- American Physical Society Sites
- Meetings & Events
- Policy & Advocacy
- Careers In Physics
- About APS
- Become a Member
For over a decade, I have advocated that current research efforts directed at developing fusion reactors shift their focus from pure fusion to a design known as fission-suppressed hybrid fusion [1-5]. The easiest fusion reaction combines a deuteron and a triton to produce a neutron of kinetic energy 14 MeV and an alpha particle of energy 3.5 MeV. However, because of the Coulomb repulsion between the deuteron and the triton, this can only be achieved in a hot, dense plasma. Production of electric power with pure fusion uses the neutron’s kinetic energy to boil water. Fission-suppressed hybrid fusion follows all the same steps, but also uses the neutrons to breed ten times more fuel, to be burned elsewhere in conventional fission reactors. My reasons for advocating this innovation are twofold: first, I believe that the world will need 10-30 terawatts (TW) of additional carbon- free power by mid-century, and secondly I feel that the progress of pure fusion is too slow to meet this need . Fission-suppressed hybrid fusion (also called fusion breeding), just might, since it makes many fewer demands on the fusion reactor, and it also fits well into today’s nuclear infrastructure. References 1-5 spell out in much more detail development paths which might make this a reality by mid century or later. They also give very rough cost estimates based on International Tokamak Experimental Reactor (ITER) costs. In this article I review the possibilities for various approaches to meeting world energy demand, the status of current fusion programs, describe the fission-suppressed hybrid fusion concept, and offer a proposal for how an ‘energy park’ comprising both light-water reactors (LWRs) and a fusion reactor could be configured.
World energy demand
Let us first consider how one might achieve 10-30 TW of carbon free energy by midcentury. The options are few. Unfortunately there is no risk-free, universally agreed upon approach. However, many argue that any solution must have a very large nuclear component. I concur. But like any other energy option, nuclear has its own set of issues including fuel supply, proliferation and disposal of spent fuel, the subject of this and my earlier work [1-5]. I, and many others had assumed that nuclear energy’s nearly impeccable safety record in the west over the last 30 years, as well as a new generation of even safer reactors, had put the issue of reactor safety to rest. With the recent disaster in Japan involving the Fukushima nuclear complex, the entire issue of reactor safety must be reexamined. But nuclear power certainly cannot be simply abandoned either, any more than oil can because of the Gulf BP spill, or coal can because of innumerable coal mining disasters.
Today, once-through nuclear produces about 350 GWe, or about 1 TWth with light water reactors (LWR’s). Freidberg and Kadak make the case that LWRs are so well established that they will be the nuclear reactor of choice for quite some time . They estimate that there is about 500-1000 TWyrs of uranium fuel: “For the foreseeable future, the most economical way to obtain fuel for LWRs is to dig it out of the ground.” Of course, this depends on how far ahead one can foresee, how long it takes to develop alternatives, and how accurate the estimate is. For instance, Hoffert, et al. estimate a fuel resource of 60-300 Twyrs . It is far beyond the scope of this work to sort through the conflicting claims as to economically available uranium ore, but there is no dispute about one thing: a once-through nuclear economy based on LWRs uses only about 1% of the available fuel. Fission or fusion breeders potentially use all of it. This author does not believe the world is so well endowed with fuel that we can afford to discard 99% of it.
While conventional ‘renewable’ sources (solar, wind and biofuel) can play some role, they can never produce power on the necessary scale. The world will learn this, but with so much hoopla and government support, it will learn the hard way. Many European countries, including Spain and Denmark, which mandated large subsidies for renewable energy, are backing off as the cost becomes apparent. These costs are not difficult to estimate. Wind and solar currently receive government subsidies of about $0.1/kwhr . Without it, large parts of the industry would simply collapse, now and for the foreseeable future. In the unlikely event that it could even be done, the complete transition in the USA (0.5TWe) to renewable sources would cost the government about $500 billion per year — real money. Realistically, the only options for providing truly sustainable carbon-free power on the scale required at any time in the foreseeable future are fission and fusion breeding (and possibly pure fusion many decades later). While fission breeding is much closer at hand, fusion breeding has the potentially overwhelming advantage of being about an order of magnitude more prolific as a fuel producer.
