

Fuel Rod Assemblies

The cylindrical fuel pellets are stacked in a fuel-clad tube with a nominal diameter of 1 cm and length of 3.6 m. Then the tubes are backfilled with helium at about three atmospheres to improve the thermal conductivity and heat transfer. The fuel rods are bundled with a thin encasing tube to prevent density variation that may affect the thermohydraulics of the reactor core. In a typical BWR fuel bundle, there are about 96 fuel rods. The BWR reactor core contains about 360–800 fuel assemblies depending on the capacity of the nuclear reactor. In PWR reactors, the fuel assembly contains a matrix of 14 × 14 and 17 × 17 fuel rods. The PWR bundles are about 4 m long. The reactor core contains about 120–193 bundles, depending on the capacity of the reaction. In case of heavy water reactions, U-235 enrichment is not required. Natural uranium (U-238) is loaded into the fuel bundles that have 380 tubes (Knief 1992).

Thorium

The isotope ²³²Th is about 3–4 times more abundant on earth than uranium. This fertile isotope can be used for producing fissile ²³³U isotopes. The neutron absorption cross section of ²³²Th is also about three times that of U-238 (7.4 barns vs. 2.7 barns, where 1 barn = 10⁻²⁴ cm²). The number of neutrons liberated per neutron absorbed is greater than 2.0 over a wide range of thermal neutron spectrum for U-233. Therefore, the ²³²Th–²³³U fuel cycle can operate in wide spectra of neutrons such as fast, epithermal, and thermal. Thorium oxide is chemically more stable than UO₂ and has a low fission product release rate. ThO₂ also has a better thermal conductivity than UO₂. Furthermore, (Th, Pu)O₂ mixed oxide fuel is more attractive than (U, Pu)O₂ fuel because Pu is not bred in the (Th, Pu)O₂ and the presence of ²³²U makes the spent fuel more proliferation resistant (Thorium fuel cycle).

Some of the limitations of ThO₂ fuel are as follows: Melting point of ThO₂ is 3350 °C. Therefore, the sintering temperature is higher than 2000 °C. Reprocessing requires heavy shielding because of the presence of remitting ²³²U with 73.6 years of half-life. Because of its high chemical stability, ThO₂ cannot be easily dissolved in HNO₃ for reprocessing. It requires addition of .005 M HF to 13 M HNO_3, which makes the stainless steel containers used in reprocessing more prone to corrosion attack even after adding 0.1 M Al(NO₃)₃ as inhibitor.

Fuel Burnup

Fuel burnup gives information about the fuel utilization as how much energy has been extracted from the fuel as a nuclear fuel source. It is expressed as megawatt-days per metric ton (MWd/ton). The generation (II) commercial reactors were designed for a burnup of 40 GWd/ton. During nuclear fission, fission products build up and poison the sustainability of the chain reaction. For example, xenon-135 (cross section 2 million barns), samarium-149 (75,000 barns), and Gd-157 (200,000 barns) poison the reactor core which is a serious problem (Knief 1992).

In order to operate the reactor for longer time/cycle, between shutdown for fuel change, excess fuel is added. This high reactivity of the excess fuel needs to be balanced by neutron-absorbing materials in addition to control rods. These are called burnable poisons. Compounds of B and Gd act as burnable poison whose neutron absorption capacity decreases with fuel burnup. By using neutron poison, the burnup of fuel in a modern reactor can achieve a burnup of 60 GWd/t (Murray 2001; Knief 1992).

Higher-enriched fuel in the advanced light water reactors can achieve a burnup rate of 90 GWd/t. Since the fast neutron reactors can handle fission product accumulation without affecting its performance, the burnup rate can be around 200 GWd/t. The deep burn modular helium reactor that utilizes ceramic-coated plutonium-based fuel can achieve a burnup rate of around 500 GWd/t (Rudling et al. 2008).

Types of Nuclear Reactors

Depending on the coolant and moderator, the reactors can be classified as:

• Light water reactor

  • Boiling water reactor

  • Pressurized water

• Heavy water moderated reactor

  • CANDU

  • Advanced heavy water reactors

• Graphite-moderated reactors

• Thermal breeder reactors

  • Molten salt breeder reactor

  • Light water breeder reactor

• Fast neutron reactors

  • Liquid metal fast breeder reactor

  • Gas-cooled fast breeder reactor

Boiling Water Reactors

The original design was developed by the General Electric Company. Figure 3 schematically illustrates the major components of a BWR system. Different power generation capabilities are 200 MW_e, 650 MW_e, and 1250 MW_e. Here, the subscript “e” in the MW_e represents electrical energy as the output. The capacity is also expressed in terms of thermal energy as output. The efficiency of converting thermal energy into electrical energy is around 30 %. Therefore, for a given MW_e output, the corresponding thermal energy will be at least three times higher. It should be noted that feedwater enters the reactor vessel at a pressure of about 70–85 bar and leaves as steam at 288 °C to run the steam turbines. The active core height is about 3–8 m that is placed in a reactor vessel of 22 m height. The bottom of the vessel is occupied by the control rod and drive mechanism. The space above the reactor core is occupied by the steam separator–dryer system components. The inner diameter of the vessel is about 6.4 m (for 1250 MW_e reactor). The wall of the reactor is made of 15 cm thick carbon steel clad with stainless steel. The stainless steel surface is exposed to the water/steam phases (Murray 2001; Knief 1992). Since a single coolant loop is used for heat transfer and steam generation, the water used in the BWR system is of ultrahigh purity with an electrical resistance of 18.2 megohm-cm. In order to minimize corrosion, the dissolved oxygen in the water is controlled below 200 parts per billion (ppb) by purging with ultrahigh-purity argon or hydrogen. The dissolved oxygen content of water exposed to ambient atmosphere is around 8 parts per million (ppm). Purging with nitrogen could lead to formation of NO_x-related compounds by radiolysis. Very high electric resistance of the water and low oxygen content result in low corrosion rates of the pressure boundary components.

Fig. 3

Schematic diagram of system flow in a boiling water nuclear reactor. Only critical components are shown

Pressurized Water Reactors

Developed by Westinghouse Electric Company, they are based on the experience of nuclear submarine reactors. The power generation capability of PWR ranges from 60 MWe to 1450 MWe. Figure 4 schematically illustrates the major components of the PWR system. The PWR has two coolant circulation loops. The primary loop that removes heat from the reactor core is contained within the reactor vessel and used for transferring heat to a secondary coolant loop. Steam formation takes place in the secondary loop which runs the steam turbine for power generation. The water of the primary loop contains 1200–1800 ppm boron in the form of boric acid. Boron is added as moderator/neutron absorber. In order to balance the acidic pH due to boric acid, LiOH is added in the water to maintain a neutral pH at room temperature. The addition of boron and lithium compounds increases the ionic conductivity of the primary loop water. This may lead to increased corrosion rate. Since oxygen reduction reaction is the cathodic reaction in most of the corrosion mechanisms, the dissolved oxygen of the primary water is reduced to less than 5 ppb by purging with ultrahigh-purity hydrogen.

Fig. 4

Schematic diagram of system flow in a pressurized water nuclear reactor. Only critical components are shown

The primary loop contains lithiated and borated water at a pressure of about 150 bar (atmospheres). The core outlet temperature of the primary loop water is at about 350 °C. This primary loop heats the water in the secondary loop to above 290 °C and 70–85 bar. The secondary loop conditions are identical to that of BWR steam conditions. The main difference between the PWR and BWR is that the steam entering the turbines of PWR does not contain any radioactivity. A typical PWR reactor vessel is about 13 meters tall and 6.2 meters in diameter. The wall of the vessel is made of 23 cm thick ferritic–bainitic steel with 3 mm stainless steel clad that is exposed to the water/steam environment. The secondary loop, also called the steam generator, is 21 meters high and 4.5 meters in diameter (Murray 2001; Knief 1992).

Heavy Water Moderated Reactors

In light water reactors, the water is used as a moderator and coolant. An ideal moderator will have low mass, high neutron scattering cross section, and low neutron absorption cross section. Light water more rapidly moderates neutrons than heavy water. Therefore, fissile material such as U-235 is required for sustaining the chain reaction. In case of heavy water moderated reactors, high-energy neutrons are available for fission of U-238. Therefore, enrichment of U-235 is not required. However, preparation of heavy water is also a tedious process since deuterium is available in nature with a ratio of 1:7000 along with regular hydrogen in water. The separation of deuterium from water is an energy-intensive process.

In CANDU-PHW system, the primary heavy water coolant is pumped through a series of pressure tubes that also housed the fuel rods. The heavy water is contained in a vessel called calandria, which is about 7.6 m in diameter and about 4 m tall. There are about 380 calandria tubes. Each calandria tube contains one pressure tube through which the coolant flows (CANDU Reactors).

Graphite-Moderated Reactors

These reactors are fueled by natural uranium and moderated by graphite, and the coolant is an inert gas. Advanced gas-cooled reactors use helium as a heat transfer medium. Helium flows through the reactor core and exits at 740 °C with a pressure of about 50 bar. The heated gas is fed through the steam generator which produces superheated steam at 510 °C under a pressure of 170 bars. The reactor vessel is of prestressed concrete strengthened with vertical and circumferential prestressing steel cables. The dimensions of the vessel are 28 m high × 30 m in diameter. The reactor core size is about 6.3 m × 8.5 m. There are six sets of helium loops and steam generators that produce about 1160 MWe power. Running the generators by gas turbines using the helium is possible (Murray 2001; Knief 1992).

Nuclear Reactor Safety

Radiation Effect

Three types of radiation damage can occur when high-energy particles impinge on a target. These are (Shapiro 1990):

• Transmutation of the constituent atoms of the target to new atoms of new material

• Displacement of atoms from their normal lattice positions in the structure of materials

• Ionization–removal of electrons of atoms that are present in the trajectory of the charged particles and formation of ion pairs

Alpha and beta particles have only low penetrating power and can be stopped by very thin layer of materials. Alpha particles produce a linear path of specific ionization per unit length of travel because of its high mass and charge. Smaller path and charge of particles result in a nonlinear path and very low specific ionization. Gamma rays have higher penetrating power than the alpha and beta particles. Therefore, a heavy shielding is required to stop gamma radiation.

Accumulation of fission products in the core of the nuclear reactor poses a potential safety threat among the public. Therefore, integrity of the fuel is important throughout the operating cycle. A set of specified operating parameters limits is determined for every reactor to assure safety of operation. These limits are generally the upper limitation on total reactor power which determines the maximum reactor core temperature that will not result in a meltdown. Fuel integrity is ensured by avoiding hot spots in the core. This is carried out by controlling the ratio of peak power to average power. The other important parameters that control the safe operation of the nuclear reactor core are the limit on the control rod position, difference between the power generation in the bottom half of the core and the top half, symmetry of power generation across the core, maximum reactor coolant temperature, minimum coolant flow, and maximum system pressure.

Under normal operating conditions, a negligible amount of radioactivity will get into the coolant (Murray 2001). Since abnormal conditions can exist, a design basis accident is postulated. In this hypothetical accident condition, a loss of coolant is assumed. Loss of coolant can occur either by a breakage of coolant piping or a failure of jet pumps supplying the coolant. Loss of coolant would result in an increase in the temperature of the nuclear core. When the temperature exceeds the melting point of fuel and cladding, the fuel tubes will be damaged and fission products will be released. The reactors are provided with safety measures such as safety rods and emergency core cooling system (ECCS) to safeguard the reactor against loss of coolant accident. The safety rods reduce the reactor power immediately upon loss of coolant flow through the reactor core. The emergency core cooling system consists of auxiliary pumps that supply cooling water to bring down the temperature.

In order to minimize the effect of radiation, the nuclear plant is normally located several kilometers away from any population centers. The fission products that might be released, in case of an accident, will be contained within the steel-reinforced concrete reactor building. The concrete building can withstand high internal pressures and is designed to have a very minimal rate of leak.

Loss of Pressure Faults: If there is a leak in the PWR system, the vessel pressure may drop from nominal 2250 psi to lower values. The safe system in the PWR is activated which pumps water from the borated water storage tank. A large rupture in the coolant system will significantly decrease the vessel pressure and increase the concrete containment pressure. When the vessel pressure drops to 600 psi, water from the core flooding tank will be pumped to the core by a pressurized nitrogen gas stream. If the primary loop pressure drops below 500 psi, then water from the borated water storage tank is pumped to the core by the low-pressure injection pumps (Knief 1992).

Probabilistic Risk Assessment (PRA)

One objective of PRA is to find the chance of an undesired event occurring in a nuclear reactor and analyze its potential causes (Fullwood and Hall 1988). The event can be damage in the core, breach of containment, release of radioactivity, etc. The first step in the risk assessment is to investigate all of the possible failure modes of equipment and/or processes. Event trees are constructed to study the process flow. Fault trees are constructed using the principles of Boolean algebra that trace causes of failures and their effects mathematically. The ultimate objective of the probability risk assessment is to determine risks to people by calculating using the relation

$$ \mathrm{Risk} = \mathrm{frequency} \times \mathrm{consequences}, $$

where frequency is the number of times per year of operation of a reactor that an undesired incident is expected to occur and consequences refers to quantifications of fatalities related directly or indirectly to the event.

Each nuclear power plant and the local government are required to have plans in place for emergency, in case of accidents occurring, and there is a potential for release of radioactivity. Mock accident drills are carried out periodically that resemble actions to be taken in a real accident. Many organizations also involve in such drills to make a coordinated effort to safeguard the life of residents in case of an accident. These organizations include radiation protection staff, police and fire departments, highway patrol, public health officers, and medical response personnel.

If the members of the public suffer loss due to accidents involving reactors or transportation of fuel, they are then compensated by nuclear insurance. The nuclear utilities pay the insurance premium to private insurances to cover any accidents to the public.

Nuclear Accidents

Safety goals are set either by regulatory authorities or by the utilities for safe operation of nuclear power plants.

The probabilistic design targets for LWR nuclear power plants are (IAEA 2001):

• A frequency of occurrence of severe core damage that is below about 10⁻⁴ events per plant operating year for existing nuclear plants

• Achievement of an improved goal of not more than 10⁻⁵ severe core damage events per plant operating year for future reactors

• Practical elimination of accident sequences that could lead to large early radioactive releases

In spite of strict safety regulations, there were accidents in the nuclear power plants. Postaccident analyses of accidents clearly indicated that these accidents were the results of severe violations of specifications and regulatory instructions.

Three Mile Island

A nuclear reactor at the Three Mile Island (TMI) facility failed on March 28, 1979, due to an accident. The accident was considered as the result of a combination of design deficiencies, equipment failure, and operation error. The accident was first triggered by a valve failure in the feedwater system. Since the feedwater system malfunctioned, the turbine generator automatically tripped, and the control rods were driven into the reactor to reduce its power. At this stage, water could have been supplied to the reactor by three backup feedwater pumps. However, there was an operator error as the valve to the steam generator was left closed by mistake. Therefore, the steam generators went dry increasing the temperature and the pressure of the primary water coolant to 2355 psi. This increase in the pressure set the relief valve on. This valve was stuck open due to a mechanical failure. The open valve let the coolant drain off from the primary system that led to a series of malfunctions and the meltdown of the core. The escape of the coolant water also resulted in spillage of some radioactivity. However, it was estimated that the highest possible dose of exposure was only 100 mrems or 1 millisievert, where 1 sievert is equivalent to 1 J/kg of energy release. It should be noted that a person taking a flight from the East Coast to the West Coast in the USA will have a radiation dose of 5 mrems due to cosmic rays.

