Economically competitive fusion
David J. Warda, and Sergei L. Dudareva,
aCulham Science Centre, Abingdon OX14 3DB, UK
Available online 27 November 2008.
Not since the oil crisis of the 1970s has the perception that energy is a crucial and precious resource been as strong as it is today. The need for a new approach to world energy supply, driven by concerns over resources, pollution, and security, is leading to a reappraisal of fusion. Fusion has enormous potential and major safety and environmental advantages, and hence could make a large difference to energy supplies.
As a result of the oil crisis three decades ago, the present generation of large fusion devices, such as the Joint European Torus (JET), the EU-funded machine located at Culham in Oxfordshire, UK, was established to move fusion from small laboratory experiments up to significant power output levels. JET can now produce many megawatts of fusion power in pulses routinely lasting many seconds and up to a minute. The next generation of fusion machines was expected to be constructed in the 1990s, and to be on the power plant scale, producing hundreds of megawatts of fusion power, but the construction of this new generation was prevented by the cuts in energy R&D funding that occurred in the 1990s. The post-oil crisis wave of R&D funding, followed by a progressive decline, is illustrated in Fig. 1, which shows UK public-sector funding for fusion, renewables and energy efficiency. Only now is construction beginning of a fusion machine of this scale. The new reactor, called ITER, is being built in the south of France, as an international venture.
Fig. 1. UK public-sector energy R&D showing how the wave of investment after the 1970s oil crisis was largely curtailed in the 1990s. The global picture is similar. There are now signs of a new wave of investment, particularly in renewables. Source: IEA.
With substantial fusion power now regularly produced in fusion machines worldwide (JET peak power output is now 16 MW), and construction of ITER now beginning, attention has recently turned increasingly to the problems associated with materials, and general issues of engineering and technology needed for a fusion power plant. These problems arise because of the hostile environment of a power plant; particularly in terms of high power load on the plasma-facing materials and of their bombardment by high-energy particles. Three fundamental questions need to be answered:
- (1) Can the plant tolerate the high power production even for reasonably short times?
(2) If it can, are there materials, either already available or those that can be developed in the near future, which will survive a sufficiently long time for the plant to be economic?
(3) Can this be achieved with structural materials which have the low activation characteristics needed to ensure there is no long-lived radioactive waste to allow reprocessing and reuse on a commercially viable timescale?
Operation of existing machines is already influenced by the first problem but the second and third are largely irrelevant in present devices. Here we will look mostly at the second problem: the implications of economics for fusion materials. (The third problem is thought to be readily solvable through the appropriate selection of the chemical composition of materials and through the elimination of impurities forming radioactive isotopes in nuclear reactions initiated by fusion neutrons.)
In designing a fusion power plant, the design choices are fundamentally dependent on the materials properties and in this way they influence the economics of the plant. For instance, if materials properties limit the power-handling capabilities, one might think that a reduced power density could be achieved by design in the form of a larger device that would inevitably be more expensive. Similarly, one could pose a question whether a reduced materials lifetime could be accommodated by design. Inevitably these interactions between material properties, design, and economic viability are quite complex, as confirmed by the analysis given below.
The ‘fusion world’
The fusion reaction of interest is between two isotopes of H, namely deuterium and tritium. These fuse to create He and a free neutron, which carries away most of the energy of the fusion reaction. The fusion process takes place at a very high temperature, of the order of 100 million degrees Celsius, in a plasma of free nuclei and electrons.
Figure 2 shows a schematic of a fusion power station. The most important elements (highlighted in Fig. 2) are the magnets, which confine the plasma and partially insulate it against heat loss, allowing the energy-carrying neutrons to escape; the vacuum vessel, which prevents the deuterium–tritium fuel from being contaminated by other gases; the blanket, which absorbs the high-energy neutrons produced in the fusion reactions in the plasma and extracts their energy through collisions with atoms in the blanket materials whilst at the same time producing, by reaction between the neutrons and Li compounds, the tritium that is subsequently burnt in the plasma; and the divertor, which, like the exhaust on a car, extracts the burnt fuel (He) thereby maintaining the purity of the deuterium–tritium mixture.