Recently, hybrid fusion has been receiving more attention [6-9]. Freidberg and Kadak have recently summarized the situation, looking into subcritical nuclear reactors, where the fission and fusion reactors are in a single reactor (also called fast fission), the use of fusion neutrons to burn actinide nuclear waste, and the fission suppressed option described below . (Actinides are the elements like plutonium which are beyond uranium in the periodic table.) Fast fission would involve a much more complicated and dangerous reactor, essentially a fusion reactor of perhaps several hundred megawatts (MW) deep inside a fission reactor of perhaps 3 gigawatts (GW). Such a design would not fit readily into the current nuclear infrastructure. Fission suppressed hybrid fusion is hardly a new idea: it was first proposed by Andrei Sakharov in 1950, and Hans Bethe advocated it in 1979 [10, 11]. Yet despite this pedigree, the idea has never really caught on. I have long argued that now is the time to reconsider it; indeed, I consider it to be the only viable hybrid fusion option. But first, let us examine the situation with pure-fusion research.
Pure fusion: The current situation and the long road ahead
Fusion research has been supported worldwide with billions of dollars for over 50 years now. For most of this time, the predictions of a single demonstration reactor have been 35 years in the future. It is at least that far off today. Never have generations of scientists worked so long on such a tough problem; never have generations of sponsors been so patient.
Fusion research has indeed made enormous progress [1- 5] and is now is concentrated at two facilities, the ITER in France, and the National Ignition Facility (NIF) at the Lawrence Livermore National Laboratory (LLNL) in California. Each is gigantic in size, is years behind schedule, and is billions of dollars more costly than original estimates. ITER was first proposed in 1985, approved in 2005, and should be constructed by 2020. It is a tokamak, which utilizes magnetic confinement of the plasma. NIF is a Megajoule laser which took over a decade to design and build. Fusion devices require input power, typically neutral beams or microwaves for magnetic fusion, or lasers for inertial fusion. At breakeven the ratio of fusion power to driver power, known as Q, is unity.
A tokamak contains a plasma confined by both a large toroidal magnetic field and a smaller poloidal field. Laser fusion works by irradiating a target with intense radiation. As the outer part of a millimeter sized target ablates, an opposite inward force compresses the core. While the compressed target is transparent to fusion neutrons, the alphas are locally absorbed, producing a propagating fusion burn wave as the target flies apart. This is called ignition. If all goes well, ITER could give breakeven in about 2025, and perhaps achieve a planned Q of 10 some years later. NIF could achieve ignition in 2-3 years. But as typical for any power plant, the efficiency of the conversion of thermal (i.e., fusion neutron) power to electrical power is itself of order 1/3. Beyond that, the efficiency of the conversion of this electrical power to microwaves or neutral beams is itself about 1/3, so any magnetic fusion device needs a Q of at least 10 before it can do any more than power itself. Laser efficiency is much lower, so a laser fusion power plant would need a still higher Q.
Tokamaks have at least two difficult problems to overcome, which, while likely solvable, are still outstanding after 50 years of research. First, they are inherently pulsed devices. Their current is driven by discharging an inductor, which can store only so much energy. However, they must be run in a steady state. Perhaps more important, tokamaks frequently disrupt; that is, they can suddenly release all or a large part of the plasma and poloidal magnetic field energy. Up to now tokamaks have stored only about 10 megajoules; ITER will store about 800 megajoules, about the energy of a 400 pound bomb. If this energy is suddenly released in a major disruption, it is not unreasonable to fear that the superconducting toroidal field coils might suddenly and uncontrollably quench, as has happened already once in the Large Hadron Collider in CERN. If the toroidal field energy is released in a major disruption, this 400 pound bomb becomes 4000 pound bomb. Clearly, for pure fusion or fusion breeding operation, ITER must be designed so as not to allow even a single major disruption. For either a fast fusion configuration or an actinide burning configuration, the tokamak would be surrounded by a ton or so of plutonium, a requirement that makes lack of disruptions all the more pressing a requirement.