The Chernobyl Accident

The Chernobyl accident occurred not because of operational abnormality but because of gross violation of operating rules and regulations. People who were in the vicinity of the damaged reactor received a radiation dose of 100 to 1500 rems. More than 135,000 people were evacuated from a 30 km zone.

Design Considerations and Life Prediction of Nuclear Components

In nuclear reactors, passive components such as pressure vessels and piping systems show very low failure probabilities. Therefore, failures of these components only have limited contributions to plant risk. On the other hand, components within a reactor core must tolerate high-temperature water, stress, vibration, and an intense neutron field. Degradation of materials in this environment can adversely affect the performance and in some cases lead to sudden failure. Degradation of materials in a nuclear power plant is very complex. There are over 25 different metal alloys within the primary and secondary systems. In addition, there are additional systems with complex nature such as the concrete containment vessel, instrumentation and control, and other support facilities. The combination of diverse set of materials, complex and harsh environment, and load conditions makes the degradation process very complicated.

The service failures of passive components occur because of the following mechanisms:

• Corrosion fatigue

• Thermal fatigue

• Stress corrosion cracking

• Corrosion attack

• Erosion and cavitation

• Flow-accelerated corrosion – i.e., erosion corrosion

• High-cycle vibration fatigue

• Water hammer

The in-core components are subjected to radiation-assisted failure mechanisms such as radiation-induced segregation, swelling, radiation creep relaxation, radiation-induced hardening, etc. The degradation of materials by radiation is considered first.

Fuel and Fuel Cladding

The fuel cladding helps contain the radioactive fission products. If the integrity of the cladding is maintained intact without cracking, rupturing, or melting during a reactor transient, the radioactive fission products are contained within the fuel rod. Transients in reactor operation or hypothetical accidents could weaken the cladding because of a temperature increase, embrittlement by oxidation, or overstressing by swelling of the fuel because of mismatch in thermal expansion (pellet–clad mechanical interaction) or high fission gas pressure. These events alone or in combination can cause cracking or rupture of the cladding and release of the radioactive products to the coolant. The fuel cladding failure models consider several mechanisms such as cracking due to hydride formation, swelling of the clad because of differential pressure between the fuel rod gap and the coolant, thermal expansion of the fuel and clad, elastic and plastic straining of the clad, interference between adjacent ballooning rods, the effect of grids and azimuthal variations in temperature and straining caused by pellet eccentricity, and stress corrosion cracking induced by pellet–clad interaction in the presence of corrosive fission products such as iodine.

Zirconium-Based Alloys

Zirconium alloys are widely used in nuclear reactors as fuel cladding and other core internals because of their low thermal neutron absorption cross section and high corrosion resistance (Grobe et al. 2009; Choo et al. 1995). Zircaloy-2 (Zr-2) and Zircaloy-4 (Zr-4) are used in low-burnup fuel cladding applications. Modern Zr alloys such as ZIRLO^TM (Zr–1Sn–1Nb–0.1Fe) and M5^TM (Zr–1Nb–0.05Fe–0.015Cr) are candidate materials for high-burnup fuel claddings. During in-reactor irradiation, the Zr alloys undergo a number of changes. Most notable is corrosion by reaction with coolant water and subsequent formation of a thick oxide layer and hydrogen intake in solid solution. A large volume of data is available on the corrosion, creep, high-temperature oxidation, and hydride formation behaviors of Zircaloys that are used in the low-burnup fuel cladding application (Shoesmith 2006). On the other hand, only a limited amount of data is available on the oxidation and degradation behaviors of the modern Zr–Nb alloys intended for high-burnup application.

The available data for Zircaloy-2 and Zircaloy-4 cannot be directly correlated with ZIRLO^TM (Zr–1Sn–1Nb–0.1Fe) and M5^TM (Zr–1Nb–0.05Fe–0.015Cr) because of distinctly different microstructures. The Zircaloy-2 and Zircaloy-4 contain α-solid solution (hcp structure) at the normal operating and dry storage temperatures, whereas the Zr–Nb alloys show a mixture of α(Zr) + β(Nb) phases at the normal operating and dry storage conditions. The presence of the β(Nb) phase significantly alters the oxidation rate of the Zr–1Nb alloys, especially at the phase boundaries (Kiran Kumar et al. 2010). However, a detailed mechanism of the oxidation kinetics at the Nb-rich phase and phase boundary (where depletion of Nb is anticipated) is not available. It should be noted that thermodynamically Nb has a lower affinity to oxygen than Zr. Zr–1Nb alloy (M5) will have lower creep strength than Sn-containing alloys. It was postulated that the mechanical strength would be drastically increased upon irradiation in the reactor (Holt 1974). Furthermore, Zr alloy is capable of dissolving oxygen up to 29 at% during the oxidation process according to the zirconium–oxygen phase diagram reported by Domagala and McPherson (Domagala and McPherson 1954). Inspection of phase diagram of the Zr–O binary indicates high amount of oxygen in the solid solution. The dissolved oxygen in Zr would act as an α-phase stabilizer and also impart creep resistance by solid solution strengthening. Yilmazbahyan et al. (2006) reported the Zr alloy–oxide interface enriched with oxygen up to its solid solution limit (29 at%) for a depth of 160–560 nm inside the metal substrate. Recently Ni et al. (2011) reported the presence of a suboxide layer in addition to an oxygen-rich region at the metal/oxide interface of the ZIRLO during the pretransition steam oxidation stage (prior to breakaway oxidation). The suboxide contained 50 at% oxygen for a thickness of 130 nm, and the metal beneath the suboxide contained 29 at% oxygen for a layer thickness of about 150 nm. Mechanical behavior of the oxygen-rich regions was not reported in detail in any of the literature; however, it is anticipated that these localized solid solution strengthened layers play a significant role in the integrity of the fuel cladding. Furthermore, the presence of a layer with oxygen concentration gradient across the metal/oxide interface greatly improves the adherence of the oxide layer. During temperature spikes, as experienced in the accidental conditions, the oxide layer and the Zr-alloy substrate undergo phase transformations. Volume changes associated with the phase transformations result in residual stresses that cause breakdown of the oxide. The presence of a graded layer would be beneficial since a large strain can be accommodated by this layer. Zr alloys that showed breakaway oxidation essentially had sharp oxygen profiles across the metal/oxide interface without any concentration gradient.

Silicon Carbide

Recent accidents in the Fukushima nuclear reactors of the Tokyo Electric Power Company, Japan, and associated worldwide concerns about the safe operation of nuclear power plants have reignited the interest in developing fuels that can contain fission products under meltdown conditions. Development of a passive LWR design that can withstand loss of coolant condition without replenishing the primary coolant, as occurred in the Fukushima nuclear power plant (I), requires substantial changes in the configuration of the reactor core geometry (Hejzlar et al. 1997). However, such design changes are very expensive and cannot be implemented easily in the existing LWRs. Recently, high-melting reactor core design based on sintered silicon carbide encasement of the UO₂ pellets within the Zircaloy cladding has been proposed (Lippmann et al. 2001). Since the proposed design changes occur only within the standard Zircaloy cladding, application of such modifications in the nuclear reactors that are already in service will be viable and less expensive.

Silicon carbide is a candidate material for one of the four coating layers of tristructural isotropic (TRISO) fuel particles to be used in high-temperature gas reactors. Among various polymorphs of SiC, β-SiC, having a cubic (zinc blende) lattice structure with two interpenetrating fcc lattices, one of carbon and the other of Si displaced by ¼<111>, is considered for various nuclear applications because of its refractoriness, low activation, good mechanical properties, and resistance to radiation damage (Henager et al. 2008). SiC is also considered as a fuel cladding material for Generation III light water reactors (LWRs), specifically by Westinghouse Electric Company. The Westinghouse concept for SiC-based cladding has a monolithic SiC tube wound with SiC fiber tows. The SiC fibers are infiltrated with SiC deposited using chemical vapor infiltration (CVI) (Hallstadius et al. 2012). Silicon carbide is a ceramic material that has a good thermal conductivity (120 Wm⁻¹ K⁻¹), a high sublimation point, and a high bending strength up to 1873 K. It is resistant to oxidation in steam and air up to 1300 °C and stable under neutron irradiation (Bloom 1998; Senor et al. 1996). In order to contain the fission products (especially the gases) under aggressive service conditions, the SiC overlay needs to be gas tight. The manufacture of UO₂ pellet encased with a thin SiC layer is complicated because enough space in between the SiC and UO₂ should be provided in order to accommodate the fission products and minimize the interaction between SiC and UO₂ at high off-normal temperatures.

The SiC-based fuel cladding is considered to be superior to Zr-based cladding because of the following advantages:

• SiC absorbs 25 % less thermal neutrons than the Zr alloy.

• No hydrogen pickup by SiC during normal operation.

• Superior mechanical properties such as SiC retains its mechanical strength even at very high temperatures (1100 °C), failure temperature in excess of 2000 °C, much slower degradation, and no meltdown.

All these factors make this material tolerant to loss of coolant accident (LOCA) conditions. However, stability of SiC in high-temperature water relevant to the LWR conditions should be demonstrated before deploying this material for commercial Generation III water reactor applications.

In high-temperature environments relevant to the very high-temperature gas-cooled reactors, the oxidation of SiC is limited by oxygen concentration and formation of a protective SiO₂ layer by following the reaction (Opila and Hann 1997):

$$ 2\mathrm{SiC} + 3{\mathrm{O}}_2\left(\mathrm{g}\right)\ \to\ 2{\mathrm{SiO}}_2 + 2\mathrm{C}\mathrm{O}\ \left(\mathrm{g}\right) $$

(4)

However, in water vapor–containing environments, in addition to the oxidation reaction (2), a simultaneous volatilization reaction (3) was considered in order to estimate the steady-state oxide layer thickness (Opila 2003):

$$ \mathrm{SiC} + 3{\mathrm{H}}_2\mathrm{O}\ \left(\mathrm{g}\right)\ \to\ {\mathrm{SiO}}_2 + 3{\mathrm{H}}_2\left(\mathrm{g}\right) + \mathrm{C}\mathrm{O}\ \left(\mathrm{g}\right) $$

(5)

$$ {\mathrm{SiO}}_2 + 2{\mathrm{H}}_2\mathrm{O}\ \left(\mathrm{g}\right)\ \to\ \mathrm{S}\mathrm{i}{\left(\mathrm{O}\mathrm{H}\right)}_4\left(\mathrm{g}\right) $$

(6)

The oxidation rate was limited by the transport of water vapor through the SiO₂ layer, and the thickness of the SiO₂ was a function of water vapor pressure. Since water availability is not a limiting factor in the light water reactors, the stability of the protective SiO₂ layer on the SiC surface in high-temperature water should be investigated in detail. Corrosion behavior of SiC in oxygenated and deoxygenated water at 290 °C was reported by Hirayama et al. (1989) in 1989, as a function of pH. The corrosion rate in oxygenated high-pH conditions was observed to obey a linear rate law indicating absence of a protective film on the SiC surface. On the other hand, formation of amorphous, graphitic, and diamond-like carbon films was observed (Kraft et al. 1998; Gogotsi et al. 1996) on the chemical-vapor-deposited (CVD) SiC exposed to water at 100–200 MPa and 400–800 °C. Formation of a protective silica scale was not observed under the supercritical water oxidation conditions. In the oxygenated high-temperature water, the following reactions can occur (Dahlkamp 1993):

$$ \mathrm{SiC} + 2{\mathrm{H}}_2\mathrm{O} + 1/2{\mathrm{O}}_2\left(\mathrm{dissolved}\right)\ \to\ {\mathrm{SiO}}_2 + \mathrm{C}\mathrm{O} + 2{\mathrm{H}}_2 $$

(7)

$$ \mathrm{SiC} + 2{\mathrm{H}}_2\mathrm{O} + {\mathrm{O}}_2\left(\mathrm{dissolved}\right)\ \to\ {\mathrm{SiO}}_2 + {\mathrm{CO}}_2 + 2{\mathrm{H}}_2 $$

(8)

$$ \mathrm{SiC} + 2{\mathrm{H}}_2\mathrm{O} + 2{\mathrm{O}}_2\to\ \mathrm{S}\mathrm{i}{\left(\mathrm{O}\mathrm{H}\right)}_4 + {\mathrm{CO}}_2 $$

(9)

$$ {\mathrm{SiO}}_2 + {\mathrm{H}}_2\mathrm{O}\ \to\ 2{\mathrm{H}}^{+} + {{\mathrm{SiO}}_3}^{2-} $$

(10)

$$ {\mathrm{CO}}_2 + {\mathrm{H}}_2\mathrm{O}\ \to\ 2{\mathrm{H}}^{+} + {{\mathrm{CO}}_3}^{2-} $$

(11)

In the deoxygenated high-temperature water, the proposed surface reactions are

$$ \mathrm{SiC} + 2{\mathrm{H}}_2\mathrm{O}\ \to\ {\mathrm{SiO}}_2 + {\mathrm{CH}}_4 $$

(12)

$$ \mathrm{SiC} + 4{\mathrm{H}}_2\mathrm{O}\ \to\ \mathrm{S}\mathrm{i}{\left(\mathrm{O}\mathrm{H}\right)}_4 + {\mathrm{CH}}_4 $$

(13)

$$ \mathrm{S}\mathrm{i}{\left(\mathrm{O}\mathrm{H}\right)}_4\to\ {\mathrm{H}}_2{{\mathrm{SiO}}_4}^{2-} + 2{\mathrm{H}}^{+} $$

(14)

The above reactions indicate that the corrosion rate of SiC is determined by the kinetics of the formation of a protective SiO₂ layer and its dissolution. Formation of gaseous phases like CH₄, CO, and CO₂ affects the morphology and defect structure of the surface layer that in turn detrimentally affects the protectiveness of the SiO₂ layer. Furthermore, corrosion resistance of the SiC is affected by the physical and chemical properties of the grain boundaries in addition to the purity and microstructure of the material. High-purity CVD SiC showed higher corrosion resistance than sintered SiC in distilled water at 360 °C (Kim et al. 2003). Localized corrosion was reported (Tan et al. 2009) on the SiC samples exposed to supercritical water at 500° C and deoxygenated high-purity water at 300 °C. The localized corrosion in supercritical water was attributed to high-intensity residual strains present in the small grains. Low-angle grain boundaries (Σ3, coincidence site lattices) showed better corrosion resistance than the general boundaries. The pit formation in deoxygenated (and subsequently hydrogenated) water at 300 °C during long exposure times was attributed to soluble silicon hydroxides and loss of silicon. The pits disappeared after reexposure of the samples to the test solution. Therefore, the pitting could be a transition corrosion mechanism of SiC exposed to hydrogenated water for an extended period of time (>4000 h).

