Dangerous Thermonuclear Quest

One of the central military and disarmament issues facing the international community today is to decide whether pursuit of research whose aim it is to achieve pure fusion explosions in the laboratory is compatible with disarmament goals and treaties, including, most importantly, the Comprehensive Test Ban Treaty. The development of pure fusion weapons is now a distinct possibility, though it is not a certainty, since their scientific feasibility remains to be established. One central challenge to disarmament and non-proliferation today is that the scientific feasibility of such weapons could be established using the same devices that are being promoted as essential for the ratification of the Comprehensive Test Ban Treaty (the treaty has been signed by about 150 countries, with the notable exceptions of India and Pakistan, since September 1996). The nuclear weapons powers, notably the United States and France, have programs for the "stewardship" of their existing stockpiles of nuclear weapons. As part of their stewardship programs they are building or operating facilities that will be used to maintain the skills of nuclear weapons designers, and which could be sued to develop a qualitatively different class of nuclear weapons. ICF facilities and research are an important part of these programs. Since its May 11 nuclear tests, India has also announced its own stockpile stewardship program.

The stated goals of the US stockpile stewardship program are to maintain the safety and reliability of existing weapons. We have shown in a previous report that most of the US program of SBSS is marginal or irrelevant to nuclear safety.¹We have also argued that fusion facilities such as NIF and the proposed X-1, are not relevant to maintaining the reliability of current nuclear weapons, particularly if the United States were to adopt a nuclear policy based upon deterrence rather than first-strike. The evidence for this conclusion is summarized below.

Pursuit of programs with explicit potential for designing new nuclear weapons is counter to Article VI of the NPT and to the CTBT. This applies whether the new weapons follow on current generation fission-triggered weapons or are part of an entirely new class of weapons, such as pure fusion weapons. In this context, it is worthwhile to recall that Article VI of the NPT relates, among other things, to the "cessation of the nuclear arms race at an early date."

ICF researchers claim that their research could also lead to commercial power production from fuels that are widely available and plentiful. However, the energy applications of any explosive fusion research should be justified on their own merits and in comparison to other energy projects. Many environmentally sound energy technologies are much further ahead than ECF and yet receive far fewer resources. Further, ECF approaches will take decades to develop into economical energy sources, if they prove feasible at all. The fact that large resources have been spent over decades on fusion power research without even establishing scientific feasibility needs to be more carefully considered, given the urgency of reducing greenhouse gas emissions. Military rationalizations and the relatively great pull of nuclear bureaucracies on governmental energy programs seem to be the forces driving ECF programs rather than serious evaluations of the world's energy and environment needs.

There is no question that NIF, pulsed power devices like the wire array z-pinch, and MTF have nuclear weapons design as one of their goals, as the following two quotes from a 1986 National Research Council Report Illustrate.

Weapons physics and nuclear weapons design is still a goal of the ICF program, now under the rubric of the Stockpile Stewardship program. Allowing nuclear weapons designers to gain greater experience in design is one of the goals that has been declared necessary for the stewardship of existing weapons. While the weapons design goals that have been announced relate to fission-triggered warheads, the same research will also advance the establishment of the scientific feasibility of pure fusion weapons. That goal has not been announced, as it would be provocative to do so and would make international opposition far more likely and the pursuit of pure fusion weapons research far more difficult.

The legality of ICF, wire-array z-pinch, and MTF programs under the Comprehensive Test Ban Treaty remains in question. The DOE has determined that the Stockpile Stewardship program in general and NIF in particular are not proliferation risks. The JASON committee, which evaluated the stockpile stewardship program for DOE, concluded that NIF is "an extremely sophisticated challenge, not one which could conceivably be undertaken by, or be useful to, a potential proliferator," especially since the basics of simple nuclear weapons designs are already widely available.⁴ The JASON report also concluded that the SBSS program would actually contribute to the goals of non-proliferation by allowing the United States to sign the CTBT.⁵ A specific review of NIF done for the DOE by the Department's Office of Arms Control and Nonproliferation reached similar conclusions.⁶ A report done by Dr. Ray Kidder, a former nuclear weapons designer and one of the originators of the ICF program at Lawrence Livermore National Laboratory, for the Arms Control and Disarmament Agency concluded that an ICF research team constituted a de facto weapons research team and that information would come from the following three sources:

• A technical library with information on the basic science necessary for designing nuclear weapons. This would be the most important source of information;

• Publication of ICF research from non-nuclear weapon states with advanced ICF research projects;

• The ICF programs of the nuclear weapons states. This would be the least useful due to the classification of large portions of ICF research in the nuclear weapons states.⁷

Kidder concluded that information relevant to nuclear weapons design would become public. However, he concludes that projects such as NIF don't form a proliferation risk because they are not replacements for full-scale nuclear tests. He also concludes that existence of an ICF research team in a non-nuclear weapon state would increase the readiness of that state to design nuclear weapons, but would not "represent nuclear weapons proliferation per se."⁸ However, it should be noted that all of these conclusions, whether one agrees with them or not, were made in the context of present generation thermonuclear weapons and did not take into account the relevance of ICF research to advanced weapons, such as pure fusion weapons.

Though significant technological hurdles to successful ICF development still exist ICF and other fusion programs pose a number of important proliferation problems that deserve far more public debate than they have received:

• In the short-term, inertial confinement fusion programs can be used to develop new thermonuclear weapons with fission triggers, thereby undermining the spirit of the Comprehensive Test Ban Treaty and Article VI of the Nuclear Non-Proliferation Treaty which commits the recognized nuclear weapons states to good faith efforts to end the nuclear arms race and towards disarmament.

• Fusion power for commercial applications is likely to be more technically feasible and more economical if it is combined with fission power development and the production of plutonium. Thus, the development of fusion power technologies could, in the long-term, provide new arguments for creating an infrastructure for plutonium production, processing and use.

• The achievement of ignition in ICF and similar devices is likely to result in the injection of even larger amounts of money into technological improvement of ICF technology, making pure fusion weapons more feasible.

A. The Science Based Stockpile Stewardship Program

The Science Based Stockpile Stewardship (SBSS) program is a multi-billion dollar effort encompassing a variety of facilities and sites, including all three weapons laboratories, and the Nevada Test Site. Facilities will be built or upgraded to allow weapons physicists to study all stages of a nuclear explosion, as well as providing the capabilities to create realistic 3-D models of weapons through the Accelerated Strategic Computing Initiative (ASCI). One of the main stated objectives for the SBSS program is to maintain the safety and reliability of the existing arsenal as the weapons age. This would be accomplished by developing a complete understanding of the physics involved in thermonuclear explosions and the modeling of weapons. (Another is to maintain weapons design teams and give them interesting work to do.)

A detailed examination of the safety and reliability justification for SBSS, based upon DOE's historical data concerning problems found with warheads in the arsenal, can be found in Zerriffi and Makhijani 1996. In the context of the present discussion, it suffices to note that fusion facilities such as NIF play no role in maintaining the safety of aging weapons. Nuclear weapons safety is an issue which affects the primary of the warhead (specifically, preventing accidental detonation of the primary). Fusion reactions (whether they be D-T fusion in the boosted primary or in the secondary) do not occur until after the fission detonation has already occurred. Safety is, at that stage, a moot point. Furthermore, DOE's own data has shown that aging has not affected the safety of a single nuclear component (either primaries and secondaries) in the entire history of the weapons program. Aging can affect the safety of some non-nuclear components, but this is a separate issue. The National Ignition Facility, the MTF program, the wire-array z-pinch are also all irrelevant to these non-nuclear safety issues. This leaves the issue of warhead reliability.

1. Reliability

The DOE's reliability justification for the SBSS program is problematic on three counts:

the DOE's definition of reliability
the expectation of future reliability problems
the relevance of NIF to addressing reliability problems of stockpiled weapons.