As regards lasers, the construction of NIF is now complete and the ignition campaign is underway. However, NIF’s role is nuclear weapons research. LLNL has chosen glass lasers, which have no average power capability; their sponsor does not require any. If they wish to extrapolate to a power plant they would need a completely different laser system. Furthermore, consistent with their goal of weapons research, their target chamber is hard wired into a configuration called indirect drive; symmetric laser illumination of a spherical target is not readily possible in NIF. The target is in a hohlraum, which is itself irradiated by the laser (a hohlraum is a cavity whose walls are in radiative equilibrium with the radiant energy within the cavity). The walls are heated to 250-300 eV and black body radiate. This radiation symmetrically illuminates and implodes the target. Calculations show that gain is sacrificed with indirect drive. But even granting a gain of 100, a NIF pulse would produce 100 megajoules of neutron energy, or about 33 megajoules of electric energy, i.e. 9 kilowatt hours, worth about a dollar. Current hohlraums cost about $10,000 and use a great deal of expensive material such as gold. Once LLNL completes its first ignition campaign, if it can get support for an energy program, perhaps it could reconfigure its optics and target chamber so symmetric illumination is possible. While certainly a costly and time consuming effort, this would be the fastest and cheapest route to megajoule direct drive experiments. Also, the NIF laser has an efficiency of order 1%. LLNL has proposed solid state diode-pumped lasers and the Naval Research Laboratory (NRL) has proposed KrF lasers, both of which could have sufficient average power and efficiency. Possibly one or the other could be built on the scale of NIF in a decade or so.
Figure 1: Schematic of an energy park: Inside a low security fence (A), five 900 M-We light water reactors (B), electricity going out (C), hydrogen and/or liquid fuel pipeline (D), cooling pools for radiation products (E), hydrogen and/or liquid fuel factory (F). Inside a high security fence (G), unburned or undiluted actinides; the separation plant (H), the actinide burner (I), and the fusion reactor (J).
Clearly, pure fusion has a very long and difficult road ahead; it has no chance of producing large scale, economical power by mid-century; its most optimistic advocates admit this. This author even has asserted that because of inherent limits on density, pressure and current in tokamaks, it is doubtful that they will ever be economical pure fusion devices, but could well be economical hybrid fusion fuel suppliers .
Fission suppressed hybrid fusion: A possible short cut
Today’s nuclear infrastructure is based on LWRs, and this will probably be the case in mid-century as well . Thus it is important that fission suppressed hybrid fusion fits in as readily as possible into current nuclear technology. The central idea of fission-suppressed hybrid fusion is to use the fusion-created neutrons to breed nuclear fuel while minimizing fission reactions in the fertile blanket. This means that the blanket must be, for instance, a liquid, perhaps a molten salt with the fertile material dissolved in it. The fissile material produced is continuously removed.
As to the fuel, there are two alternatives: to breed Pu-239 from U-238, or to breed U-233 from Thorium. The proliferation risks associated with the use of plutonium in the raw fuel likely dictate the thorium cycle [1-5]. The fertile material, in which the fissile material is dissolved, could be either U-238, Th-232, or a mixture. There are advantages and drawbacks to each. The former raw fuel would be much more proliferation resistant, while the latter produces many fewer actinides.
Through Monte-Carlo calculations, one can determine the ultimate fate of a fusion neutron in blankets of various materials and geometries. In one fission-suppressed configuration, Moir has calculated that each fusion neutron produces about 0.6 U-233 atoms, the triton sustaining the fusion reaction, and about eight additional MeV for a total of about 24 MeV, the breeding reactions being exothermic . However, when this U-233 is burned in fission reactor it releases about 200 MeV, so in this particular case a single fusion reactor can fuel about 5 LWR’s of equal power. In contrast, it takes two fission breeders to fuel one LWR of equal power. This is the tremendous advantage of fusion breeding over fission breeding. Fusion breeding has an additional advantage as well. A fission breeder needs a large amount of fissile material to start up, whereas a fusion breeder requires none. This is why, in this author’s opinion, it is essential to attempt to develop fusion breeding even though fission breeding is presently much closer to reality.
To see the enormous potential of fusion breeding, consider the originally proposed ITER (Q=10, 1.5 GW of fusion power) as a potential commercial reactor. Assume both electricity and the fusion driver are produced with an efficiency of 1/3. The 1.5 GW of fusion power then produces about 500 MW of electric power. However the microwaves or beams needed to drive the reactor take 150 MW (recall Q=10), that is 450 MW of the raw electric power, leaving all of 50 MW for the grid. But now consider the same reactor, but run as a fusion breeder. The output power is now increased by the breeding to 2.4 GW, or 800 MWe. But it would also produce about 13 GW of nuclear fuel, enough for five 900 MWe conventional reactors. Now 5 GWe goes out to the grid, an increase of two orders of magnitude over pure fusion. Thus, instead of being a stepping stone to who knows what sort of demonstration, decades and decades after completion of ITER, an ITER-sized reactor could be an end in itself.