It is well established by molecular dynamics simulations and experimental results that at room temperature water (both in gas and liquid phases) adsorbs preferentially on Si sites by a dissociative chemisorption process (Cicero et al. 2004; Liu et al. 2012). The reactivity of Si-terminated SiC surface with water manifests into a corrosion process (by formation of Si–H and Si–OH bonds, reactions 10 and 11). Recently, nanocrystalline 3C–SiC has been used as electrodes for high-efficiency electrochemical hydrogen evolution (He et al. 2012). On the other hand, water dissociation on the C-terminated surface is reported to be energetically unfavorable even at high temperature (Liu et al. 2010). The relaxed H…H distance between H₃ ⁺O and Si–H site is reported to be 0.125 nm, whereas for C–H site the corresponding distance of H…H is 0.275 nm. The larger H₂O–SiC distance of the C-terminated surface and relatively small binding energy (<0.05 eV) result in physic-sorption or no bonding at all (Shen and Pantelides 2013). Therefore, it is possible to improve the corrosion resistance of the SiC by having only the C-terminated surface. However, it should be noted that the structure of carbon is important in determining its activity toward water. For example, if the carbon is present as diamond form, then water adsorption by dissociative chemisorption is a possibility (Okamoto 1998). Furthermore, atomic hydrogen will etch both graphitic and diamond forms of carbon. The reaction of water with Si-terminated surface is associated with the interaction of lone pair of electrons of O atom in the H₂O with the 3-d orbitals of the nearest Si atoms (as electron acceptors). Therefore, the corrosion resistance of SiC could be improved by tying the oxygen with an oxide-forming element that will have more stability in the PWR conditions, for example, Ti. Li et al. (2013) reported that, based on the ab initio calculations, Ti–C bonds are thermodynamically more stable than Ti–Si bonds and the β-SiC (111)/α-Ti(0001) interface shows large work of adhesion. Both titanium silicon carbide and titanium carbide showed better resistance to hydrothermal corrosion than SiC (Zhang et al. 2010; Shimada et al. 2006).

Irradiation Effects

Irradiation-assisted stress corrosion cracking (IASCC) has been drawing more attention over the years and has become potentially a critical phenomenon for core internals in light water reactors (LWRs). Alloys of iron and nickel base and oxygen-free copper are the materials found to be affected by IASCC. It is widely recognized that IASCC is a result of the interaction of irradiation, material, environment, temperature, and stress. The complexity of IASCC arises from the fact that irradiation has an impact on all the other variables listed above so that the knowledge available on SCC of materials in unirradiated environmental conditions is not sufficient to solve the IASCC problem. Because irradiation can alter the microstructure and microchemistry of the material, can affect the aggressiveness of the environment by water radiolysis, can increase the temperature of the parts by gamma heating, and can change the component stresses through relaxation of creep or by radiation hardening, interpretation of wide range of issues influencing IASCC requires specialized knowledge covering fracture mechanics, electrochemistry, physical metallurgy, and core neutronics. IASCC may have a higher potential to occur in fusion reactor components because of the higher dose rate of neutron irradiation than in LWRs. Components of the blanket and first wall cooling system, divertor cooling system, and vacuum vessel cooling system are potential problem sites where IASCC could occur. Though the mechanism of IASCC is not fully understood, factors affecting it are well documented, especially the effect of radiation on the environment and on material properties. Among the radiation effects, some are fluence dependent and some are flux dependent, while both fluence and flux cause joint effects. Radiation-induced segregation (RIS), radiation-induced microstructures, and radiation creep relaxation are fluence dependent, while radiolysis and to some extent RIS are flux dependent. The ratio of thermal to fast neutron flux affects transmutation.

Environment-Assisted Cracking (EAC) of Austenitic Stainless Steel in High-Temperature Water

EAC of stainless alloys in high-temperature water occurs due to the synergistic interaction of stress, environment, and material. Generally, active path corrosion cracking and hydrogen cracking are the mechanisms involved in EAC. Crack initiation and crack propagation are two distinct events, which are controlled by environmental, mechanical, and material variables. Water chemistry and microchemistry of the material play a vital role in initiating the SCC. Dissolved oxygen and CO₂ and the presence of Cl⁻ and SO₄ ²⁻ are deleterious from an environmental point of view, and inclusions such as MnS, segregation of Si and P, sensitized microstructure, and the presence of secondary phases such as sigma, laves, chi, etc., are detrimental from a material point of view.

EAC of Unirradiated Ferritic/Martensitic Steels

Ferritic stainless steels (>17 % Cr) are considered to have better SCC resistance than austenitic stainless steels. This is true only when the Ni, Cu, and Co contents are below certain levels (Bond and Dundar 1977). However, 8–12 % Cr steels are subjected to both SCC and hydrogen embrittlement. Apart from the environmental factors such as dissolved oxygen, presence of sulfate and chloride ions, etc., microstructural condition of the material also controls the cracking behavior. Untempered martensite and acicular bainite phases are found to be more prone to hydrogen cracking than tempered martensite and bainite + ferrite phases (Kerr et al. 1987). Generally it is observed that pitting is associated with the initiation of SCC or corrosion fatigue in this type of material. Mostly intergranular cracking is observed along the prior austenite grain boundaries. However, it is not very clear why only the prior austenite grain boundaries are the most preferred site for cracking and not other boundaries such as interlath boundaries or interfaces between two martensite packets. Probably certain solute elements segregated in the austenite grain boundaries may have more affinity to hydrogen, as discussed by Leslie (1977). But, Auger electron spectroscopy carried out on these fracture surfaces did not throw much light on this aspect. Hydrogen cracking resistance of ferritic/martensitic steel is significant for fusion wall application because direct transmutation, water–lithium interactions, radiolysis of water, and corrosion could charge hydrogen into the steel. Hydrogen cracking could be enhanced by other irradiation damage mechanisms such as RIS, increased defect density, etc.

Environmental Aspect

Radiolysis

Radiolysis is a complex issue affected by water chemistry, neutron flux (not fluence), flow rate, temperature, etc. Radiation causes decomposition of water into many species which affect the corrosion potential. At high hydrogen levels (>1 ppm), radiolysis is sufficiently suppressed so that it has very little effect on changing the corrosion potential (Maziasz and McHargue 1987). The interior of the cracks was not found to be polarized by radiation, as the corrosion potentials of cracks and tight crevices were not altered.

Flux Dependence

The structural materials are exposed to temperatures of 290–350 °C in water reactors. In the case of a BWR, the temperature is constant at 288 °C, whereas in a PWR, the temperature varies with location to a maximum of 400 °C in the baffle plates. The fast flux in a BWR is around 7 × 10¹⁷ n/m² s (E > 1 MeV), and in a PWR, it is 20–30 % higher than in a BWR. Radiation damage in materials is quantified in terms of displacements per atom (dpa) as calculated by approved methods. Empirically 1.4 dpa per 10²⁵ neutrons (n)/m² (E > 1 MeV) is used for LWRs. From this, the fast flux can be back calculated to be 10⁻⁷ dpa/s in the core of LWRs and 1.5–4 × 10⁻⁷ dpa/s in test reactors. In fast reactors, the fast flux is given approximately as 10⁻⁶ dpa/s, and the temperature also is higher (>370 °C) in fast reactors. So, the data generated in fast reactors cannot be compared with those of LWRs. The thermal to fast flux ratio also is an important issue. The thermal neutrons are those which are in thermal equilibrium with neighboring atoms and with energies below 0.5 eV.

Radiation Water Chemistry and Corrosion Potential

Radiation causes breakdown of water into primary species (H⁺, e⁻ _aq) and molecules such as H₂O₂, O₂, H₂, etc. The concentration of species is proportional to the square root of the radiation flux. Fast neutron radiation has a stronger effect on water chemistry than other types of radiation such as thermal neutrons, beta particles, and gamma radiation (Suzuki et al. 1991). This feature is because of the higher linear energy transfer (LET) and the higher neutron flux of fast neutrons.

It is generally believed that the corrosion potential has more influence than the concentration of oxidizing and reducing species in controlling SCC. The initial concentrations of oxygen and hydrogen are found to be important in determining the final corrosion potential after irradiation. Though a large increase in concentration of some species occurs after irradiation, the change in corrosion potential is not drastic. When hydrogen is present at more than 200 ppb and at 0 ppb O₂, there is no radiation-induced elevation of corrosion potential, whereas the presence of H₂O₂ increases the corrosion potential.

Crack Initiation and Propagation

It is generally observed that SCC initiation preferentially occurs at sites like pits and second phase particles. Preferential dissolution of secondary phases or inclusions creates a crevice where the local electrolyte chemistry and local strain level become more favorable for SCC initiation by a slip dissolution mechanism. In the case of IASCC, irradiated microstructural features (like Cr depletion, Si and P segregation, etc.) and the presence of hard phases such as oxides make the crack initiation process much easier. Oxide particles effectively participate in IASCC initiation by two proposed mechanisms as follows: (1) Oxides are hard to deform. So, under load, the shear stress at the interface of the oxide matrix increases to very high levels as the ductile matrix around the particle deforms. This results in failure in the bonding, creating a crevice where the local chemistry of the electrolyte changes to more a conducive condition for promoting SCC. (2) Alternately, the oxide could fracture creating a microcrack which can either extend into the matrix or create a very high stress intensity for easy SCC initiation.

Strain at crack initiation (SCI) was proposed as the definition for IASCC initiation in slow strain rate tensile testing (SSRT) at 10⁻⁷ s⁻¹ strain rate. It was defined as the strain at which the stress–strain curve of SSRTs began to depart from that of tensile tests, when plotted using the same coordinates. Higher SCI means SCC initiation starts at higher strain. Though the intergranular (IG) fracture ratio decreases with decreasing dissolved oxygen (DO), it increases inversely below 10 ppb of DO. These phenomena may indicate the continuum of initiation of IASCC from BWR conditions to PWR conditions.

Crack Propagation : Gamma ray irradiation is not expected to affect the microstructure or microchemistry of the material. However, it decomposes water into many kinds of radiolytic products of which hydrogen peroxide (H₂O₂) is very important to IASCC. In the 288 °C BWR environment, gamma irradiation accelerated the crack growth to varying degrees depending on the water chemistry, flux, etc. For example, the average crack growth rates in unirradiated, irradiated with gamma ray for fluxes of 5 × 10⁶, and 9 × 10⁶ R/h were 7.2 × 10⁻¹⁰, 1 × 10⁻⁹, and 1.3 × 10⁻⁹ m/s, respectively. From these values, the crack growth rates in low-conductivity pure water could be observed to be marginally affected by gamma ray irradiation.

The effect of dissolved oxygen (DO) on crack velocity with additions of Na₂SO₄ is similar in both irradiated and unirradiated test conditions. Addition of sulfate ions showed more effect in accelerating the crack growth than did irradiation. DO also had a similar effect. Suppressing the DO content decreased the crack growth rate. Though crack velocity increased with sulfate ions as in the case of the unirradiated condition, DO had a major effect in controlling the crack behavior in the irradiated condition also. Nitrate additions were found to be less aggressive than sulfate additions in a BWR environment for 304 SS. Dissolved hydrogen showed greater beneficial effect in suppressing crack growth. The mechanism of crack growth mitigation by hydrogen injection could be explained by analyzing the corrosion potential of the system. The presence of molecules like H₂O₂ and O₂ increases the free corrosion potential which falls into the cracking range, and hence, the crack velocity is enhanced following the slip dissolution model and Faraday’s law. Whereas when hydrogen is introduced into the environment, it helps the recombination of species and thus reduces the corrosion potential well below the cracking range. IASCC tests were carried out on irradiated stainless steel samples under BWR condition using the slow strain rate testing method. They presented average crack growth data by dividing the maximum crack depth by total test duration. The maximum crack growth rate divided by the test time was suppressed by hydrogen water chemistry (HWC) below 3 × 10²¹ neutrons (n)/cm², but not above 3 × 10²¹ n/cm². It was observed that variations in either fluence level (3 × 10²⁰–9 × 10²¹ n/cm²; E > 1 MeV) or flux level (1.5 × 10¹³–7.6 × 10¹³ n/cm² s) did not affect the crack velocity drastically (a maximum of a factor of two).

Critical Issues on Selection of Candidate Materials for Advanced Nuclear Reactors

Advanced systems selected for Generation IV reactors require high operating temperatures in the range of 500–1000 °C, depending on the coolant and longer service life. The fuels of the advanced reactors will have very high-burnup capabilities and fast neutron spectra. The construction materials of Generation IV reactors will be exposed to severe environmental conditions in combination with increased radiation damage. Therefore, selection of structural materials for advanced reactors requires a thorough understanding of materials’ behavior in the extreme service conditions.

The structural materials of advanced nuclear reactors will undergo degradation primarily due to three factors, viz., (1) exposure to high temperature and service stresses (high-temperature degradation), (2) irradiation damage, and (3) interaction with service environments. The first two factors are common among all the types of reactors, and therefore, the data generated at high temperatures and irradiation levels relevant to the service conditions can be used for material qualification for different type of reactors, as the operating temperature of most of the advanced reactors is in the range of 500–800 °C. However, the third factor, interaction with environment, is reactor specific. The material should possess higher resistance to corrosion attack in the service environment. Among the various types of advanced reactors, liquid metal (particularly liquid sodium and lead–bismuth eutectic)-cooled fast reactors are considered in this study. Some of the critical issues pertaining to each major degradation modes will be discussed in this presentation.

The materials considered for advanced reactor structural applications can be classified into three major categories viz., (1) ferritic–martensitic-type Fe–Cr alloys, (2) austenitic alloys (stainless steels and Ni–Cr–Mo alloys), and (3) oxide dispersion strengthened (ODS) alloys. Refractory metal-based alloys are not considered in this work. Merits and disadvantages of the first two categories of the materials will be analyzed based on the critical degradation issues.

High-Temperature Degradation

Major Issues and Temperature Limits

The major issues of high-temperature degradation are phase stability, oxidation, and creep–fatigue interaction. It is widely believed that thermal effects will offset the irradiation effects at high temperatures because of increased diffusivities and stress relaxation effects. This may be true for annihilation of point defects. However, effect of radiation-induced segregation could be aggravated at high temperatures.

Available literature data indicate that the maximum service temperatures of different alloys are limited by chemistry and microstructure. For example, ferritic/martensitic steel with a maximum Cr content of 12 % can service up to 650 °C and austenitic stainless steels up to 800 °C, nickel-based alloys up to 900 °C, and ODS alloys up to 1050 °C. The interaction of creep–fatigue is considered to be of primary importance.

