This chapter discusses the potential proliferation risks associated with inertial fusion energy (IFE). Many modern nuclear weapons rely on a fusion stage as well as a fission stage, and there has been discussion of the potential for nuclear proliferation—particularly vertical proliferation1—in a country where an IFE power plant is sited.
The panel begins by providing some background on nuclear proliferation and inertial confinement fusion (ICF) and continues with discussions of several related topics: classification concerns, the relative proliferation risk associated with different target designs, weapons production in ICF facilities, knowledge transfer, other proliferation risks associated with ICF, and, finally, the importance of international engagement on this issue.
The term “nuclear proliferation” refers to the spread of nuclear weapons knowledge, technology, and materials to countries or organizations that did not previously have this capability. Proliferation has been of increasing concern in recent years, particularly following the successful detonation of a North Korean nuclear weapon, and the signals that Iran may also be pursuing an illicit nuclear weapons program. With the breakup of the Soviet Union, special nuclear material (SNM)
1 Vertical proliferation refers to the enhancement of a country’s capability to move from simple weapons to more sophisticated weapons.
became available at lightly guarded facilities; it is unclear how much was lost to theft, but proliferation concerns remain. Another concern arises from the many nuclear weapons in Pakistan, and whether they are controlled adequately.
Proliferation could occur in several ways: (1) the spread of knowledge about how to build nuclear weapons to other countries, (2) knowledge of—and access to—the physical technology used to construct nuclear weapons, (3) access to the materials from which a nuclear weapon could be constructed (e.g., SNM), and (4) access to people who have been engaged in nuclear weapons technology in other nations.
Because the first nuclear weapons were built using technology that was later adapted for use in civilian nuclear power plants and the civilian nuclear fuel cycle, the role that fission power could play in proliferation has been considered for decades. An international safeguards regime to detect attempts at proliferation is currently in place and operated by the International Atomic Energy Agency (IAEA). This regime, which is based on the Treaty on the Non-Proliferation of Nuclear Weapons (NPT), involves cooperation in developing nuclear energy while ensuring that nuclear power plants and fuel cycle facilities are used only for peaceful purposes.
The risk of nuclear proliferation could also be associated with ICF research facilities or, possibly in the future, IFE plants. For example, IFE plants and ICF research facilities provide an intense source of neutrons, which could, in principle, be used to generate 239Pu from 238U. In addition, information that could help countries develop more advanced boosted weapons or thermonuclear weapons could be gained from a thorough understanding of a fusion facility’s operation.
While the effect of a fission-only weapon can be devastating, the development of two-stage (both fission and fusion) thermonuclear weapons can provide much higher yield per weapon. By using an ICF facility to improve its understanding of the physics of fusion, a nation might glean information useful in transitioning its weapons program into a much more complex, modern, and threatening system. In fact, the U.S. research program in laboratory-based ICF has been largely funded by the nuclear weapons program, because valuable information can be learned from ICF that can otherwise be learned only from nuclear testing.2
Because IFE is still at an early stage as a potential energy source, international treaties related to nuclear weapons and proliferation do not clearly apply to IFE at this time. However, given the value of ICF to the U.S. nuclear weapons program
2 The moratorium on nuclear testing announced on October 2, 1992, by President George H.W. Bush and extended by the Clinton administration remains in effect. It was reinforced by the 1996 U.S. signing of the Comprehensive Nuclear Test Ban Treaty, which, however, has not been ratified by the U.S. Senate. The information gained by the nuclear weapons program is related to improving our understanding of weapons components built during the cold war, including the effects of aging on component performance.
and the programs of other nations, the applicability of some treaties to ICF has been considered.