As another example, now consider an ITER-sized reactor driven by a 1 MJ laser. Take the laser efficiency to be 0.1 and Q = 100 and a repetition rate of 15 Hz. At an electrical generation efficiency of 1/3, it sends 350 MW to the grid. This is possibly economical, but in a system of its size is more likely marginal. However, run in a fission suppressed hybrid mode, it would also produce 4.5 GWe of nuclear fuel, an order of magnitude increase. But suppose technical development stops short, so that laser efficiency is 0.05 and Q = 50. Then in a pure fusion mode there would be no power for the grid, but in a fission suppressed hybrid mode, it would still produce 2.2 GWe of nuclear fuel.
Obviously, the demands of fission suppressed hybrid fusion on the fusion reactor are much less than for pure fusion; so much so that fusion breeding even has a reasonable chance of supplying large scale power by mid century [1-5].
The energy park
One possible sustainable model for world development is “the energy park”, introduced by the author and shown schematically in Fig (1) [3-5]. The concept is preliminary, and is introduced as what might be described as more than a dream but much less than a careful plan. In it, a single ITER sized, 800MWe fission suppressed hybrid fusion reactor (“J” in the Figure) fuels five LWR’s of about 900 MW each (“B”). The spent fuel is reprocessed and is separated into three categories, fission products such as cobalt 60, the original fertile material (say U-238), and actinides such as plutonium. The energy park would store the fission products, which would have half-lives of about 30 years, in cooling pools or dry cask storage on site until they became inert, perhaps after 300-600 years. This is a time human society can reasonably plan for. The U-238 is recycled.
Because of its 24,000-year half-life, plutonium will always be a significant proliferation risk, so geological repositories such as Yucca Mountain must be extremely secure and certified for hundreds of thousands of years. This is difficult to do with any credibility. Instead, in the energy park, the plutonium and other actinides are burned in a sixth reactor (“I” in the Figure), a fast neutron reactor, such as an integral fast reactor (IFR), but run with as low a conversion ratio as possible. (The conversion ratio is the ratio of bred fuel to input fuel; above unity it is a breeder, below it is a burner.) The IFR can run with any transuranics as fuel and can be run either as a breeder or burner. Furthermore, it has demonstrated passive safety [13,14]. That is when a component fails, the reactor, without human intervention safely shuts off. Once burned, the ashes would have no proliferation risk. Hence the need for geological disposal would be eliminated or greatly reduced.
Energy parks could provide economical, carbon free power, with no long-lived waste and no proliferation risk. They could sustain the world at 30 TW at least as far into the future as the dawn of civilization was in the past.
1. W. Manheimer, Fusion Technology, 36, 1 (1999).
2. M. Hoffert et al, Science, 298, 981 (2002).
3. W. Manheimer, Journal of Fusion Energy, 4, 131 (2001; published 2003), 23, 223 (2004), and 25, 121 (2006).
4. W. Manheimer, J. of Fusion Energy, 28, 60 (2009).
5. R. Moir and W. Manheimer, Hybrid Fusion Fission, chapter 14, Fusion Technology, T. Dolan, Springer Verlag, to be published 2011.
6. J. Freidberg and A Kadak, Nature Physics, 5, 370 (2009).
7. W. Stacey et al, Nucl. Technology, 162, 53 (2008).
8. M. Kotschenreuther et al, Fusion Eng. Design, 84, 83 (2009).
9. E. Moses et al, Fusion Science and Technology, 56, 547 (2009).
10. A. Sakharov, Memoirs, Vintage Books, 1990, p 143.
11. H. Bethe, Physics Today, May 1979.
12. R. Moir, Fusion Technology, 8, 465 (1985).
13. Y. Chang, Physics & Society, April, 2002.
14. W. Hannum, G. Marsh and G. Stanford, Physics & Society, July 2004.
Retired from the U.S. Naval Research Laboratory