Fatigue, Creep, and Creep–Fatigue Interaction

Creep or creep–fatigue interaction of structural materials at elevated temperatures over a long period of time in advanced reactor environments is a critical issue. High temperature and the temperature gradient during start-ups, in-services, and shutdowns induce both static and cyclic thermal stresses. These constitute the stress factors that generate creep and creep–fatigue interaction. In addition, components such as thread roots in steam turbine casing bolts, pipe, and branch connections in reactors endure multiaxial stresses.

The earlier studies (Brinkman and Korth 1973) investigated the effect of heat-to-heat variation on fatigue and creep–fatigue resistance of type 304 stainless steel at 593 °C. Carbide precipitation was considered as the reason of increasing low-cycle fatigue (LCF) resistance. Additionally, a fairly uniform distribution of inter- and intragranular carbides M₂₃C₆ was considered to increase the resistance to the tensile hold time effect. Generally, zero hold time tests revealed transgranular fracture surfaces, while intergranular features were obtained even with hold times as short as 0.01 h. This is also illustrated by the studies of Schaaf (1988) (Fig. 5). The creep–fatigue failure can be categorized into three modes: fatigue-dominated failure with almost transgranular features, creep–fatigue interaction (both transgranular and intergranular), and creep-dominated failure with mainly intergranular cracks.

Fig. 5

The three failure modes: fatigue dominated (left), creep–fatigue interaction (left), creep dominated (middle) (Van Der Schaaf 1988)

In recent years, creep–fatigue properties of liquid metal fast breeder reactor (LMFBR) candidate structural materials, such as austenitic 304 L, 304NG, 316LN, and AISI 321, were investigated at 600 °C (Rho and Nam 2002; Nilsson 1988; Min and Nam 2003). It was observed that nitrogen addition improved fatigue life under creep–fatigue condition. The density of Cr-rich carbides formed at the grain boundary of 304NG (0.08 % N) was lower than that of 304 L (0.03 % N). Planar slip planes of 316LN initiated under creep–fatigue interaction probably enhanced stress concentration immediately next to grain boundaries and promoted intergranular fatigue fracture. In the case of AISI 321, it was observed that the creep–fatigue life of TiC-aged specimen was 40 % longer than that of Cr₂₃C₆ aged, although the two carbide densities at grain boundaries were similar. It is suggested that the interfacial free energy between TiC and grains is lower than that between Cr₂₃C₆ and grains in AISI 321.

In addition, irradiation creep accumulates in reactor materials. It is known that irradiation creep has very weak temperature dependence. However, creep remains high at temperatures as low as 60 °C (Grossbeck et al. 1990). It is postulated that migration of vacancies and migration of interstitials are two independent mechanisms of irradiation creep. The effect of irradiation is to lower the endurance of plastic strain range.

So far, most of the experimental studies on creep–fatigue interaction were conducted by using low-cycle fatigue tests with and without tensile strain hold in air at temperatures ranging from 400 to 600 °C. The accumulated data in simulated reactor environments at high temperature up to 800 °C is inadequate for a better understanding of the creep–fatigue interaction mechanism. For example, oxidation and solubility of alloying elements in high-temperature liquid metal have to be considered as possible factors affecting creep–fatigue behavior. Also, carbide precipitation at component weld joints and heat-affected zone (HAZ) may have different behaviors from base metals.

Creep–Fatigue Life Prediction

In this section, selected creep–fatigue life prediction methods are reviewed without considering the irradiation effects.

Suauzay et al. (2004) analyzed their experimental results of creep–fatigue behavior of 316LN at 500 °C using linear damage accumulation model. This model is based on Miner’s rule, expressed as

$$ \frac{N_F}{N_F^{pf}}+\frac{t_F}{t_F^{relax}}=1 $$

(15)

• N _F : number of cycles to failure for a t_h hold time (t_n > 0)

• N _F ^Pf: number of cycles to failure in pure fatigue, based on the Coffin–Manson relation (t _n = 0)

• t _F  = N_Ft_h

• t _F ^creep: failure time in pure creep condition given as

• t _F ^creep = H / σ^r, where H and r are creep coefficients

$$ {t}_F^{relax}={N}_F{\displaystyle {\int}_0^{t_h}\frac{dt}{t_F^{\mathrm{creep}}\sigma (t)}} $$

(16)

Tsuji and Nakajima (1994) evaluated the damage accumulation of Hastelloy-XR in HTGR environment at 700–950 °C by applying the life fraction rule and ductility exhaustion rule. The creep damage during strain holding time was given as

$$ {D}_{cL}={{\displaystyle \sum_i\left(\Delta {t}_i/\Delta {t}_{Ri}\right)}}^n $$

(17)

• D _CL : creep damage by life fraction rule

• Δt _i : strain holding period for particular temperature and stress

• t _Ri : rupture time based on the Larson–Miller parameter

• n: number cycles for failure in the experimental condition with trapezoidal strain wave form (fatigue–creep components)

The ductility-exhaustion rule is given as

$$ {D}_{cd}={{\displaystyle \sum_i\left({\dot{\varepsilon}}_{\min}\Delta {t}_i/{\varepsilon}_{Ri}\right)}}^n $$

(18)

• D _cd : creep damage by ductility exhaustion rule

• Δt _i : strain holding period for particular temperature and stress

• $\dot{\varepsilon}$ _min: minimum creep rate calculated from the Larson–Miller parameter

• ε _Ri : strain at rupture

It was observed that the ductility exhaustion rule predicted the fatigue life under the effective creep condition more successfully than the life fraction rule.

Most of the creep–fatigue life prediction models are based on phenomenology of failures. For example, ferritic/martensitic steels and nickel-based superalloys showed damage accumulation at the crack tip or crack process zones. In these materials even compressive stress hold times were found to affect the damage accumulation. In case of austenitic stainless steels, creep–fatigue damage occurs by grain boundary cavitation, and tensile hold time is considered to be more important. The proposed damage accumulation function based on grain boundary cavitation phenomenon is given as (Nam 2002)

$$ {D}_{C-F}=\Delta {\varepsilon}_p^m{\left\{\frac{ \exp \left(-{Q}_g/RT\right)}{T}{\displaystyle {\int}_0^t\sigma (t)dt}\right\}}^{2/3} $$

(19)

• Δε _p : plastic strain range

• m: strain exponent

• Q _g : activation energy for grain boundary diffusion

• R: gas constant

• T: temperature

When the damage function is plotted against experimental creep–fatigue life, as shown in Fig. 6, a linear relationship was observed with a slope of −1.66. Consideration of creep–fatigue life of different types of austenitic stainless steels at 600–700 °C revealed that the slope varies from −1.62 to −1.66. Therefore, the proposed damage function can be used for life prediction in the given experimental conditions.

Fig. 6

Normalized Coffin–Manson plot for AISI 316 stainless steel (Nam 2002)

It should be noted that most of the life prediction methodologies are based on the data generated using smooth tensile specimens. In these cases, the events of crack initiation and crack propagation were not distinguished. In service conditions, inspection and monitoring methods require data on crack initiation time and crack velocity to evaluate the integrity of the components.

Yokobori (Yokobori and Yokobori 2001) considered the crack initiation and crack propagation issues for creep–fatigue interaction based on critical notch opening displacement criterion. According to these authors, the crack initiation event is complete when the defect size reaches 5 μm, a dimension that can be resolved by optical microscopy. Crack initiation is considered to occur when a critical strain is reached by some atomistic rearrangement by time-dependent plastic flow. This process in creep–fatigue interaction is thermally activated and aided by stress field. Based on these considerations, the time for crack initiation was given as

$$ 1/{t}_i=\frac{A_1}{\varepsilon_c} \exp \left(\frac{-\Delta H-\Phi \left({\sigma}_g\right)}{RT}\right) $$

(20)

• t _i : time for crack initiation

• ε _c : critical strain required for crack initiation

• A ₁: material constant

• ΔH: activation energy for plastic flow

$$ \Phi \left({\sigma}_g\right)=\Delta f \ln \left(\frac{K_I}{G\sqrt{b}}\right) $$

(21)

• Δf: energy coefficient

• K _I : stress intensity factor = α a^1/2 σ _g

• G : modulus of rigidity

• b: burgers vector

Similar expressions can also be written for crack propagation. The expression for crack propagation is

$$ \frac{da}{dt}=B{\sigma}_g^m{K}_I^n\left[\frac{-\left\{\Delta {H}_g-\Delta f \ln \left(\frac{K_I}{G\sqrt{b}}\right)\right\}}{RT}\right] $$

(22)

The major advantage of the above approach is its ability to take various boundary conditions, such as environmental degradation (reduction in ΔH_g due to liquid metal) and irradiation effects, into account.

Radiation Damage

Radiation Damage of Microstructure

Among the radiation effects, some are fluence dependent and some are flux dependent, while both fluence and flux cause joint effects. Radiation-induced segregation (RIS), radiation-induced microstructures, and radiation creep relaxation are fluence dependent, while RIS is flux dependent to some extent.

Figure 7 illustrates schematically the collision of an energetic particle (either a neutron, an electron, or a proton) with a lattice atom generating radiation damage (Maziasz 1993). If the energy transfer of the elastic collision is greater than the displacement threshold (E_d), a primary knock-on atom (PKA) is generated. PKA can displace additional atoms through the lattice if it has sufficient energy until the energy of all the atoms has been reduced below E_d (Capdevila et al. 2008). The Frenkel pair, consisting of a vacancy and self-interstitial atom (SIA), could be considered as the fundamental component of radiation damage (Fullwood and Hall 1988).

Fig. 7

Schematic illustration of generation of a primary knock-on atom (PKA) (Maziasz 1993)

The extent of radiation damage is a function of temperature. Several extensive reviews on microstructural evolution in irradiated austenitic stainless steels are available, which report a transition of microstructural damage approximately at 300 °C. At high temperatures (>300 °C), vacancy clusters in austenitic stainless steels become thermally unstable. The presence of voids and swelling is observed at higher temperatures. Under certain conditions small gas-filled bubbles can grow to form voids, referred to as swelling, as the volume of material increases beyond the size limitation dictated by the thermodynamic equilibrium of gas. Both hydrogen and helium play an important role in swelling of a material. A swelling rate of 1 % per dpa is maintained at temperatures above 425 °C. The lower limit of temperature for swelling is observed to be affected by displacement rate.

Radiation-Induced Microchemistry

In austenitic stainless steels, depletion of Cr and Fe and enrichment of Ni have been observed. The Cr and Fe have higher diffusivity than Ni. Therefore, they migrate away from the interface, enriching the boundary with Ni. This could be attributed to the inverse Kirkendall segregation. Segregation of Si and P at grain boundaries is observed by an uphill diffusion process. Along with Cr and Fe, minor alloying elements such as Mn, Ti, and Mo also get depleted at grain boundaries. Mn levels drop to 0.5 at% at grain boundaries in type 304 SS. In type 316 SS, more than 50 % depletion of Mo after irradiation to 3 dpa has been reported (Cookson and Was 1995). For the same level of irradiation, enrichment of Si occurred to levels of about 6–8 at%. Nickel-silicide precipitation also has often been reported to form at dislocation loops at temperatures >380 °C and at higher doses (>20 dpa) (Kimura et al. 1996). At higher doses (PWR relevant, >10 dpa) sulfur segregation can be expected due to the burnup of Mn in MnS inclusions and subsequent release of S. Radiation-induced Cr depletion could retard carbide formation at grain boundaries. Radiation-induced segregation of Ni and Si could lead to formation of γ′ or G phase at higher temperatures (Shiba et al. 1996).

Mechanical Properties

In general, it is observed that with increases in irradiation dose, the yield strength of the material increases. The ultimate tensile strength also increases, but the increase is not as great as for the yield strength. Formation of higher densities of vacancies and interstitials is attributed as the cause for this increase. Suzuki et al. (Holt 1974) reported increases in strength for various grades of austenitic stainless steels with increase in neutron fluence as shown in Fig. 8. However, a saturation level is reached at the 3 × 10²⁵ n/m² fluence level (E > 1 MeV) beyond which no significant increase in strength could be observed. The increase in yield strength (Δσ) of the 304 SS irradiated in BWR environment at 288 °C showed a relation of Δσ = 1.1 × 10⁻³ × (neutron fluence, n/m²)^0.27. It was observed that type 304 SS was more prone to irradiation hardening than was type 316. Composition has two effects, viz., (1) certain alloy elements help nucleate Frank loops and (2) stacking fault energy (SFE) is altered. Low SFE results in more hardening. Also a low SFE can lead to nucleation of twins as an alternative deformation mechanism to dislocation glide. Alloying elements such as Ni, Mo, and C increase the SFE in austenitic stainless steel, and Cr, Si, Mn, and N tend to decrease the SFE.

Fig. 8

Typical relation between the increase of the 0.2 % yield stress of austenitic stainless steels and neutron fluence (E > 1 MeV) after irradiation in BWR environment at 288 °C (Koyama et al. 2007)

Loss of work hardening and uniform elongation is observed after irradiation. The elongation decreases significantly with increasing dose. This kind of loss in work hardening and hence uniform ductility could be attributed to the irradiated microstructure, where annihilation of barriers occurs due to their interaction with dislocation. Interacting with obstacles, dislocations multiply in unirradiated material which results in development of back stresses and hence work hardening of the material. However, in irradiated conditions, the obstacles such as loops and voids can be destroyed when they interact with moving dislocations, resulting in work softening. This behavior causes flow localization, and hence, the slip band spacing increases, ultimately reducing the macroscopic deformation. At higher temperatures (above 600 °C), the ductility is observed to be severely affected by He embrittlement. When a large void population develops near 400 °C, the fracture mode is observed to be transgranular channel. The reduction of fracture toughness of irradiated SS can be attributed to the higher population of voids so that fracture occurs at an early stage by dislocation channeling or highly heterogeneous deformation–decohesion ahead of the crack tip²³. RIS of Ni at voids also results in brittle behavior of a material. This preferential segregation of Ni at voids results in matrix depletion of Ni and hence destabilizes the austenite. The strain-induced martensite transformation, possible in the destabilized austenite, acts as low-energy path for crack propagation ²⁴. This mechanism for cracking resulted in quasi-cleavage fracture with an overall fracture toughness of 80 MPa m^1/2 after the austenitic material has been irradiated to high dose (1.6 × 10²³ n/cm²) at 425 °C.

Irradiation hardening and softening are important factors in determining the fusion reactor life limits as creep properties are affected by these changes. In ferritic steels, the irradiation hardening is attributed to the formation of small defect clusters and dislocation loops, with associated precipitation of small carbides such as M₂C, M₆C, etc. Kimura et al. (1996) studied the irradiation hardening behavior of 9Cr-2 WV steel and reported saturation of irradiation hardening at a dose level of about 10–15 dpa. Irradiating at above 430 °C resulted in softening at dose levels of 40–60 dpa. Swelling was found to be associated only with hardening, in this study.

Shiba et al. (1996) investigated the response of F82H steel to irradiation at low damage levels (<1 dpa) for 300–500 °C. They observed hardening only at 300 °C. However, no softening was observed when irradiated at 520 °C. Interestingly for these test conditions, no change in DBTT was observed between irradiated and unirradiated Charpy impact tested samples. However, the upper shelf energies were lower for irradiated samples.