We will discuss these three issues in turn.

a. Reliability definition

DOE considers a warhead unreliable if it does not explode at its stated yield and at the correct target parameters (e.g. burst height). This definition of reliability is only necessary if the stated objective is to eliminate the hardened silos containing an adversary's nuclear weapons. Moreover, the DOE also considers a warhead unreliable even if the yield is above the design rated value, but if the accuracy is less than the warhead's specifications. In brief, the DOE's definition of reliability corresponds to a nuclear force maintained for the purpose of counterforce strikes, first use, and nuclear war-fighting capability.

If the objective is simply to deter a nuclear attack, then it is reasonable to assume that the belief on the part of a potential adversary that US warheads would be used in retaliation for an attack and that they would perform reasonably well would be sufficient. Declines in primary yield up to a certain point would be unlikely to cause the failure of the secondary to detonate. Hence, any overall performance decreases (resulting from lowered primary yield which is still sufficient to detonate the secondary) would be irrelevant to second strike deterrence.

This leaves the issue of threshold effects that may cause the failure of the secondary if the yield of the primary drops below a certain level. Adopting this level as a criterion, coupled with a substantial relaxation of accuracy requirements, would provide a different approach to reliability that would not be so clearly linked to counterforce doctrine. This would be more than sufficient for second-strike deterrence. Even in the case of the failure of the secondary and the primary fusion booster, the fission portion of the primary would still provide a huge explosion, estimated to be between several hundred tons and a few kilotons of TNT equivalent.⁹ Such explosions, while far smaller than design basis explosions of several hundred kilotons typical for strategic warheads, would be hundreds of times larger than the terrorist bomb that destroyed the Alfred P. Murrah Federal Building in Oklahoma City. Thus, second-strike deterrence does not actually require consistent secondary detonation or even the consistent functioning of the booster in the primary of a nuclear warhead. No adversary of the US would strike first under the assumption that the secondaries of the weapons used for retaliation would be less likely to go off.

Another major issue is the US nuclear weapons posture in light of its obligations under Article VI of the NPT. In its advisory opinion on nuclear weapons and war, the World Court unanimously decided that this article of the NPT required nuclear weapons states parties to actually achieve complete nuclear disarmament.¹⁰ One reasonable way to approach this goal so far as nuclear weapons reliability is concerned would be to remove permanently from the stockpile weapons that are deemed unsafe or unreliable.

Finally, we should also note the destabilizing effects of pursuing the SBSS program with a counterforce reliability definition. A counterforce definition of reliability would be dangerous at any time, but is especially so in a time when command and control in Russia are thought to be deteriorating. Fear of a first strike is a central reason for the US and Russia to keep their nuclear forces on hair-trigger alert. This launch on warning posture (commonly called a "use-it-or-lose-it" policy) is highly dangerous because it could result in large-scale accidental nuclear war.¹¹ The United States, in its own self-interest should abandon a policy of that increases fears in Russia of a possible first strike because it would reduce the incentive for Russia to keep is forces on hair-trigger alert. DOE's definition of reliability sends the contrary message and hence increases nuclear dangers.

b. Future reliability problems

The historical data analyzed by IEER in its report The Nuclear Safety Smokescreen¹²indicate that aging-related reliability problems do not appear to be significant when it comes to nuclear components, particularly secondaries. Of the 186 different types of reliability problems found with the arsenal, only eight affect secondaries. However, even that overstates the problem. Seven of those eight were actually operation (performance/yield) problems. These problems are not related to whether or not the weapon will explode, simply whether it will explode at its rated yield. This returns to the question of the definition of reliability used by DOE. It should be noted that only one reliability problem type affecting secondaries was related to aging. Therefore, since the early 1950's when the Stockpile Evaluation Program was started, there has been only one type of aging-related reliability problem found. What should be more relevant to the DOE is the larger number of aging-related reliability problems found with non-nuclear components, components which can be replaced and/or redesigned and tested separately from the rest of the warhead. NIF would have no relevance to these problems.