The NPT does allow for laser fusion experiments, both in states that already have nuclear weapons and those that do not. As noted in 1998, this position is based on the unopposed, U.S. unilateral statement at the 1975 NPT Review Conference stating that “nuclear reactions initiated in millimeter-sized pellets of fissionable and or fusionable material by lasers or by energetic beams of particles, in which energy releases, while extremely rapid . . . are nondestructively contained within a suitable vessel . . . [do] not constitute a nuclear explosive device within the meaning of the NPT . . .” (U.S. DOE, 1995). Even so, the status of pulsed-power fusion experiments under the NPT remains unclear (Paine and Mckinzie, 1998).
In the 1990s, there was discussion in the United States about whether the Comprehensive Nuclear Test Ban Treaty (CTBT) also banned the use of ICF.3 Ultimately, the Clinton administration took the position that ICF is not a prohibited activity under the CTBT (Jones and von Hippel, 1998), and this position continues to be that of the Obama administration. However, some experts still debate the applicability of this treaty to ICF (Paine and McKinzie, 1998).
ICF research has received a great deal of specifically directed funding in the United States in recent years, even though IFE per se has not. This research is funded primarily through the U.S. nuclear weapons program, which envisions using ICF experiments and modeling as a method of verifying codes and calculations related to the current U.S. nuclear weapons stockpile. Because many of the topics involved in ICF are related in some way to nuclear weapons, much of the work is classified. The next section provides a brief introduction to the history and current status of the classification and declassification of various ICF concepts.
The primary reason stated by the U.S. government for classifying information related to ICF is to protect information relevant to the design of thermonuclear weapons. The possibility of using lasers to ignite fuel was first considered by the Atomic Energy Commission (AEC) and the national weapons laboratories in the early 1960s. At that time, concerns about the potential for laser fusion weapons as well as close ties between ICF concepts and nuclear weapons design (particularly physics and simulation codes) led the AEC to classify research on ICF. The first classification guidance for inertial confinement fusion information was issued in 1964. Initially, all aspects of ICF were considered to be classified.
3 It should be noted that the United States is not currently a party to the CTBT but as a signatory is bound not to act in violation of the fundamental restrictions of the CTBT.
Declassification of fusion concepts began slowly in the 1970s, and by August 1974, essentially all work with directly irradiated fusion targets was declassified. After a long pause, declassification began again in the late 1980s and continued through the early 1990s. Most notably, in late 1990, an Inertial Confinement Fusion Classification Review was requested by the Secretary of Energy with the intent of eliminating unnecessary restrictions on information relevant to the energy applications of inertial confinement fusion. The panel included representatives from the Department of Energy (DOE) national laboratories, the Department of State, the Arms Control and Disarmament Office, and other stakeholders, and the report was issued on March 19, 1991. The key panel recommendations included these: (1) “For laboratory capsules absorbing <10 MJ of energy and with maximum dimension <1 cm, all information should be declassified with some exemptions” and (2) “Some Centurion-Halite declassification would be desirable to gain the scientific credibility needed to advance the energy mission of ICF” (U.S. DOE, 2001). Later, on December 7, 1993, nearly all information on laboratory ICF experiments was declassified.4
At present, much of the information related to ICF targets has been declassified, with several notable exceptions. First, some aspects of computer codes and certain target designs remain classified, as well as the details of some historical experiments related to ICF (in particular, the Centurion-Halite program). Some aspects of classified targets are discussed in the classified Appendix F.
Whether or not aspects of ICF are classified is highly relevant to the future of IFE. If essential parts of an IFE plant are classified, this could create significant complexities for commercialization. Although some commercial facilities rely on classified concepts (such as those involved in the enrichment or reprocessing of nuclear fuel), there are likely to be export controls or specific regulations involved in dealing with this situation.
It is important to realize that classification or export controls could themselves indirectly cause proliferation risks if denial of information, technology, or materials causes some nations to mount covert programs or withdraw from the NPT.