Klueh and Alexander (1996) conducted a detailed study on Charpy impact toughness of different types of low-activation steels irradiated at 0–24 dpa. They observed that 9Cr-2 W-type steels were least affected by irradiation. In terms of microstructure, they found that steels with 100 % martensitic structure showed superior impact toughness properties after irradiation than steels with dual microstructures such as bainite + ferrite or martensite + ferrite. By changing the composition and microstructure, the effect of irradiation on toughness could be favorably modified.

Typical 12 % Cr steel shows very little void swelling (only 0.1 % volume change for a dose of 90 dpa at 400 °C). It is generally observed that 9 % Cr steel shows better impact toughness and DBTT values after irradiation than 12 % Cr steel. However, by controlling the phase content (either 100 % martensite or at least less than 20 % delta ferrite) and with uniform distribution of carbide/carbonitride phases, much improved mechanical properties could be achieved. In the case of 9 % Cr steel, the volumetric swelling was around 0.1 % at 100 dpa and 400 °C which is similar to that for 12 % Cr steel. After 90–100 dpa, the rate of swelling of 9 % Cr steel was reported to be approximately 0.01 %/dpa.

Irradiation Creep

Two groups of mechanisms have been proposed for irradiation creep, viz., (1) irradiation-induced creep and (2) irradiation-enhanced creep.

Irradiation-Induced Creep. Depending on the nucleation and growth of dislocation loops, two types of mechanisms are observed, viz.:

• 1. Stress-induced preferred nucleation (SIPN)

• 2. Stress-induced preferential absorption (SIPA)

For the case of SIPN, interstitial loops are assumed to nucleate preferentially on planes perpendicular to the tensile stress, and vacancy loops nucleate preferentially on the planes parallel to the applied stress.

For the SIPA mechanism, interstitials are preferentially absorbed by the loops oriented perpendicular to the tensile stress so that there is an elongation in the direction of stress.

Irradiation-Enhanced Creep. It is postulated that irradiation accelerates the thermal creep by producing excess vacancies and interstitials and thus facilitating the easier dislocation movements. For example, jogs in the form of nonglissile edge dislocation segments on a screw dislocation can be moved by the point defects produced by irradiation, otherwise not possible.

Fretting

Fretting results from low-amplitude motion between the tube and the support plates of the steam generators in the nuclear power plants. These flow-induced, low-amplitude vibrations occur during the normal operation of steam generator, in spite of the use of anti-vibration supports, wherein two fluids are flowing in countercurrent manner. Several parameters including fluid flow, vibration frequency, and impact force affect fretting. The flow-induced vibrations also lead to cyclic loading that will be random in nature. As per the US Nuclear Regulatory Commission, the USA has 27 nuclear power plants that use Inconel Alloy 600 (Ni, >72 %; Cr, 14–17 %; and Fe, 6–10 % form major constituents) and 42 that use Alloy 690 (Ni, >58 %; Cr, 27–31 %; and Fe, 7–11 % form major constituents) as tubing material. To improve mechanochemical properties, these materials are subjected to mill annealing (Alloy 600) or thermal treatment (Alloy 600 or 690), which forms the important factor, other than the alloy composition, in determining its degradation. The tube support plates are typically fabricated using 405 ferritic stainless steels. While the primary reason for degradation and failure of the tubes used to be thinning of tubing material due to water flow, the recent failures and inspections indicate that accelerated degradation is becoming an issue of concern. At the center of this is the failure of steam generator tubes in January 2012, after less than 3 years of operation, at the San Onofre plant in California which led to the leakage of radioactive material from inside the tubes to the outside water. While the migration from Alloy 600 to 690 was primarily conducted due to improved corrosion resistance of Alloy 690 (provided by higher chromium content), the mechanical properties of Alloy 690 are not superior to that of Alloy 600. Therefore, Alloy 690 would be expected to be more susceptible to mechanically induced failure such as fretting and fatigue. Moreover, since the steam generator transfers excess heat from reactor core to outside, these tubes are exposed to extreme temperatures and (~320 °C) and pressures (~150 bar). Preliminary reports from the San Onofre nuclear plant indicated that the accelerated degradation was in part due to increased fretting from flow-induced vibration. This type of cyclic loading, in addition to the normal load (contact stress) due to fretting conditions, results in damage accumulation beneath the contacting surface of Alloy 690. The mechanism of fretting in LWR environments is complex because the failure occurs due to a combination of several synergistic processes such as fretting fatigue, fretting corrosion, and fretting wear. The material removal occurs in the following stages: (1) formation of highly plastic deformed surface layer, (2) fracture of the work-hardened layer, and (3) removal of wear debris and propagation of cracks in the deformed subsurface. The localized material loss due to fretting has two consequences in the LWR environment, namely, (1) accelerated corrosion of small worn-out areas that become anodes and large unaffected areas that act as cathodes and (2) fatigue crack initiation from the worn-out area that acts as a stress concentrator. Another important aspect is the microstructure of the alloy. Greater resistance to wear was observed with the large grain structures and coarse carbides along the grain boundaries of nickel-based alloys. Carbide morphology also influenced the wear resistance. Continuous grain boundary carbides showed increased propensity to crack formation (and hence low wear resistance) as compared to discrete grain boundary carbides.

Cast Stainless Steel Components

Cast stainless steels are extensively used in light water reactors (LWRs) as alloys for coolant piping and auxiliary piping components such as pump casings, valve bodies and fittings, elbows, and nozzles. Similar to the weld microstructure of austenitic stainless steels, the cast microstructure also contains delta ferrite. The ferrite content varies from 3 % to 12 % in welds and up to 40 % in cast austenitic stainless steel components. The delta ferrite is required to mitigate hot cracking during solidification and control the intergranular corrosion. Mechanical strength and stress corrosion cracking resistance are improved by the ferrite phase present in the austenite matrix. Depending on the chemical composition, the primary solidification phase could be austenite or ferrite. When the primary solidification phase is austenite, the ferrite is present as interdendrites. Partitioning of the solute elements occur in the interdendritic regions that affect the chemical and mechanical properties when compared to the equiaxed wrought microstructures. The heterogeneity in the chemical composition also results in detrimental microstructural changes such as spinodal decomposition and precipitation of topologically close packed (TCP) phases during long time exposures to service temperatures that lead to thermal embrittlement.

The popular grades of cast austenite + ferrite duplex structure stainless steels in nuclear service are the CF3 and CF8 series of alloys. Among these, the CF3, CF3A, CF3M, CF8, CF8A, and CF8M are the most widely used alloys (equivalents of 304 and 316 wrought grades). These alloys typically have 17–21 wt% Cr and 8–13 wt% Ni. The digit following the letters CF refers to the carbon content of the alloys “3” for 0.03 % and “8” for 0.08 %. The fourth letter “A” denotes higher ferrite control which raises strength above that of the normal CF grades, and the letter “M” denotes addition of Mo to the nominal compositions of CF grade alloys. The macroscopic cast structure is generally divided into two categories depending on the casting process, namely, (1) static cast structure which contains columnar grain structure at the ends and equiaxed (randomly speckled) grains at the center (Calonne et al. 2004) and (2) centrifugally cast structure which contains long columnar grains at the outer wall and a mixture of equiaxed and columnar structures in the inner regions (Anderson et al. 2007).

Embrittlement due to thermal aging of cast stainless steels at service conditions in the temperature range of 280–320 °C has been a major concern (Chung and Leax 1990). The main transformations are the spinodal decomposition of α into α and a chromium-rich phase α′, precipitation of a G phase (Ni₁₆Ti₆Si₇), ε, and π (a nitride phase). Primarily, the formation of Cr-rich α′ (martensite) phase strengthens ferrite and decreases the toughness. With increased temperature (>550 °C), other embrittling phases such as σ, χ, η, M₂₃C₆ carbide, and γ₂ austenite are form aided by the presence of the ferrite/austenite interfaces. Sigma phase is a tetragonal crystal composed of (Cr,Mo)_x (Ni,Fe)_y. The chi (χ) phase is a body-centered cubic with a typical composition of Fe₃₆Cr₁₂Mo₁₀. The typical stoichiometry of the Laves (η) phase is Fe₂Mo with a hexagonal structure. These topologically close packed (TCP) phases have large lattice parameters and large number of atoms in a lattice that show directional properties. Since these TCP phases nucleate at the high surface energy sites (grain boundaries and phase boundaries), cohesive strength of the grains is significantly reduced and brittle failure is often observed. It is important to note that the cold working accelerates the formation of the TCP phases by increased diffusion. Therefore, formation of Laves phase in cold-worked structure can be a high possibility even at reactor service temperatures.

Corrosion fatigue data for cast stainless steels in water containing 200 ppb and 8 ppm of dissolved oxygen (DO) at 289 °C have been generated and compiled by Shack and Kassner of the Argonne National Laboratory (Shack and Kassner). In general, the corrosion fatigue crack growth rate is assumed to be related to the air fatigue crack growth through a power law given as

$$ {\left(\mathrm{d}\mathrm{a}/\mathrm{d}\mathrm{t}\right)}_{\mathrm{env}} = \mathrm{A}\ {{\left(\mathrm{d}\mathrm{a}/\mathrm{d}\mathrm{t}\right)}^{\mathrm{m}}}_{\mathrm{air}} $$

(23)

For stress ratio R < 0.9, A = 4.5 × 10⁻⁵ for DO = 200 ppb, and A = 1.5 × 10⁻⁴ for DO = 8 ppm, m = 0.5.

Kawaguchi et al. (1997) studied the thermal embrittlement behavior of centrifugally cast CF8M duplex stainless steel after aging at 300–450 °C for up to 40,000 h. The aging treatment was quantified by a temper parameter denoted as P and given as P = log(t) + 0.4343(Q/R)(T ₁ ⁻¹− T ₂ ⁻¹), where t = aging time, Q = activation energy for the embrittlement (typically 100 kJ/mol), T = temperature, and R = gas constant. The ferrite content of the samples varied from 15 % to 17.5 %. Spinodal decomposition of δ-ferrite to Cr-rich α′ phase (size, 5 nm) was observed after the following aging conditions: 300 °C for 10⁴ h, 350 °C for 3000 h, and 450 °C for 300 h. The precipitation of larger (~50 nm) G phase was observed only at longer aging times than that required for spinodal decomposition and at higher temperatures. For example, aging at 300 °C for 40,000 h did not show the presence of G phase. Thermal aging at 350 °C for 10⁴ h and 450 °C for 3000 h showed occurrence of the G phase. Spinodal decomposition was considered the main reason for the thermal embrittlement behavior of the CF8M cast stainless steel based on the Charpy V-notch energy of 230 J that decreased from the 300 J of the as-cast samples.

The use of subsize CT samples for the evaluation of the fracture toughness and the validation of the results with 1 T-CT samples was investigated by Jayet-Gendrot et al. (1998). Mini-CT specimens (5 mm thick) were extracted from the skin of the cast stainless steel elbows of a PWR unit that underwent 86,898 h of service at around 323 °C. The J-integral values of the mini-CT specimens (82 kJ/m² at 0.2 mm of Δa offset) were observed to be in good agreement with those derived from the 1 T-CT specimens. The effect of thermal aging on the low-cycle fatigue (LCF) behavior of the cast stainless steel in room temperature air was evaluated by Kwon et al. (2001). The samples were evaluated in as-cast and aged conditions (430 °C for 300 and 1800 h), and the LCF behavior was described by the relation

$$ \frac{\Delta {\varepsilon}_t}{2}=\left[\frac{\sigma_f^{\prime }}{E}\right]{N}_f^b+{\varepsilon}_f^{\prime }{N}_f^c $$

(24)

where Δε _t  = total strain range, σ _f ′ = fatigue strength coefficient, E = Young’s modulus, b = Basquin’s exponent, ε _f ′ = fatigue ductility coefficient, c = fatigue ductility exponent, and N _f  = cycles to failure. The values of (σ _f ′/E), (−b), (ε _f ′), and (−c) of the 300 h aged samples were higher than that of un-aged samples. However, increasing the aging time to 1800 h resulted in lower values than that of the un-aged samples.

Jeong et al. (2009) evaluated the effect of strain hardening on the environmental fatigue behavior of CF8M under PWR conditions. The material was taken in the as-cast condition with 25 % ferrite. The tests were carried out at 316 °C and 15 MPa with 30 ml of dissolved hydrogen per kg of H₂O and < 5 ppb of DO. Cyclic hardening was observed during the initial 200 cycles that showed peak loads which increased with increase in the strain amplitude. The fatigue test data points were scattered within the ASME design curve and the ASME mean curve. The same group also evaluated the effect of strain rate on the fatigue behavior (Jeong et al. 2011). The strain rate was varied from 0.004 % s⁻¹ to 0.04 % s⁻¹. The number of cycles to failure increased with increase in the strain rate almost by an order of magnitude. The increase in the strain amplitude from 0.4 % to 0.8 % decreased the number of cycles to failure (from 2750 to 150 cycles at 0.004 % s⁻¹ and from 13,500 cycles to 1500 cycles at 0.04 % s⁻¹ strain rate).

Cicero et al. (2009) analyzed a CASS CF8M component (motor-operated valve of the reactor water cleanup (RWCU) system of a BWR unit) that was in service for 40 years using FITNET-FFS procedure and the ASME code. The ferrite content of the component was about 15 %. If the ferrite content was more than 10 %, the aging effect due to service temperature was needed to be considered for structural integrity analysis. The RWCU system was subjected to more than 60 major thermal cycles in the temperature range of 30–300 °C and a stable operating temperature of ~250 °C in 14 years. There were other minor temperature excursions at around 250 °C. The maximum service stress calculated at the neck of the valve was about 86 MPa, and the critical flaw size was much larger than that could be detected by inspection techniques. Wang et al. (2010) used nano-indentation technique to evaluate the thermal aging damage mechanism of the CASS. The specimens were aged at 400 °C for 100–3000 h representing service life of 0.7–21.48 years according to the corresponding Arrhenius relation. Dislocation pileup at the Cr-rich clusters of α′ spinodal decomposed phase was attributed to the observed embrittlement.

Cost–Benefit Analysis

Nuclear power is highly competitive with other forms of power generation such as fossil fuel power and renewable energy-based power generation. The cost of fuel is much less than that of fossil fuels. However, the capital cost is high because of increased margin of safety precautions and cost involved toward storage of spent fuels. While calculating the cost of nuclear power, the cost involved in waste management and decommissioning cost are fully considered ( Economics of Nuclear Power).

In 2010, the cost of 1 kg of uranium as UO₂ reactor fuel is calculated as $ 2555. At 45,000 MW-day/ton burnup, 360,000 kWh electrical energy can be generated per kg of fuel. Therefore, the fuel cost per kWh energy is 0.77 cent. The US electricity production cost using different fuel sources in the year 2008 is given in Table 1. This includes cost of fuel, operation, and maintenance. Capital cost is not considered.