c. Relevance of NIF to reliability of the current stockpile

Even if reliability problems do occur in secondaries and they are deemed important enough for action to be taken, it has not been demonstrated that fusion facilities such as NIF would be of material benefit. The most prudent approach in such a case would be to either replace the secondary with a spare or to remanufacture the component. Remanufacture as a means of maintaining the stockpile has been put forward by a number of experts in the field as the proper means to achieve the goal. Furthermore, using NIF to fix problems with the secondaries could actually result in more problems, and perhaps create a push for resumption of nuclear testing. If a problem were found with a secondary and NIF experiments were designed to study aspects of the problem, it would result in modification of the computer codes used to model the weapons. Currently, the computer codes have been validated by comparison to data from underground tests. It is possible that, as the codes are modified to reflect experimental data from NIF (or other facilities), they will deviate more and more from the phenomena occurring in the weapons, which are different and, in some ways, more complex. In particular, NIF fuel pellets will be driven by lasers, while nuclear weapon secondaries are driven by the nuclear explosions in the primaries of the warheads. While the geometries of indirect-drive ICF and thermonuclear weapons are similar, and both are compressed by x-rays, there remain differences which need to be accounted for in transferring knowledge about one to the other. These differences, such as the geometry of the devices and the use of uranium casings in weapons, will likely create uncertainties that could lead to a push to re-validate the revised computer codes with new underground tests. This would require a US withdrawal from the CTBT - a step that would be likely to have major adverse proliferation consequences. It is crucial that the ban on testing be maintained by CTBT signatories and extended to all other states. Factors that would aggravate the risk of a CTBT breakdown should be eliminated as far as possible. Stopping the construction of NIF would be one step in that direction.

2. The US laser fusion program as a weapons development program

The US laser fusion program has traditionally been a component of the weapons development program. Prior to the cessation of nuclear testing, NOVA and other facilities of a similar nature were used to explore the same physical processes as NIF, albeit at lower pressures, temperatures, and energies. This information was used in the weapons design process; final proof of designs slated to enter the US nuclear arsenal was always through nuclear testing.

Official DOE statements on the SBSS program and the National Ignition Facility demonstrate the inconsistencies inherent in the program. On one hand, there are many statements that refer to the necessity of underground testing for developing new weapons, thereby implying that new designs cannot now be developed in the context of a CTBT. DOE also point to the fact that there are no requests from the DOD for new weapons designs as "proof" that SBSS is not aimed at weapons design. On the other hand, one of the stated purposes of the SBSS program is to attract and retain weapons designers and to provide them with the opportunity to practice their skills. According to the DOE, this is necessary in the eventuality that requirements for new nuclear weapons (presumably originating from the military) develop.

The DOE's Stockpile Stewardship and Management Plan, also known as the "Green Book," provides a good indication of the view DOE takes towards future weapons design work. It states:

The National Ignition Facility is supposed to play a key role in maintaining the design expertise for new weapons. This desire to maintain and exercise design capabilities is not an abstract conjecture. The Green Book also discusses two replacement nuclear designs, one of which would require fabrication of new plutonium pits. It would not, however, require nuclear testing:

In view of the statements, it is quite conceivable that DOE weapons scientists would conduct at least preliminary design investigations of pure fusion weapons if and when the necessary data become available. DOE's rationale allows nuclear weapons scientists the opportunity to practice their design skills. The design of pure fusion weapons would fit in with this DOE policy.

3. Other countries

The United States is not the only country with an active research program into ICF or MTF. All five of the declared nuclear weapons states have facilities for conducting experiments in ICF. Other countries also have ICF research facilities. A half dozen additional countries have either laser or particle beam facilities that are, at least partly, devoted to the study of ICF. In some cases, these facilities are fairly small, such as the single beam facilities in India, South Korea, and Israel. (Nonetheless, India and Israel have probably used the research done at these facilities to design their nuclear weapons.) However, other operating or planned facilities rival the capabilities present and planned for the US ICF program. In France, a facility very similar to NIF is planned. Laser Mégajoule would be built near Bordeaux, and like NIF in the US, it is considered a part of the French program for stewardship of their nuclear arsenal (the equivalent of SBSS, called Palen).