There are four possible scenarios for future classification of IFE concepts. The first possibility is simple—the target will be classified or other key aspects of the concept will be classified. The second possibility is that the target is unclassified, but the expertise needed to make or assess it will involve classified information or codes. A third possibility is that other parts of the plant (e.g., lasers) will be considered to be dual use and subject to export controls. Any of these three outcomes could be very troublesome at a commercial plant. On the other hand, a fourth possibility is that the target and expertise will be unclassified, and none of the key elements of the plant are subject to export controls. If this is feasible, then
4 R. Johnson, LLNL, “The History of ICF Classification,” a document provided to the panel on February 24, 2011.
it would be the simplest configuration and a highly desirable goal for the future commercialization of IFE.
Any kind of ICF seeks to achieve thermonuclear ignition and burn. As noted previously, this goal relates ICF to thermonuclear weapons, and for this reason ICF (whether in a research facility or a power plant) is seen to pose some proliferation risk. However, this risk is mitigated by the fact that (1) nuclear weapons are much larger than ICF targets and (2) their operation presents some different engineering challenges.
Indirect-drive targets are associated with some proliferation concerns because the physics involved is more closely related to the physics associated with thermonuclear weapons than is the case with direct drive. In particular, the functioning of indirect-drive targets involves the use of X-rays in the hohlraum to drive the capsule implosion. ICF using indirect drive was declassified in 1991.
In any case, the processes involved in heavy-ion deposition (for heavy-ion-driven fusion) and the beam-plasma interactions that occur in direct-drive capsules are physically much more remote from conditions in existing thermonuclear weapons. In addition, these processes do not relate to any feasible design for a weapon that the panel is aware of. For these reasons, it is the judgment of the panel that heavy-ion fusion and direct-drive fusion pose (arguably) fewer proliferation concerns.
The Z-pinch fusion concept is likewise remote from existing weapons. However, during the cold war, the Soviet program in explosively driven magnetic implosion (MAGO) progressed further than any other approach to pure fusion, though like all such approaches, it was still very far from ignition (Garanin et al., 2006; Velikhov, 2008). Since the 1990s, Los Alamos National Laboratory and the All Russian Research Institute of Experimental Physics (VNIIEF) have carried out joint experiments on MAGO (Lindemuth et al., 1995).
In the future, as processing power for desktop and academic computers continues to increase, and as knowledge of plasma physics continues to accumulate in the open literature, many of these concerns may become less relevant, including the proliferation risk distinction between indirect drive and other forms of ICF that might be used for IFE. Enough physics knowledge may accumulate in the public arena that the use of indirect-drive IFE would not be able to add much to publicly available knowledge. In such a world, codes would be classified according to their direct use for (and calibration from) nuclear weapons, not according to the physics that they model. However, if an IFE plant were to rely on classified codes for target design or other operational aspects, and knowledge of these technologies could be used to gain information about the codes’ details, proliferation would be a concern.
CONCLUSION 3-1: At present, there are more proliferation concerns associated with indirect-drive targets than with direct-drive targets. However, the spread of technology around the world may eventually render these concerns moot. Remaining concerns are likely to focus on the use of classified codes for target design.
One of the key proliferation risks associated with any fusion plant (ICF or magnetic confinement fusion) is that it is possible to use the plant to create materials that are essential for the construction of nuclear weapons. These materials fall into two primary categories: special nuclear materials and tritium. Both types of material can be produced without the use of fusion facilities, but commercial fusion plants may be a more convenient source for these materials for those who cannot acquire them easily in another way. The potential for the production of each type of material is discussed next.
Special Nuclear Materials
As noted previously, it is technically possible to utilize the significant neutron flux emanating from a fusion reactor core to produce 239Pu from 238U. To accomplish this task covertly, it would be necessary to:
• Move quantities of uranium into the immediate vicinity of the fusion core and
• Acquire technology for—and construct—the appropriate reprocessing facilities to separate the plutonium from the uranium and fission products.
The first task is likely to be operationally cumbersome. In addition, the transfer of large quantities of uranium into and out of a fusion power plant would likely be detectable, because such conveyance would not be a normal operation for such a plant. The development and construction of a reprocessing facility—assuming that it had not already been built and brought into operation—would also be necessary. The technology is not new, but it requires significant radiation-handling capability. The construction and operation of such a facility would probably be detectable by the current safeguards regime.