Table 1 Cost comparison of electricity generation in the USA using different fuel sources (for the year 2008)

The capital cost includes:

• Bare plant – engineering, procurement, and construction (EPC)

• The owner’s cost (land, cooling infrastructure, administration and associated buildings, site works, switch yard, transmission, project management, license, etc.)

• Cost escalation due to increased labor and materials

• Inflation

• Financing and interest of financing

The typical construction period of a nuclear power plant is about 48–54 months. Decommissioning cost is about 9–15 % of initial capital cost, which is about 0.1–0.2 cent per kWh of energy generated in the USA. The EPC cost in the year 2008 was about $ 3000/kW.

Spent Fuel and Reprocessing

When the spent fuel assembly is removed from the reactor, it is stored at the reactor site and allowed to cool before reprocessing or disposal. Typical compositions of fresh and spent fuels are listed in Table 2.

Table 2 Typical composition of nuclear fuel and spent nuclear fuel

Most of the commercial reactor spent fuels are in water-filled swimming-pool-type structures. This type of arrangement is chosen because water is inexpensive, has good heat transfer coefficient by convection, and provides shielding, and visibility in water gives an opportunity to detect undesired events, if any. The limitation of water as a cooling medium in spent nuclear fuel is that water is a neutron monitor and active electrolyte for corrosion reactions.

The typical PWR operating cycle is about one year when 1/3 of the core is replaced with new fuel. After one year of operation, the fuel assembly, which weighs about 1300 lbs, is removed from the core and transferred to an interim storage facility. The radiation levels of the unshielded fuel assembly are more than millions of rems per hour.

The spent fuel assemblies are placed in vertical stainless steel racks. In order to prevent reaching critical conditions of the spent nuclear fuel assemblies, these are stored in well-separated conditions. Furthermore, neutron-absolving materials such as boron carbide or boron rods are inserted to inhibit neutron multiplication. The pool storage facility is designed only for interim storage – until the spent fuel is cooled down to low temperature. The remnant radioactive decay has subsided. Afterward the spent fuel will be taken for reprocessing or, in the absence of reprocessing, to a long-term storage facility.

Dry Storage

As an alternate to wet pool storage, dry storage using metal casks and concrete modules is practical. The heat generated during radioactive decay of the spent fuel is removed by the force convection of air, in case of modular concrete vault storage. Metal casks are provided with fins for faster heat transfer. These metal casks, if properly designed, can also be used for transportation of spent nuclear fuels.

For transportation of spent nuclear fuel, the metal casks are provided with (1) protection against direct radiation exposure to workers and the public, (2) provision for radioactive heat removal, and (3) neutron absorbers to prevent criticality. The metal casks can contain about 7 PWR assemblies or 18 BWR assemblies. The body of the cask is made of stainless steel of 5 m long and 1.5 m wide. Shielding is provided by depleted uranium or lead metal. It has an outer stainless steel shell and a corrugated stainless steel jacket that circulates water as neutron shielding fins are provided for external air forced cooling and minimal impact damage.

The spent fuel casks for transportation are constructed so sturdily that it can withstand the impact of being dropped from a height of 10 m onto an unyielding surface (metal anvil) and pass the crash test of a 130 km/h locomotive crash on a stationary cask-loaded tractor-trailer rig. It can also withstand fire for up to a 125 min burn in JP-4 fuel at 980–1150 °C.

Transmutation

Transmutation of transuranic elements such as plutonium, neptunium, americium, and curium can be conducted by irradiating with fast neutrons. In this process, the original actinide isotopes are transformed to radioactive and nonradioactive fission products. This process is important for nuclear waste management, since the isotopes of actinides have half-lives of thousands of years and alpha emitters. Transmuting these isotopes to short-lived fission products helps eliminate the radioactive hazardous associated with long-lived radionuclides.

Reprocessing

The spent fuel contains about 3.5 % fission products that predominantly contain neutron poisons such as Xe¹³⁵ and I-137. Accumulation of fission products and depletion of fissile U-235 in the nuclear fuel make the sustainability of the nuclear chain reaction very difficult. Therefore, the nuclear fuel is removed from the reactor core. Currently, about 10,500 tons (of heavy metal) of spent fuel is disposed every year from nuclear reactors. The purpose of reprocessing is to separate the actinides from the fission products so that it can be reused as nuclear fuel. This decreases the burden on uranium mining and results in a more sustainable use of nuclear energy as a renewable energy source. Reprocessing can be carried over using aqueous or nonaqueous processes.

Aqueous Reprocessing

The aqueous process is based on the solvent extraction. Figure 9 illustrates the process flow. First, the spent nuclear fuel is dissolved in nitric acid. The Zircaloy cladding is removed separately. The aqueous solution containing dissolved spent fuel is taken for the solvent extraction in an organic solution of kerosene containing tributyl phosphate (TBP). When the aqueous solution comes in contact with the organic TBP, hexavalent uranium (U⁶⁺) and tetravalent plutonium (Pu⁴⁺) are extracted by TBP. Almost all the fission products remain in the nitric acid solution which is extracted as high-level liquid waste. In the solvent extraction partitioning step, Pu ⁴⁺ is reduced to Pu³⁺ by adding Li⁺ as a reductant. The Pu³⁺ is removed by dissolving in nitric acid solution. The recovered Pu can be used as a raw material for fast breeder reactor fuels in the future. The uranium species remaining in the solution can be recovered by processing through a series of scrubbing columns and purification columns. The purified uranium can be enriched and used as a fuel after converting to UO₂. The ability to separate plutonium from uranium is considered a potential proliferation concern. Therefore, modifications are made in the PUREX process to avoid separation of plutonium.

Fig. 9

Flow diagram of PUREX process of reprocessing spent nuclear oxide fuels

In the modified processes, uranium is separated while keeping Pu, minor actinides, and fission products in the waste solution. Later, the actinides are separated as a group. Another modification of the PUREX process is coprocessing. If the intent of reprocessing of spent fuel is to use the recovered actinides for producing mixed oxide fuel (MOX), then coprocessing is the right method. In this process, partitioning of U or Pu does not take place. Therefore, proliferation of Pu for weapon is not a concern. In the coprocessing method, 30 vol.% TBP in n-dodecane is used as solvent and a 2.5 M HNO₃ solution is used as scrub solution. The aqueous feed solution containing 4.2 M HNO₃, 2 M UO₂, + Pu, and 1.25 M FP is fed through solvent extraction column of TBP in n-dodecane. Uranium and plutonium are complexed with the TBP, and thus, fission products are separated. The U + Pu complexed with organic phase is washed with dilute nitric acid. The resulting nitrate solution of U + Pu is treated with peroxides or oxalates to form precipitates of U + Pu peroxide or oxalate. These oxalate precipitates are calcined to form UO₃ or U₃O₈ and reduced in hydrogen atmosphere to form UO₂. There are several variations in the PUREX process. Table 3 lists these modified PUREX processes.

Table 3 Variations of aqueous–organic reprocessing of nuclear spent oxide fuels (Adopted from the Nuclear Technology Review Supplement, International Atomic Energy Agency, Vienna, 2008)

Pyroprocessing

Pyrochemical or pyrometallurgical processing using LiCl–KCl molten salt systems is considered one of the most feasible alternatives to the PUREX process for safe and proliferation-resistant recovery of nuclear fuel elements from the spent fuels. This technology may also be useful for separating actinides from the high-level waste generated by the PUREX process. Pyrometallurgical process is preferred because of the stability of the molten salts to high radiation and shorter cooling times (OCDE/NEA Report). Reprocessing of metallic fuels involves separation of actinides from the fission products by electro-transport in a molten salt electrolyte. Since rare earth elements (as part of fission products) have similar chemical properties as that of actinides and show neutronic poison effect, separation of fission products is important for efficiently recycling the actinides. Spent oxide fuels also can be reprocessed by the pyrometallurgical electrorefining method. In this case, the spent oxide fuel is reduced to metal form by lithium (Koyama et al. 2007) or chlorinated in the presence of a reductant such as carbon (Yang et al. 1997) before anodic dissolution or direct dissolution in the presence of an oxidizer such as CdCl₂ into the molten salt (Koyama et al. 1997).

The major advantages of the pyroprocessing spent fuel are as follows:

• The process is proliferation resistant since Pu is not separated from minor actinides.

• Interim storage of spent nuclear fuel may not be required since the pyroprocessing is capable of handling spent fuels in hot conditions as the process takes place in temperatures greater than 500 degrees Celsius.

• No liquid wastage is generated for disposal. Therefore, waste management becomes easy.

• The process can be adopted for in-line reprocessing at the reactor site.

• This process can accept several forms of fuel such as uranium oxide, carbide, nitride, mixed oxides, and pure heavy metals.

• Very short turnaround time results in cost saving.

• Generation of minimum transuranic waste.

The limitations of the process are requirements of facilities with oxygen- and moisture-free environment, arid construction materials that withstand very high temperature, and highly corrosive molten halide environment.

Reprocessing of Spent Metallic Fuel

Metallic fuels are used in experimental fast breeder reactors with liquid sodium as coolant. Reprocessing of this spent fuel (U–Zr, U–Pu + Zr alloys) is carried on by first chopping them into small pieces, loaded onto an anode basket made of SS, and dissolving them by applying anodic potential in an electrorefining cell. The electrolyte is typically an eutectic of LiCl–KCl at 500 °C. By applying an anodic potential to the stainless steel basket containing the chopped fuels, the pellets are oxidized and dissolved in the molten salt. Dissolved actinides are present as chlorides in the molten salt. Lanthanides in the fission product are converted to lanthanide chlorides and dissolved in the molten salt. Addition of CdCl₂ to the LiCl–KCl mixture helps transfer most of the actinides and lanthanides as chlorides in the molten salt bath. Gaseous fission products are out-gased. Undissolved cladding materials and noble fission products will be recovered as solids from the reprocessing cell.

During the electrorefining process, uranium is recovered from the molten salt by application of a constant cathodic current density to a steel cathode in a shape of a cylindrical rod, as shown in Fig. 10. The resultant cathodic potential is just sufficient to electrodeposit only uranium onto the steel cathode. After depositing uranium, when the ratio of plutonium to uranium is greater than 2 (Pu/U > 2), now the electrodeposition process is continued with liquid cadmium as cathode. In this step, plutonium is recovered along with americium (Am) in the form of Pu–xAmxCd6 compound. More than 10 wt% of Pu is collected using this method. A high separation factor between actinides and rare earths within a MClx–LiCl–KCl system has also been reported when liquid bismuth is used as liquid cathode. After the actinide recovery, the molten salt is solidified and scrubbed to remove fission products through a zeolite column.

Fig. 10

Schematic arrangement of electrorefining cell for pyroprocessing of spent nuclear fuel in molten LiCl–KCl

The redox potentials of actinides and lanthanides are given in Tables 4 and 5, respectively. The lanthanides show more negative potentials than actinides. Among the actinides, uranium shows less negative reduction potential than plutonium and americium. Therefore, under a sufficient cathodic polarization, uranium will be reduced first. Electrodeposition of the uranium on the solid steel cathode decreases the concentration of the uranium (III) ions in the melt. Therefore, the redox potential of U(III) will move to more negative potentials with continuation of the electrorefining process. The electrorefining process is switched to liquid cadmium cathode because of the following reasons: (1) Liquid cadmium as cathode decreases the activity of actinides other than uranium as shown in Table 6; (2) the lower activity coefficient brings the redox potentials of all actinides closer so that these elements can be deposited together; and (3) recovery of Pu along with other minor actinides gives better proliferation resistance.

Table 4 Redox potentials and activity coefficients of actinides in LiCl–KCl eutectic melt at different temperatures (Roy et al. 1996, pp 2487–2492)

Table 5 Redox potentials of lanthanides dissolved in LiCl–KCl eutectic at 450 °C

Table 6 Activity coefficients of actinides in liquid cadmium at 450 °C

The five orders of magnitude smaller activity coefficient of Pu as compared to that of U could be attributed to formation of PuCd₆ compounds in the liquid Cd cathode (Shirai et al. 2000). When Pu is electrodeposited onto liquid cadmium cathode, the reduction potential is shifted by 0.3 V in the positive direction as compared to the electrodeposition onto a solid surface. This shift in the positive direction brings the reduction potential of Pu closer to the reduction potential of U(III). The shift in the reduction potential of PU(III) in liquid cadmium cathode can be explained by using the Nernst equation:

$$ {\mathrm{Pu}}^{3+} + 3\ {\mathrm{e}}^{-}\to\ \mathrm{P}\mathrm{u} $$

(25)

$$ {E}^1={E}^0+\frac{2.3RT}{3F} \log \frac{\left[P{u}^{3+}\right]}{\left[\gamma Pu\right]} $$

(26)

Since the value of γ is 3.1 × 10⁻⁵ in the liquid cadmium, the redox potential is shifted almost by 0.25 V in the positive direction. Sustained operation of the electrometallurgical reprocessing cell results in accumulation of fission products in the electrolyte and depletion of the uranium ions in the salt. The variation in composition of the electrolyte could potentially alter the operating conditions of the cell because of the significant changes in the thermophysical properties and interfacial electrochemical behavior of the molten salt systems. For better process control, a detailed database of the electrochemical properties of the molten salt system is required. When multiple fission product elements are present in the electrolyte, the reduction behavior of the actinides could significantly be altered because of possible underpotential reduction of lanthanides and slower diffusion kinetics of actinides. This is important for determining limits on the use of the molten salt electrolyte before it needs to be purified or disposed.

Thermodynamic and transport properties of binary LnX₃–MX systems have been investigated widely (Gaune-Escard et al. 1994; Takagi et al. 1997; Gong et al. 2005) where Ln = La, Ce, Pr, Nd, Gd, Tb, and Eu; M = Li, L, Na, Cs, and Rb; and X = F, Cl, I, and Br. Addition of lanthanide chloride to alkali metal chloride results in formation of a variety of stoichiometric compounds such as M₃LnCl₆, MLn₂Cl₇, M₂LnCl₅, M₃Ln₅Cl₁₈, etc. Formation of compounds and complexes in the molten salt system affects the electrical conductivity and other thermophysical properties. Stoichiometric compounds show minimum electrical conductivity. Structural disordering increases the number of current carriers and improves the conductivity. The specific electrical conductivity of LnCl₃ ranged from 0.11 to 0.4 Sm⁻¹ at 1000–1250 K. The activation energy for electrical conduction was about 28–30 kJ/mol. Polymerization of the melt was reported to play a significant role in increasing the electrical conductivity of the molten salt system². Existence of octahedral complex anions of LnCl₆ ³⁻ in the LnCl₃ melts and formation of dimers have been proposed by the following reaction (Ikeda et al. 1988):

$$ {{2\mathrm{LnCl}}_6}^{3-}\to {\mathrm{Ln}}_2{{\mathrm{Cl}}_{11}}^{5-} + {\mathrm{Cl}}^{-} $$

(27)

Since free Cl⁻ ions are produced by the above dimerization reaction, the conductivity of the melt increases. Both polymerization of melt and presence of free chloride ions could affect the activity and mobility of the cations, and in turn the separation kinetics could be altered.