Fig. 2. Schematic view of a fusion power plant1, total height approximately 20 m. The items for which material performance is particularly crucial are the blanket, shown in blue, and the divertor, red. The vacuum vessel, gray, has access ports for maintenance which pass between the magnets, brown.
Of these systems, the only one that has not been used up to now in a fusion device is the blanket, although the demands on the other systems will be much greater in a power station than in existing experimental devices.
An important point about Fig. 2 is that the vacuum vessel separates the world outside the vessel from the ‘fusion world’ inside the vessel, where structures are subject to high heat loads and bombardment by high-energy particles – neutrons, photons, and charged particles2. Materials near the fusion fuel are subject to neutron fluxes up to 100,000,000 times higher than materials outside the vessel, such as the magnets. This article is concerned with the materials inside the vacuum vessel, those that lie in the fusion world.
The materials placed in the fusion world have novel requirements. Those in the divertor need to handle large power densities, around 10 MW/m2, due to the direct bombardment of the surface of the divertor by charged particles and photons. The side of the blanket facing the fusion plasma (known as the first wall) is exposed to lower but still appreciable power densities, of the order of 1 MW/m2, due to the flux of fusion neutrons that carry energy through it into the bulk of the blanket. Neutrons, due to their relatively small cross-section of interaction with atoms (the cross-section of elastic scattering of a fusion neutron by the nucleus of an iron atom is of the order of 10−24 cm2, see Ref. 3), penetrate fairly deeply into the blanket, gradually slowing down through collisions with nuclei of atoms in the material. These collisions initiate events known as ‘collision cascades’, in which atoms receiving initial neutron impacts collide with other atoms, resulting in the local melting and resolidification of the material, and in the formation of radiation defects (pairs of vacancies and self-interstitial atoms). The often-used measure of the effect of interaction of neutrons with a material is the number of times an atom is displaced from its position in the lattice, the displacements per atom (dpa). We see that the exposure to the fusion environment presents new challenges for materials science, which are extensively described in literature on the subject2, and are briefly elucidated below.
The range of materials that can be used in fusion applications is restricted by the requirement that they must not generate long-lived radioactive waste when exposed to the fusion environment. This immediately rules out an important subset of alloying elements routinely used in conventional steels such as Nb and Mo. Fortunately there are steels and other candidate structural materials, including V alloys and silicon carbide composites, where high-activation elements are replaced by low-activation elements, giving the radioactivity that arises in the structure of the fusion device a characteristically short half-life, substantially reducing the need to consider repository disposal of waste materials. In the medium term, low-activation materials can reduce activity levels by a factor of a 1000. An additional advantage of low-activation materials is the safety advantage of reduced levels of decay heat present in the plant. This means that even a complete loss of cooling cannot cause melting of the plant, resulting in a high degree of passive safety.
Materials for a blanket
The main functional material used in a fusion blanket is a compound of Li. Different designs use different compounds but two common choices are lithium orthosilicate, in a pebble bed, and lithium lead, as a liquid metal, the latter illustrated in Fig. 3. In either case, the purpose of the Li compound is to generate the H isotope, tritium, which together with deuterium extracted from ordinary water forms the fusion fuel. The Li compound is therefore consumed; neutron bombardment of this compound is not a concern, but rather the means for generating the tritium.
Fig. 3. An example of a complex conceptual blanket design in which the Li is in the form of lithium lead, whilst the first wall is cooled with helium1.