Overall, the panel judges that the construction and diversion of an IFE plant in this fashion is not the simplest path for a host state to produce SNM. Research reactors and commercial nuclear plants capable of serving the same purpose (irradiation of uranium for plutonium production) exist in many nations. However, a previously built and operating fusion plant could serve as a path of opportunity for a nation interested in developing weapons. Such facilities may therefore have
to be subject to inspection to assure that they would not be so used, and to IAEA safeguards in states that do not already have nuclear weapons.
However, if terrorists were to seize an IFE plant, it could provide them with neutrons for the production of material to make a weapon of mass destruction. In this case, any facility capable of producing neutrons could be useful, but it is possible that no better solution would be available. Nonetheless, as noted above, an effective form of reprocessing would still be needed to isolate the plutonium.
For these reasons, the panel believes that a fusion plant raises fewer proliferation concerns than a fission plant with respect to the production of nuclear materials. However, in a region free of nuclear facilities, siting of a fusion plant could increase the proliferation risk in that region if the fusion plant were totally exempt from inspection by the IAEA or other international body. A hybrid fusion-fission plant would have the proliferation disadvantages and the economic problems of both technologies.
In order to fuel itself, a functioning IFE plant would likely be designed to continually breed a stream of tritium in vast amounts: about 60 kg per year for a plant of 1 GW (thermal) capacity. Tritium not only is an essential fuel for a fusion power plant, but it also can be used in part to fuel modern, boosted fission weapons or thermonuclear weapons.
The diversion of some portion of the substantial tritium stream would be relatively straightforward, but such diversion does not necessarily pose a significant proliferation threat per se. However, for a state already possessing nuclear weapons the diversion of only a few grams of tritium would be significant and would be difficult to detect. In addition, tritium can be produced in other ways if a state needs it. To date, tritium for nuclear weapons and other purposes has been produced using fission reactors.
With current technologies tritium alone, unlike SNM, cannot be used to build a nuclear weapon, and only a host state with relatively advanced capabilities would find such a stream of tritium to be useful. Indeed, for primitive nuclear weapons, tritium does not need to be used at all. However, if a significant diversion of tritium is observed, then it could be a signal to the international community that the host state is considering increasing its nuclear capability to include more advanced weapons using boosting or thermonuclear burn.
A second path for a potential proliferator might be the covert acquisition of key information about fusion, drawing on knowledge gained from operating a fusion
facility. This path is discussed separately for research facilities and energy facilities in the following sections.
Inertial Confinement Fusion Research Facilities
Research facilities—such as the National Ignition Facility (NIF)—pose different proliferation concerns than a fully functioning inertial fusion power plant, and the concerns associated with a host country misusing a research facility are likely to be greater than those associated with a fusion power plant. A fusion research facility is designed for the purpose of increasing physics understanding on a range of topics, not for a specific function (i.e., energy production). A power plant, however, is likely to be highly specialized and not designed with the flexibility inherent in a research machine. In addition, research facility diagnostics by their nature will provide hints about the underlying physics that power plant diagnostics may not.
If considered fully, the proliferation risk associated with a research facility can go beyond the physical presence of the facility in one nation or another. Research facilities may cater to a range of scientific interests beyond the needs of either the power generation community or the weapons community. For example, the NIF provides the plasma physics community with a highly effective experimental test and validation for a number of codes and theories that may indirectly or directly relate to the physics required for an understanding of thermonuclear weapons. Because the research community is intrinsically both open and international, such an improved understanding of plasma physics could provide a range of potentially useful information to a proliferator.
This increase in understanding is unlikely to stop, regardless of U.S. decisions. In the coming decades, both experiments and simulation in research facilities worldwide are likely to surpass current U.S. capabilities. For example, continuing increases in computing speed and understanding in the open research community could result in extremely capable physics codes.