Standard potentials of actinides in LiCl–KCl eutectic salt and separation of the actinides from rare earths by electrorefining have been widely reported by many research groups (Sakamura et al. 1998; Roy et al. 1996; Serrano and Taxil 1999). Recently, Castrillejo and coworkers (2005c) reported electrochemical behavior of a series of lanthanide elements in LiCl–KCl eutectic melt in the temperature range of 400–550 °C.

Cyclic voltammetry results of binary, ternary, and quaternary LnCl₃-(LiCl–KCl)_Eutectic systems at 500 °C indicate that the incipient potentials of cathodic reduction waves shifted to less negative values with increased additions of lanthanide components. The positive shift in the potential of reduction wave is, in general, associated with two phenomena, viz., (1) under potential deposition, the interaction of reducing species (R) with the substrate (S) is energetically more favorable than the species–species (R-R) interaction, and (2) when two species (A and B) are present in the electrolyte, formation of a compound (A_nB_m) is more favorable by having a negative free energy (−ΔG), and the deposition potential is positively shifted from the redox potential of the more negative species by an amount (−ΔG/nF) (Cohen 1983). The CV results of binary system (single component lanthanide addition) do not show any underpotential deposition of pure lanthanide elements. However, in this investigation, addition of more than one lanthanide chloride in the LiCl–KCl eutectic resulted in considerable shift in the incipient potential of the cathodic wave. According to Hume-Rothery principles, atoms having similar size (size difference <15 %) and electronegativity will preferentially form solid solutions and not compounds. Therefore, elements of lanthanide series will not form compounds within them but can form solid solutions. The commercially available mischmetal is such an extended solid solution of various lanthanide elements that is separated from naturally occurring monazite mineral. Easy formation of solid solution of lanthanide elements indicates that this is a thermodynamically more favorable process. Therefore, any change in the free energy of formation of solid solution will be reflected in shifting the cathodic wave potential to positive direction with reference to the individual elements’ reduction potential.

The other possible reason for the positive shift in the reduction potential of multicomponent lanthanide system could be because of lowered stability of the lanthanide clusters in the molten alkali salt. Generally, solvation of Ln(III) in LiCl–KCl eutectic mixture forms clusters of [Ln(KCl)_n]³⁺ and [Ln(LiCl)_n]³⁺, with the coordination number, n, varying from 4 to 9. The presence of a single lanthanide component in the alkali molten salt results in coordination numbers 6–9. Model calculations of Hazebroucq et al. (2005) indicated that the stable coordination number for Gd (III) was 6 and for La (III) it was 7–8. When multiple lanthanide components are present, the coordination of the solvated clusters is significantly affected and their stability is reduced. Therefore, the positive shift in the reduction potential for multicomponent system could be attributed to the change in stability of the solvated clusters in the molten salt.

Recovery of Rare Earth Elements (REEs) from Fission Products

Rare earth elements are part of strategic materials used in various important energy and defense applications. Separation of rare earth elements from fission products could help reduce the burden on nuclear waste management and reduce the impact on the environment during mining operation of extraction of these strategic materials. Rare earth elements such as neodymium and samarium are used in producing high-strength permanent magnets that are used in various applications such as motors and electronic components. Rare earth oxides are used in laser applications. Almost 38 % of REEs find application as phosphors and solid-state lighting device. Mischmetals are used for manufacturing AB5-type nickel–cobalt hydrides that are used as cathodes for metal hydride batteries and hydrogen storage for electric vehicle applications. Rare earth oxides are extensively used as catalysts for various chemical processes such as methane reformation, fluid cracking, oil refining, and water–gas shift reactions. Furthermore, rare earth oxides are used in fuel cell and high-temperature battery applications as catalysts and membranes. Rare earth elements are used in thermal barrier coatings and strategic alloys as well. Therefore, separation of lanthanide elements from fission products is important from both strategic and environmental points of view.

It is observed that addition of multicomponent lanthanides (more than 5 wt%) to the LiCl–KCl eutectic shifted the reduction potential to less negative values which might hinder the effective separation of actinides from the lanthanides. One possible way of minimizing this effect is to remove the lanthanides as the electrorefining progresses so that accumulation of lanthanide is minimized. Separation of lanthanide can be made possible by using a bipolar cell and a bipolar membrane that has a high diffusion coefficient to lanthanide series. The design of the cell is given in Fig. 11.

Working Principle of the Bipolar Cell for Separation of Lanthanides: Figure 11 schematically illustrates the construction of the bipolar cell. Two compartments are connected by a thin metal diaphragm. This metal diaphragm will preferentially alloy with a particular lanthanide element, for example, neodymium. The right-side compartment will contain molten salt of multicomponent additions that needs to be electrolyzed. The diaphragm as a cathode, an inert electrode as an anode, and Ag/AgCl as a reference electrode will complete the electrical circuit. Depending on the applied potential, a particular lanthanide element that has a higher alloying affinity with the diaphragm metal will deposit onto the diaphragm and diffuse out to the left compartment because of a concentration gradient and a small electric field across the diaphragm. In the left compartment, the diaphragm is connected to a positive terminal of another electric circuit. Therefore, depending on the anodic potential applied to the diaphragm, the diffused lanthanide element will dissolve into the electrolyte of the left-side compartment. The oxidized lanthanide species will get reduced at the cathode of the left-side compartment and thus separated from the other species.

Fig. 11

Schematic of bipolar cell with metal diaphragm that preferentially alloys with select lanthanide(s) and allows diffusion of lanthanide elements to the other side

The main issue of this design is construction of a stable diaphragm that is thin enough to allow a required flux of the lanthanide. Aluminum diaphragms can be used for separating Nd. Nickel diaphragm also can be used. A thin diaphragm of Ni can be prepared by electroless deposition onto an organic substrate.

Pyroprocessing of Spent Oxide Fuels

Spent nuclear oxide fuels also can be reprocessed using molten salt refining technique as that of metal oxide fuels. The electroreduction technique used for directly reducing uranium oxide to uranium is similar to the technique of the FFC Cambridge process proposed for directly reducing TiO₂ to Ti (Chen et al. 2000). Alternately, the oxide fuels can be reduced by using lithium. The first step in chemical reduction of UO₂ is to convert the material to U₃O₈. Similarly, PuO₂ is also converted to Pu₂O₃ before reacting with lithium. The proposed reactions are

$$ {\mathrm{U}}_3{\mathrm{O}}_8 + 16\ \mathrm{L}\mathrm{i}\ \to\ 3\mathrm{U} + 8{\mathrm{Li}}_2\mathrm{O}\kern4.4em \Delta \mathrm{G} = -871.8\ \mathrm{kJ} $$

(28)

$$ {\mathrm{Pu}}_2{\mathrm{O}}_3 + 6\ \mathrm{L}\mathrm{i}\ \to\ 2\ \mathrm{P}\mathrm{u} + 3\ {\mathrm{Li}}_2\mathrm{O}\kern2em \Delta \mathrm{G} = 18.8\ \mathrm{kJ} $$

(29)

$$ {\mathrm{Am}}_2{\mathrm{O}}_3 + 6\ \mathrm{L}\mathrm{i}\ \to\ 2\ \mathrm{Am} + 3\ {\mathrm{Li}}_2\mathrm{O}\kern1.6em \Delta \mathrm{G} = 23.18\ \mathrm{kJ} $$

(30)

Mixed oxide spent fuels can be electrochemically reduced more easily than UO₂ in a LiCl bath at 650 °C. In the uranium oxide electroreduction cell as presently conceived, a platinum (Pt) wire is used as the anode.¹ The Pt wire is slowly etched away as the electrolytic reduction proceeds because of the highly oxidizing conditions and attack by lithium (Li). The cell is operated at 650 °C with a LiCl–Li₂O electrolyte. The overall cell reaction is as follows:

$$ {\mathrm{UO}}_2\to\ \mathrm{U}\ \left(\mathrm{cathode}\right) + {\mathrm{O}}_2\uparrow \left(\mathrm{anode}\right) $$

(31)

The Li₂O content varies from 1 to 8 wt% for the Li-assisted chemical reduction of UO₂ by participating in the following reactions:

$$ 2{\mathrm{Li}}_2\mathrm{O}\to 4\mathrm{L}\mathrm{i} + {\mathrm{O}}_2\uparrow $$

(32)

$$ 4\mathrm{L}\mathrm{i} + {\mathrm{UO}}_2\to \mathrm{U} + 2{\mathrm{Li}}_2\mathrm{O} $$

(33)

Although the reduction of Li₂O to Li increases the reduction rate of the spent oxide fuel, the Li diffuses through the salt electrolyte and attacks Pt, thereby degrading the anode. To decrease degradation of the Pt anode, a secondary electric circuit is provided to oxidize the Li to Li(I). Provision of a more refractory anode that withstands the highly oxidizing conditions and attack by Li would greatly simplify operation of the cell.

Therefore, finding an alternative anode material is vital before the electrolytic cell is put into production. Platinum metal is highly resistant to oxidation under normal conditions, but not under high anodic potential at 650 °C in a corrosive Li containing molten chloride electrolyte. Several materials are worthy of consideration as inert anodes. Various nickel-based super alloys such as Haynes 263, Haynes 75, Inconel 718, Inconel X-750, Inconel 713 LC, Inconel MA 754, Nimonic 80A, and Nimonic 90 have been investigated by KAERI.² It was noted that increasing the chrome (Cr) contents of the above alloy increased the stress on the surface layer and increased the corrosion rate. Elements such as aluminum (Al) and titanium (Ti) improved the corrosion resistance by forming a more protective layer. It should be pointed out that these test conditions did not include metallic Li in the electrolyte. The solubility of Li in LiCl was more than 0.6 mol%. Oxidizing environment was created by purging argon + 10 % O₂ gas into LiCl + 3 wt% Li₂O electrolyte at 650 °C. Boron-doped diamond as a potential anode material has also been reported (Storm van Leeuwen and Smith 2005).

Inert anode material for electrolytic reduction of the spent fuel in LiCl + Li₂O electrolyte should satisfy the following requirements:

• The material should have comparable electrical conductivity to that of platinum.

• The material should not be consumed by the electrochemical/chemical reaction so that the dimensions remain stable throughout the electrolytic process without altering the cell voltage significantly.

• The material should be resistant to corrosion by lithium in the molten salt.

• The protective oxide layer of the material must not dissolve in the basic flux of Li₂O present in the molten salt.

• The material must be stable in the highly oxidizing potentials encountered at ~650 °C in LiCl + Li₂O electrolyte.

• The material should possess sufficient mechanical properties (such as creep strength and fracture toughness) in order to withstand thermal shock and service-related stresses.

The material fulfilling the above requirements can be a metal or an alloy with high electronic conductivity. The material can also be a composite (cermet) or made of metal/alloy coated with external oxide/carbide/nitride. This article describes the investigation of inert anode materials for use in an electrochemical cell by simulating the corrosive environment of the spent nuclear fuels with LiCl and LiCl + Li₂O molten mixtures.

Containment of Radionuclide

Development of next-generation nuclear waste forms capable of reliable and safe immobilization of radionuclide is a significant challenge that limits the imminent nuclear renaissance to meet the energy demands. Among the nuclear wastes, Cs and Sr are considered critical because of the high activities of ¹³⁴Cs (half-life, 2 years), ¹³⁷Cs (half-life, 30.17 years), and ⁹⁰Sr (half-life, 28.8 years). Removal of these isotopes will significantly reduce the radioactive load on a geological repository (Wigeland et al. 2006). These radionuclides are separated from the low-level nuclear wastes using an ion-exchange process and combined with high-level waste fraction and then immobilized in a glass or ceramic form. The following transmutation reactions take place during radioactive decay of ¹³⁷Cs and ⁹⁰Sr:

$$ {}^{137}\mathrm{C}\mathrm{s}\ {\to}^{137}\kern-0.3em *\mathrm{Ba} + \beta \left(\mathrm{half}\ \mathrm{life}:\ 30.17\ \mathrm{years}\right)\ {\to}^{137}\mathrm{Ba} + \gamma \left(2.55\ \min \right) $$

(34)

$$ {}^{90}\mathrm{S}\mathrm{r}\ {\to}^{90}\mathrm{Y} + \beta \left(\mathrm{half}\ \mathrm{life}:\ 30.17\ \mathrm{years}\right)\ {\to}^{90}\mathrm{Z}\mathrm{r} + \beta \left(64.1\ \mathrm{h}\right) $$

(35)

The decay of ¹³⁷Cs emits beta particles with energy ranging from 0.089 to 1.454 MeV and gamma rays with 0.475–1.168 MeV energy (Lide 1997). The beta and gamma decays of the radionuclides could detrimentally affect the stability of the host material. Furthermore, formation of the transmuted elements in the host materials alters the electronic structure and chemical composition of the waste form. Most of the investigations so far have concentrated on the effect of radiation damage such as swelling, bubble formation, and leachability on the structural stability of the waste form. Very little attention has been paid on the effect of electronic structure changes during the decay of constituent radioisotopes in the waste form. For example, when the Cs⁺ transmutes to Ba²⁺ + e⁻(β) + γ, the single valent Cs is replaced with divalent Ba. Therefore, the host material should be able to accommodate the valence and associated ionic radius changes. Recently, Jiang et al. (2009) from LANL reported for the first time the modeling of the chemical evolution of CsCl to BaCl due to radioactive decay using ab initio calculations. In the modeling calculations, these authors considered CsCl crystal as a representative waste form.

Crystalline silicotitanate (CST) and (Ba,Cs) hollandite ceramics are considered potential candidates for the specific immobilization of Cs. The sodium silicotitanate (Na₂SiTi₂O₇. 7H₂O) is ion-exchanged to form hydrogen crystalline silicotitanate (H₂SiTi₂O₇.1.5H₂O). This material effectively sequesters both Cs and Sr (Celestian et al. 2008). The hollandite for Cs sequestration has a formula of (Ba_xCs_y)(M_2x+yTi_8−2x−yO₁₆), where M is trivalent cations such as Fe³⁺, Al³⁺, and Ti³⁺ and x + y < 2. Celestian et al.⁴ showed that selective ion exchange of Cs in the H-crystalline silicotitanate (H-CST) is achieved by repulsive forces between the Cs⁺ and H₂O dipole that lead to a series of events at the molecular-scale level such as rotation of H₂O by 159° followed by bending away of hydroxyl group by 0.055 nm displacement which makes TiO₆ to rotate about 5.6°. The rotation of the TiO₆ columns results in a structural transformation that changes the initial P4₂/mbc space group of the H-CST to P4₂/mcm of Cs-CST. During this transformation, an initial elliptical eight-membered ring (8MR) channel becomes a circular one that opens up a new site for Cs⁺ occupancy at the center of the channel (known as Cs1 site) in addition to the initial Cs2 site which is outside the 8MR window. A minimum occupancy of 0.15 at the Cs2 site is required to initiate the structural transformation and formation of the new Cs1 sites. The above hypothesis for the selective ion-exchange mechanism of Cs in the H-CST material indicates that the interaction of Cs⁺ and H₂O dipole/hydroxyl groups is very important not only for the site selectivity but also for the stability of the Cs-CST structure. It is not clear how the radioactive transmutation of Cs⁺ to Ba²⁺ will affect the stability of the Cs-CST structure over an extended period of time. In case of the hollandite, structural transformation from tetragonal to monoclinic at room temperature is reported with increased Ba occupancy in Ba_xFe_2xTi_8−2xO₁₆ (Carter 2004).