The remainder of the blanket is a load-bearing box structure, to contain the Li compound, with cooling channels, to remove the deposited heat. In the example shown in Fig. 3 there are effectively two coolants, where the lithium lead itself flows and carries heat away, whilst He is used to cool the first wall. The two areas where maintaining integrity of the structure is most crucial are the first wall part of the box, which absorbs energy coming directly from the plasma and is subject to a large neutron dose, and the remainder of the blanket, including coolant channels, which are primarily subject to neutron bombardment. The coolant channels in the above design are inserts of silicon carbide composite which have no structural function but act as thermal barriers allowing operation of the coolant at temperatures higher than the bare steel structure would tolerate.
Figure 4 shows the radial variation of power deposition within the blanket structure, calculated using the blanket design shown in Fig. 3. It is primarily material near the plasma edge that is subject to heating, and hence damage, by neutrons. Analysis based on collision cross-sections for fusion neutrons with nuclei of atoms in the material3, and on computer models describing generation of defects in cascades initiated by neutrons,  and , shows that the most exposed plasma-facing parts of the blanket will be sustaining radiation damage at the rates of the order of 10 dpa per year, while the level of exposure to fusion neutron irradiation in other parts of the blanket is much lower and . Fig. 5 and Fig. 6 follow this up, showing recent experimental data that illustrate changes in microstructure and material properties resulting from exposure to irradiation. Perhaps the most significant point illustrated by Fig. 5 and Fig. 6 is the fact that properties of materials change dramatically over a narrow interval of variation of parameters characterizing the environment in a fusion power plant, such as the irradiation dose, temperature, or mechanical load,  and . The deterioration of mechanical properties of candidate fusion steels saturates with increasing dose,  and , but it is clear that very high neutron fluences approaching 150 dpa will be very challenging for materials to tolerate15. In a typical design of a fusion power plant, but not in ITER, this means that the blanket must be replaced typically every 5–10 years. As an experimental device, ITER will not run for months at a time and so will not accumulate the large neutron dose that will be inevitable in a fusion power station.
Fig. 4. Power density, deposited by neutrons, falls off quickly with distance into the fusion blanket. Most of the energy is deposited in the first 30 cm of material1.
Fig. 5. Dislocation microstructure formed in Fe at two different temperatures, 300°C (top) and 500°C (bottom), under irradiation to 6.5 dpa10. Dislocation loops seen in these electron microscope images, as well as small defects invisible in these images, impede plastic deformation, resulting in hardening and embrittlement of the material. The difference between microstructures formed at two different temperatures is due to the extreme elastic anisotropy of iron at high temperature11.
Fig. 6. Increase in the Ductile–Brittle Transition Temperature (DBTT), i.e. the temperature above which the material is ductile, and below which it is brittle, for EUROFER97 steel shown as a function of irradiation dose at 300°C,  and . The origin of the vertical axis is approximately at −70°C. Note the sharp rise in DBTT as a function of irradiation dose in the interval between 0 and 3 dpa, saturating at 30 dpa.
The above studies do not address the effects of He production in the lattice due to transmutation nuclear reactions initiated by fusion neutrons. In this respect the possibility of performing tests still remains limited, but it is already clear that the presence of He will make materials more brittle due to accumulation of He at boundaries between crystal grains in a metal. He embrittlement is not expected to saturate as a function of irradiation dose. In addition, production of He stimulates dimensional changes, giving rise to radiation-induced swelling. A new facility, IFMIF, intended for testing materials in neutron spectra very similar to fusion, is under design16. Imaginative mock-ups of fusion conditions through artificial He production by, for instance, irradiation of B-doped materials, give a range of results, from discouraging to more promising. Only by building a test facility such as IFMIF, or a fusion power plant itself, will we be able to carry out these tests convincingly.
There are equally drastic positive changes in the behavior of materials that suggest viable practical solutions for the technology of fusion power generation. For example, by using high operating temperatures it is possible to significantly reduce the effects of radiation hardening and embrittlement (i.e. to decrease the Ductile–Brittle Transition Temperature, DBTT). In qualitative terms this can be explained by the enhanced diffusion and recombination of radiation defects at higher operating temperatures.