However, it should be clear that information about physics is not the same as information about weapons design. For a nation that has never successfully (or unsuccessfully) detonated a thermonuclear weapon, no fusion research facility or power plant can adequately replace experimental physics and engineering knowledge gained from nuclear testing.
IFE Power Plants
An IFE power plant, as noted above, is unlikely to be highly flexible, and a research facility is likely to provide more information to a potential proliferator. By the time a design is commercialized, the physics will likely have been well understood (or engineered around), and the designs of the individual components will
have been optimized to the extent possible for power production. In addition, the diagnostics will be likely to be optimized for the needs of a power plant operator, not for the needs of a physicist attempting to learn useful weapons information.
However, knowledge transfer remains a concern if an IFE power plant is deployed overseas in a country where proliferation is a concern, because local expertise will be needed to operate the plant. The plant may not yield useful information about the physics involved in the reaction, but could provide information about energies needed and other technological details that must be known to obtain ignition in a fuel pellet. Moreover, personnel would gain practical experience in handling tritium. Whether this knowledge would be greater than that obtainable in the open literature is unclear.
CONCLUSION 3-2: The nuclear weapons proliferation risks associated with fusion power plants are real but are likely to be controllable. These risks fall into three categories:
• Knowledge transfer,
• SNM production, and
• Tritium diversion.
CONCLUSION 3-3: Research facilities are likely to be a greater proliferation concern than power plants. A working power plant is less flexible than a research facility, and it is likely to be more difficult to explore a range of physics problems with a power plant. However, domestic research facilities, which may have a mix of defense and scientific missions, are more complicated to put under international safeguards than commercial power plants. Furthermore, the issue of proliferation from research facilities will have to be dealt with long before proliferation from potential power plants becomes a concern.
One proliferation concern associated with ICF is the potential for the development of a laser fusion weapon, as discussed briefly in the section on classification earlier in this chapter. However, owing to the size, complexity, and energy requirements of existing or planned driver systems, the panel does not consider this to be a credible and immediate concern with respect to current concepts for inertial fusion energy, such as laser-driven fusion energy. However, in the distant future, advances in laser technology could change this picture.
In a 1998 declassification decision, DOE stated that “the U.S. does not have and is not developing a pure fusion weapon and no credible design for a pure fusion weapon resulted from the DOE investment” (U.S. DOE, 2001). According
to information released after the cold war, the Soviet experience was similar. However, this concern might someday materialize with currently unforeseen technology developments. For this reason and to alleviate any current concerns, it will be important to address the possibility (or impossibility) of pure fusion weapons in policy discussions and in the safeguards regime.
As described in the previous sections, there are proliferation risks associated with the use of ICF facilities around the world, and—should IFE concepts prove to be fruitful—with IFE plants themselves.
Managing proliferation, whether it is associated with fission concepts or fusion concepts, is intrinsically an international problem. While one country may not allow the export of certain technologies, other countries that do not consider the technology as sensitive may choose to allow it. In addition, the result of proliferation—the successful construction of a nuclear weapon by one more state—is international in its consequences.
For this reason, preventing proliferation associated with fusion energy requires international agreement on methods for managing the risks of the technologies involved, including safeguards. The IAEA defines the purpose of its safeguards system as follows:
to provide credible assurance to the international community that nuclear material and other specified items are not diverted from peaceful nuclear uses. Towards this end, the safeguards system consists of several, interrelated elements: (i) the Agency’s statutory authority to establish and administer safeguards; (ii) the rights and obligations assumed in safeguards agreements and additional protocols; and (iii) the technical measures implemented pursuant to those agreements. These, taken together, enable the Agency to independently verify the declarations made by States about their nuclear material and activities.
This safeguards system has been in place for decades to verify compliance with the NPT for fission plants and fuel cycle facilities around the world. If new facilities that also pose a proliferation risk—such as fusion facilities—were to be deployed around the world, it would be sensible to either include them in the current regime or to design a similar safeguards regime for them.