Nuclear Waste Management

All types of radioactive waste can be disposed of if the disposal method provides protection for the health and safety of people and the environment. Members of the European Union (EU) produce about 7000 m³ of high-level waste from 143 nuclear power plants. Disposal in deep underground engineered facilities is considered the best solution for managing high-level and long-lived radioactive wastes. High-level waste is mixed with glass and vitrified. The vitrified waste is stored for 30–50 years and allowed to cool. After cooling, the waste is placed in an iron shell container. The iron container is placed inside a copper shell, which is evacuated and backfilled with an inert gas and sealed. The copper canister is buried in a deep permanent repository with engineered barriers. Swedish and Finnish nuclear waste repository models are based on copper canisters, since copper is considered to be stable in clay environments devoid of oxygen. Several countries plan to use this approach of permanent storage of the high-level radioactive waste.

When the amounts of radioactive waste in surface storage increase, the sustainability of storage in the long term and the associated safety and security implications are of concern. Geological disposal promises to provide containment and isolation of radioactive waste from the human environment for the very long periods required. Safety concerns due to possible human intrusion into the waste are very much reduced as compared to surface storage, owing mainly to the significant depths under the surface at which geological repositories will be located. In the USA, long-term storage of nuclear waste in Yucca Mountain was considered an option. The Yucca Mountain repository was based on different layers of engineered barriers to nuclear waste storage such as a 300 m deep geological barrier, a Ti-alloy drip shield to prevent water seeping through rock faults falling on the canister surface, an outer wall of canister made of Ni–22Cr–13Mo–3Fe alloy, and a thicker inner wall of canister made of type 304 stainless steel. Recently the US Department of Energy filed a motion with the Nuclear Regulatory Commission to withdraw the license application for a high-level nuclear waste repository at Yucca Mountain. Therefore, the spent fuel assemblies from the nuclear power plants will be stored on-site in the utility facilities for longer time.

Future Direction

Meltdown of the nuclear core is considered the severest form of nuclear accident, since the probability of release of radioactivity is high in this condition. In order to prevent core meltdown, Western nuclear power plants are provided with two or four emergency core cooling systems (ECCS). This system consists of high-pressure coolant injection system, depressurization system, low-pressure coolant injection system, core spray system, containment spray system, isolation cooling system, and emergency electrical system. The emergency electrical system consists of diesel generators, motor generator flywheels, and batteries. In case of an emergency situation, the control rods are moved completely inside the reactor core and the power is reduced considerably. Any loss of coolant will trigger the ECCS. There are two or four ECCS in a reactor to ensure that at least one will respond and meltdown will be avoided. However, if all ECCS fail as in the case of Fukushima nuclear plants in March 11, 2011, meltdown of the core is initiated. It should be noted that the Fukushima nuclear disaster is not due to operator error or gross violation of safety regulations unlike the accidents reported in the Three Mile Island or Chernobyl. The plant suffered a major damage because of an earthquake of 9.0 in a Richter scale. The reactors were designed for a maximum ground acceleration of 0.18 g (1.74 m/s²), whereas the earthquake caused a ground acceleration of 0.35 g (3.43 m/s²). Therefore, the reactors were shut down automatically. This would have triggered the insertion of control rods and ECCS. However, a tsunami of 20 m tall waves followed the earthquake and flooded the power plant. The plant was designed only for a tsunami of 5.7 m. The flooding situation knocked down the emergency power and prevented any assistance reaching the power plant. The batteries provided in the emergency electrical system were not adequate to pump the required volume of coolant to the core. The power required for running coolant pumps or other electrical systems triggered a cascade of accidents in the Fukushima nuclear plant that released radioactive gases into the environment causing displacement of several thousand people living around the power plant.

In order to cool the core, several measures were taken by the Fukushima power plant officials. Adding water to the degraded core can result in several consequences such as (Kuan and Hanson 1991):

• Hydrogen production

• Change in the geometry of core

• Pressurization of the system due to high rate of steam generation

• Steam explosion

• Re-criticality of the core when enough neutron absorbers are not present

The core damage occurs in several stages as explained below (Cookson and Was 1995):

• Pre-damage Stage: If the core is not fully immersed in water, then the upper portion of the core will be exposed to steam in the reactor. Now the core will start to heat up at a rate of 0.3–1 °C/s.

• Ballooning and Bursting of Fuel Rods: When the temperature reaches > 1100 K, the Zircaloy cladding will balloon up because of the rapid heating and burst. This altered geometry of the fuel rod will affect the geometry of the coolant flow channels in the core. Some locations will have restricted access to the coolant because of the ballooning effect. If sufficient water is added, core damage can be suppressed at this stage.

• Rapid Oxidation: This stage is initiated at 1500 K. When Zircaloy reacts with steam, hydrogen is produced as given by the following reaction and a large amount of heat is released:

  $$ \mathrm{Z}\mathrm{r} + 2{\mathrm{H}}_2\mathrm{O}\ \to\ {\mathrm{ZrO}}_2 + 4{\mathrm{H}}_2 + 6.5\ \mathrm{M}\mathrm{J}/\mathrm{kg}\ \mathrm{of}\ \mathrm{Z}\mathrm{r} $$

• If water is added at sufficient rate and volume, the core will be quenched and progression of damage could be stopped. If the water is not sufficient or the rate of heat removal is less than the rate of heat generated, the damage propagates to the next stage.

• Debris Bed Formation: When the temperature reaches 1700 K, the molten control materials will flow to the lower part of the core (which is submerged in the water) where the temperature is low and solidify. At 2150 K melting of Zircaloy occurs. Molten Zircaloy along with dissolved UO₂ may flow downward and solidify at the lower portion of the core. These solidified debris will form a cohesive bed leading to restricted flow of coolant in the lower region of the core.

• Relocation of Lower Plenum: When molten core materials (which are experiencing 1500–2150 K) fall to the lower region of the core which is at ~550 K, steam is generated rapidly leading to occurrence of steam explosion. Furthermore, this steam oxidizes any unoxidized molten Zircaloy which generates hydrogen at a faster rate. These reactions lead to overpressurization of the system. Re-criticality also may occur in the relocated core debris when the control materials are not present in the required concentration.

Understanding of the sequence of core damage is necessary to design preventive measures of core meltdown. Future work on nuclear safety should concentrate on a reliable ECCS that can be operated even in the worst-case scenario as experienced in the Tohoku Tsunami. Future work also should focus on a reliable system, with public acceptance, for a long-term safe storage of nuclear spent fuel.

Future Fuel Cladding Materials: Zr–Sn alloys such as Zircaloy-2 and Zircaloy-4 are currently used as fuel cladding tubes in the current light water reactors because of their low neutron absorption cross sections for thermal neutrons, reasonable creep resistance, and corrosion resistance in high-temperature high-pressure water (Wray and Marra 2011). These cladding materials perform well under normal operating conditions and give a reasonable safety margin under design basis accident (DBA) scenarios. However, under beyond design basis accident (BDBA) conditions, such as a loss-of-coolant accident event that occurred in the Fukushima Daiichi power plant, zirconium-based cladding materials undergo severe degradation because the peak clad temperature (PCT) exceeds the design limit of 1204 °C (Charit and Murty 2008). When the Zr-alloy cladding is exposed to high-temperature steam environment, an exothermic Zr-steam reaction generates more heat than that of radioactive decay which in turn oxidizes the entire cladding material. The current US design regulation (10 code of Federal Regulation 50.46) limits the equivalent cladding reacted (ECR) thickness to 17 % of the initial cladding thickness under DBA conditions. Furthermore, copious amount of hydrogen is generated during the steam oxidation reaction of zirconium that may result in explosion. Therefore, one of the goals of the Fuel Cycle R&D program is to develop high-performance LWR fuel and cladding materials that are resistant against different severe accident scenarios. In addition to the enhanced safety margin, the next-generation fuel clads should have the required properties to perform under high-burnup operating conditions (> 40 MWd/kg of U). At a high level of burnup, high fission gas pressures are realized along with higher creep deformation. In addition, neutron damage to the cladding makes it more susceptible to failure. There could be situations of fuel cladding chemical interaction (FCCI) involving fuel constituent redistribution (Carmack et al. 2009). Hence, improved cladding and matrix materials for pin-type and dispersion-type fuels with low FCCI potential, high strength, radiation tolerance, and high-temperature oxidation resistance are highly desirable for accident-tolerant fuel cladding materials.

Recently, renewed interest has emerged in aluminum-bearing ferritic alloys despite the neutronic penalty in LWR applications. For example, the APMT alloy (nominal composition, Fe-22 Cr-5 Al-3 Mo- < 0.05C, wt%) is being considered for its extreme high-temperature oxidation resistance even beyond 1200 °C due to the protective nature of alumina-based scale (Terrani et al. 2013). This alloy is conventionally used in high-temperature furnace elements. While the alloy has shown promise in terms of oxidation resistance at elevated temperatures, this alloy has not been adequately assessed for advanced fuel cladding applications. Furthermore, addition of “reactive” elements such as Y, Hf, Zr, etc., has been considered to improve the oxidation resistance of alumina-forming alloys (Guo et al. 2014). The details of growth stresses during steam oxidation of alumina layers and the effect of reactive elements on the diffusion and electronic behavior of the oxide layers are not studied in detail. Such an understanding is pertinent for the design of new FeCrAlRE cladding materials that show improved LOCA resistance. In addition to FeCrAl alloys, other materials such as Mo (Nelson et al. 2013) and ferritic ODS alloys (Klueh et al. 2005) are also actively investigated for fuel cladding applications. The design of the new cladding alloy will be based on the following considerations (Knief 1992; Pint et al. 2013):

• The target mechanical properties under unirradiated conditions:

  • The tensile strength at room temperature will be greater than 600 MPa.

  • The yield strength at 1200 °C will be about 100 MPa (versus ~50 MPa of the Zr-4 alloy at 800 °C).

  • A 100 h creep rupture strength at 1200 °C will be about 50 MPa (versus 5 MPa at 800 °C of the Zr-4 alloy).

  • Elastic modulus ~100 GPa at 1200 °C.

• Understanding irradiation effects:

  • Formation of dislocation loops and α′ phase; phase stability.

  • Fracture toughness after irradiation to 20 dpa level is ~50 MPa√m (compared to 12–15 MPa√m of Zr alloy); dimensional changes <1 % at 20 dpa.

• Understanding of corrosion behavior:

  • Cr-rich oxide layer that is expected to impart resistance to aqueous corrosion

  • Al₂O₃ layers associated with the high-temperature oxidation resistance (>600 °C)

  • Effect of reactive elements (actinides, Zr, Hf, Sc, etc.) on the diffusivity of Al³⁺, V_Al ³⁻ V_O ²⁺, and O²⁻ and adhesion of oxide layer

  • Understanding the origin of oxide growth stresses during steam oxidation, electronic properties, and the stability of oxide layer under LOCA condition

It is well documented (Lim et al. 2013) that higher concentration of Cr in ferritic steel leads to Cr-rich α′ and σ phase formation during thermal aging between 350 °C and 550 °C. Since the normal operating temperature of the LWR falls in the embrittling temperature range, the effect of spinodal decomposition should be considered. It is observed that Al partitions to Fe-α phase and the partitioning factor increases with the aging time in Fe–20Cr–5Al ODS alloy (Capdevila et al. 2008). Under the LOCA condition, the α′ phase would be dissolved in the matrix, and therefore, spinodal decomposition may not be an issue. Since the formation of α′ does not affect the distribution of scale-forming Al, high-temperature oxidation resistance of the alloy may not be impaired by the embrittlement aging at low temperatures. However, ductility will be severely affected. The low-temperature (up to ~350 °C) corrosion resistance of the FeCrAlRE alloys will be imparted by a Cr-rich oxide layer in the high-temperature high-pressure water under normal operating conditions. The required oxidation resistance under LOCA conditions could be attributed to the formation of an impervious α-Al₂O₃ film which is stable at temperatures above 1040 °C. Transient aluminum oxides such as γ-Al₂O₃ and δ or θ-Al₂O₃ are stable at temperature ranges 500–800 °C and 800–1040 °C, respectively. The transformation of transient oxides into α-Al₂O₃ is accompanied by a 10 % volume contraction that results in accumulation of tensile stresses. If the oxide scale contains multiple oxide phases, the mismatch in the coefficient of thermal expansion again leads to build up of stresses. When starvation of oxygen occurs during high-temperature exposure, generation of oxygen vacancies (V_O ²⁺) is expected at the expense of oxygen sublattice (O_O ^x) following the reaction:

$$ {{\mathrm{O}}_{\mathrm{O}}}^{\mathrm{x}}\to\ \mathrm{\frac{1}{2}}\ {\mathrm{O}}_2 + {{\mathrm{V}}_{\mathrm{O}}}^{2+} + 2{\mathrm{e}}^{-} $$

(36)

Similarly, under oxygen-rich conditions, aluminum ion vacancies could be generated by incorporating the oxygen atoms into the lattice from the adsorbed oxygen molecule following the reaction:

$$ \mathrm{\frac{1}{2}}\ {\mathrm{O}}_2\to\ {{\mathrm{O}}_{\mathrm{O}}}^{\mathrm{x}} + 2/{{3\ \mathrm{V}}_{\mathrm{Al}}}^{3-} + 2{\mathrm{h}}^{+} $$

(37)

These cation and anion vacancies are important in the formation of oxide layer through the reaction

$$ {{2\mathrm{V}}_{\mathrm{Al}}}^{3-} + {{3\mathrm{V}}_{\mathrm{O}}}^{2+} + {{2\mathrm{A}\mathrm{l}}_{\mathrm{Al}}}^{\mathrm{x}} + {{3\mathrm{O}}_{\mathrm{O}}}^{\mathrm{x}}\to\ {\mathrm{Al}}_2{\mathrm{O}}_3 $$

(38)

However, when the concentration of the vacancies reaches a nonequilibrium condition, the stability of the oxide layer is affected by forming porosity either at the oxide/atmosphere interface due to condensation of oxygen vacancies or at the oxide/metal interface due to condensation of cation vacancies. Since grain boundaries act as short circuit diffusion paths for the transportation of atoms and ions, the presence of aliovalent ions in the oxide layer and reactive elements at the grain boundaries of the alloy could significantly alter the diffusivities of both oxygen and aluminum species. Hindering the diffusion of species that form an oxide will significantly decrease the oxidation rate. In addition to affecting the diffusivities, the RE can also modify the electronic states of the oxide layer and thereby affect the oxidation kinetics (Heuer et al. 2011).