Together with other minor factors, it is tokamak fuel kinetics that determines the fraction of fusion energy released as particle kinetic energy useable as thrust, and that in the undesirable form of radiation (see Section 7.4.1).
Among the many technical issues associated to fusion and fusion propulsion, that of radiation and its shielding occupies a special place. Some of the radiation is useful, e.g., neutrons are indispensable to convert the liquid lithium coolant blanket behind the first wall into tritium needed by reactions 2, 4 or 5a of Figure 8.8; with D-T kinetics, in fact, most of the energy is deposited inside the lithium coolant by the neutron flux, of order of MW/m2; however, most other effects damage structural materials and body tissue (see Appendix A). Particles, especially high-energy neutrons, and gamma photons radiated during fusion carry enough energy to penetrate solid material and dislodge atoms from their crystal lattices. With respect to fission, fusion kinetics produces neutrons with higher average energies; see Figure 8.8. Some of these interactions with solid matter create He or H atoms directly inside lattices, embrittling the material: this was the reason for the limited life of fission nuclear thermal reactors tested in the 1960s and 1970s. The effect of high-energy neutrons on stainless steel, for instance that of the first wall, is to reduce
ductility to about 1% of the original after 2 years [Kulcinski and Conn, 1974]. This is the result of forming inside the steel about 1,000 atoms of helium and hydrogen per million structural atoms. Correspondingly, steel tends to swell, about 7% to 9%, if untreated. Apparently cold working the steel tends to reduce swelling to below 1%, but these figures are revealing.
Shielding technology has come a long way since the 1970s; there are new promising and light materials, based on carbon, for instance. However, traditional shielding still must rely on quantity of matter to stop radiation, and this adds mass to fusion engines and inevitably implies radiation damage. Figure 8.15 shows, from left to right, the layers of matter going outwards from the fusing plasma at the center of a tokamak torus [Kulcinski and Conn, 1974]. Although somewhat dated, the structure shown is realistic and may be divided into three main zones: the torus inside, the blanket and the shield. The magnetic coils form the reactor outside. The magnetic field permeating the torus keeps plasma 50 cm away from the solid first wall, in this example made of 0.4 cm thick stainless steel (S.S.). Ideally, nothing should exist between the edge of plasma and the first wall. Beyond the wall is the lithium blanket and its recirculating system, extracting most of the 14MeV neutrons thermalizing inside lithium, and providing most of the thermal power. Note that lithium contains a certain percentage of steel, since it is corrosive with most metals. Tritium is bred by neutrons deposited inside the lithium blanket and is extracted (in this particular scheme) by two independent circuits, so that one may be closed while the second is in service. A thermal insulation vacuum gap separates the blanket from the shield proper, made of boron carbide and lead. The carbide slows down and thermalizes neutrons that have not been stopped by the blanket, while lead absorbs gamma-rays. In this design helium is used to cool the shield assembly. A final vacuum gap insulates the reactor from the low-temperature superconducting magnet, made of NbTi and comprising a copper "lifesaver", in case the supercon-
ducting mode of operation ceased for any reason. The shield shown is designed for a 5-GW (thermal) fusion tokamak, and the blanket + shield structure is about 172 cm thick.
A conceptual way around the radiation problem is to look for a fusion kinetics that does not release neutrons, the particles more difficult to stop. Protons carry in the average the same momentum as neutrons, but their charge means they can interact with, and be stopped by, matter (or by an external electro-magnetic field) far more easily, requiring less shielding mass. The problem with this approach is that the energy yield of "aneutronic" kinetics is lower than for D-T; see Figure 8.11, and their ignition temperature even 10 times higher. Just as outlined in Section 7.4.1, the first task of a shielding design is to slow down and stop unwanted neutrons, not all of them if one wants to breed tritium.
The cooling system integral to a tokamak for industrial power generation constitutes also the heat exchanger extracting the fusion energy deposited in the coolant by high-energy particles, and thermalized as heat. In a fusion propulsion system utilizing electricity (to power electric thrusters, but also for other on-board tasks) it seems clear that such an extraction system must be more efficient and hopefully more compact than the conventional machinery of Rankine, Brayton or Stirling cycles of terrestrial power plants. For instance, direct conversion into electricity via thermionics, although a low (< 10-12%) efficiency process, is feasible, as may be other more speculative ways based on modern advances in electronics. A tokamak MCF configuration is thus naturally suited for the second type of propulsion strategy that is called electric fusion propulsion.