Of course, these safeguards would need to take into account the design of a particular fusion power plant. Although numerous design concepts have been advanced,5 the panel did not see any credible, complete power plant designs. This
5 See, for example, “OSIRIS and SOMBRERO Inertial Fusion Power Plant Designs,” DOE/ER-54100-1, March 1992, and “Inertial Fusion Energy Reactor Design Studies Prometheus-L and Prometheus-H,” DOE/ER-54101, March 1992.
has benefits, because it provides an opportunity to consider “safeguardability” directly in the initial design of a fusion power plant.
Early international discussions on this topic could be very helpful in reaching an international consensus on the key proliferation concerns associated with the use of inertial fusion power plants as well as how to manage these concerns (Goldston and Glaser, 2011).
CONCLUSION 3-4: It will be important to consider international engagement regarding the potential for proliferation associated with IFE power plants.
Proliferation is most tied to access to SNM, e.g., using enrichment processes. Richard Meserve6 recently wrote, “There is no proliferation risk from the [fission] reactors. Proliferation risks can arise from enrichment facilities because the technology could be used for weapons purposes” (Meserve, 2011). An advantage of fusion plants with respect to nonproliferation is that SNM will not be used in the plants and SNM will not be accessible from the waste products, as it is from fission plants. This lack of direct access to SNM is the major nonproliferation advantage of a fusion plant.
The disadvantage of inertial fusion power plants is that they allow access to knowledge and experience with fusion, which will necessarily increase with the design and operation of such plants. The latest nuclear weapons use fusion as a major source of the explosion energy. These concerns were outlined in one presentation by an official (Massard, 2010):
As an EU [European Union] requirement, we keep a clear separation between IFE and “sensitive” weapons science (nonproliferation)
• No use of weapons codes in the European programs
• No benchmarking of physics code with weapons code
• Not in favor of indirect drive capsule option in the European program for sensitivity issues
European countries have strong collaborations in ICF (e.g., HiPER). The French are building a laser fusion facility, LMJ, which is broadly similar to the NIF and which will be the most capable driver available in Europe. As a matter of policy, these programs will pursue indirect-drive ICF but do not intend to pursue indirect drive for IFE (Massard, 2010) because of the perceived proliferation risk. The
6 Former Chair of the U.S. Nuclear Regulatory Commission and chair of the IAEA safety advisory group.
United Kingdom participates in LMJ and HiPER and also actively participates at the NIF in the United States, and in the latter context is pursuing indirect-drive ICF.7
The Russian program in pure fusion evolved historically from the pre-1991 Soviet nuclear weapons program (Velikhov, 2008). Its major emphasis is on magnetic confinement fusion, which is not within the scope of this report. In ICF, two methods have received continuing attention in Russia: laser fusion and magnetized target fusion (MTF). Although research supporting ICF development is ongoing with smaller lasers (Kirillov et al., 2000; Belkov et al., 2010), Russia currently has no laser facility comparable to the NIF or LMJ8 and is unlikely to achieve laser-driven ignition in the near future. As for magnetized target fusion, the Russian MAGO concept has been widely advertised, and, as mentioned, joint work with LANL is ongoing. The proliferation risks of the MAGO MTF concept have been discussed in detail (Jones and von Hippel, 1998). Little concern about the potential for proliferation in MAGO is evident in Russian publications and policy. Indeed, in general, different countries have different classification policies.
7 J. Collier, UK Science and Technology Facilities Council, “Recent Activities and Plans in the EU and UK on Inertial Fusion Energy,” briefing to the NRC IFE Committee, June 15, 2011.
8 A news report in August 2011 suggests that plans for an NIF-class laser at VNIEFF are once again going forward, with commissioning expected in 2017; however, the stated purpose is stockpile stewardship, not ICF. See http://english.ruvr.ru/2011/09/30/57370758.html.