Although far from having been discussed to the extent deserved, the description of MCF mirror thrusters above suggests the MCF propulsion system is the better when choosing between thermal and electric. A tokamak MCF reactor coupled to an electrical generator, followed by an electric thruster would probably be a more controversial (albeit feasible) configuration. Just as commented in Chapter 7 when fission NEP was being discussed, a propulsion system configuration constituted by two separate energy and thrust generators does have its merits, the main one being that each component may be optimized to some extent independently. The drawback of fusion electric propulsion is that it must include machinery for energy conversion. Thermal energy must be converted into electricity, and at the current state of technology this may be done in the simplest and most reliable way only via a thermodynamic cycle. All thermodynamic energy conversion carries an efficiency penalty. Although combining two different cycles (Brayton and Rankine, for instance) may increase conversion efficiency by a few percentage points, combined power generation further complicates an already complex conversion scheme. In the end, the efficiency of conventional cycles reaches at most 50%. The remaining thermal energy can be used for other important tasks (radar, laser telecommunications, cryogenics are the ones that come to mind) but the greatest fraction would have to be rejected somehow to a lower temperature sink. Typical terrestrial sinks are rivers, or colder air. In space, that means space radiators, because no conduction or convection may take place. Space radiators add to total mass, having a weight/ power ratio of order 0.01 to 0.15kg/kW. At a conservative 0.1 kg/kW figure, radiator mass is 100 tons per each gigawatt of thermal power.
The electric power extracted at such high price can power an electric magneto-plasma-dynamic (or perhaps even ion) thruster capable of Isp in the 104 to 105m/s in the near- or mid-term (say, 10 to 20 years from now). MPD rockets are capable of higher Isp, but have lower thrust density compared to ion engines [Auweter-Kurtz and Kurtz, 2008]. The combined fusion power source and MPD rocket will be predictably a large assembly, as shown later by estimates of mass budgets. Besides, electric power switching and conditioning for GW-class thrusters operated at high currents or high voltage or both, would certainly be extraordinary technology challenges.
This said, fusion electric propulsion based on direct conversion (i.e., entirely bypassing thermodynamics) is a future development potentially impacting in a positive way on these considerations. Direct conversion has a relatively short history, and is limited to low power (< 1 kW) applications such as the RTG (Radioisotope Thermionic Generators) built for the Galileo and Cassini missions. RTG exploit the emission of charged particles from high-temperature solid materials to produce electrical power. Their efficiency is even less than thermodynamic conversion, being in the 10-15% range at best. Their major appeal is that they are static devices (no moving parts). The AMTEC technology described in Section 7.17 is a better option.
The most investigated type of direct conversion is that based on magneto-hydrodynamics, a technology for high power developed and tested for more than 20 years in the EU, the Soviet Union and the US [Messerle, 1995]. It consists of passing a ionized hot gas in a duct between a magnetic field B. If the B vector is normal to the gas velocity u, an electric field normal to both is generated by the motion of ions, and energy can be extracted. This class of generators is therefore the exact reverse of MPD electric thrusters described in Chapter 7. In MPD thrusters applying external E and B creates an accelerating Lorentz force F; in MHD generation, slowing down u in a field B creates an electric field E and thus a voltage.
MHD generation is inherently suited to extract energy from fusion, in that fusion products are a plasma. Any fusion kinetics producing few or no neutrons, e.g., reactions 6, 8 or 9 in Figure 8.8) would be ideal in this context. Handling such energetic particles in an MHD generator would be difficult, but the extraction process would be much more efficient than others based on any thermodynamic cycles or thermionics. MHD generation was abandoned in the mid-1980s mainly because of the difficult engineering problems posed by working with high-temperature ionized gas. This gas was at the time the hot exhaust products of coal burners, at temperatures of order 1,800 K. Since spontaneous ionization at this temperature was negligible (ionizing air nitrogen needs about 15 eV), the coal combustion products exhaust was seeded with alkaline metals (K, Ba, Na,...) that ionize much more easily, at energies of order 3-4 eV. These metals are extremely corrosive, and ruined MHD extraction duct sections very rapidly. Revisiting this technology is mandatory for direct conversion of heat into electricity; in fusion propulsion the question of ionization would no longer constitute a problem (rather, the high plasma energy would).
Are there new ideas in direct energy conversion? The answer is a qualified yes. Some are actually at the stage of just ideas. For instance, interesting work has been carried on since the 1980s in converting energy from radioactive decay of radio-nuclides producing alpha and beta particles into electricity, see [Brown, 1989]. This may seem identical to the RTG process, in which energy of alphas and betas is thermalized and the heat released produces electrons; in fact, this is not so. This novel concept is based on the fact that the energy of particles emitted by radio-nuclides also includes that of the electromagnetic field they generate because of their charge and motion. The fraction of energy in the form of electromagnetic field is much greater than that present as kinetic energy and captured by RTG. Time will tell whether these new concepts are indeed practicable in an engineering sense. Success in this area hinges on the chances of fusion propulsion to be investigated with significant resources. At the moment these are slight, but continuing interest by Japan in the GAMMA-10 mirror machine (at the Tsukuba research center), by Russia in the GOL-3 gas-dynamic mirror reactor at Novosibirsk and recent (2004) interest by ESA in fusion propulsion, e.g., see [Romanelli and Bruno, 2005], may be positive signs.
To conclude this section, at the stage of our knowledge far more work is needed to reach firm conclusions concerning the best solution to convert MCF thermal into electrical energy. By all reasoning made, an educated guess is that electric fusion propulsion is probably much more complicated than direct thermal fusion propulsion, although conceptually more flexible in terms of thrust and Isp modulation.
Was this article helpful?