The Japanese in particular appear to have reached a high level of sophistication in their ICF laser program. In Osaka, the twelve beam Gekko-XII facility operates between 15 and 30 kJ with pulses between 0.1 and 10 nanoseconds. This facility has already reached a neutron production level of approximately 1013 neutrons per shot. The planned facilities in Japan are almost as ambitious as those of the National Ignition Facility or laser Megajoule. The Kongoh laser facility would use 92 beams to achieve 300 kJ in 3 nanoseconds.

Germany's research efforts have been focused largely on heavy ion beam fusion. Beam research is ongoing at a number of facilities and a working group has been organized to develop a design of a heavy ion beam facility to achieve ignition. This research group involves scientists from a number of countries but is organized at a German institute.

B. Proliferation

Controlling the proliferation of fission or thermonuclear weapons is already a challenging task. It involves complex inspection systems and the monitoring of facilities which produce the basic elements necessary for weapons: highly enriched uranium and/or plutonium.

The materials accounting process for fusion research would also pose significantly greater challenges. The enrichment of uranium or the separation of plutonium involve large, costly, and readily identifiable facilities, making access to weapons-usable fissile materials the central restraint on proliferation (hence the concern over the availability of fissile materials from the former Soviet Union). In countries with light water reactors, it is possible to determine when fuel was removed since the reactor must be shut down for refueling. The radioactivity of the fissile materials also allows for certain techniques of monitoring. In short, the difficulties of obtaining fissile materials or concealing production facilities for them are the main ways to control the proliferation of fission weapons.

While fission power and fission-based nuclear weapons share obvious links, the operation of fission power plants and civilian fission research does not involve creating fission explosions. Hence, commercial nuclear fission is one more step removed from weapons work than ICF programs.

Pure fusion weapons, if developed, would present far stiffer challenges. Unlike fission reactor research, there is no separation of weapons and energy research in ICF. Explosions are needed for both applications. Therefore, any research into fusion explosions for energy purposes necessarily and directly provides information on fusion explosions for military purposes. While there can be some separation through experiment design (as there is supposedly going to be at NIF), this is not inherent in the facilities' capabilities. Furthermore, fusion research is occurring in a large number of countries and a wide variety of institutions under the rubric of commercial energy research. Much of the literature is already unclassified and will continue to be so. Some of the most advanced machines are planned for countries that are not now nuclear weapons states.

If pure fusion weapons were developed, the restraints on proliferation via materials control would be weak and, in the long term, could disappear altogether. Initially, control of tritium production might provide an avenue for limiting proliferation. But tritium can be produced in commercial reactors (through the use of lithium target rods in light water reactors or by the extraction of the tritium produced in heavy water reactors, like CANDUS, due to the conversion of deuterium to tritium).¹⁵Separation facilities are also needed to extract the tritium for the target rods, but these are less complex than those for extracting plutonium from irradiated reactor fuel and could be more readily developed and operated.

Tritium is hard to detect if it is properly shielded and put into appropriate containers, making development of effective radioactive and monitoring systems very difficult, although not impossible. Further, tritium is currently not under international safeguards and there are no official plans for such safeguards. In fact, the US is in the process of greatly loosening restraints. It has initiated a program to produce test quantities of tritium for its weapons program in commercial nuclear reactors and may initiate a large-scale program for military tritium production in commercial reactors owned by the Tennessee Valley Authority. Even more troubling, however, is the possible future use of lithium and deuterium in either fusion research or in potential fusion weapons programs. While this is speculative at present so far as pure fusion weapons are concerned, it is important to note here that the thermonuclear component of fission-triggered nuclear weapons consists of a combination of these two elements in the form of lithium-deuteride. Both lithium and deuterium are non-radioactive and are readily available. There will be essentially no way to control their production or to keep track of it.

In the short term it is necessary to bring tritium stockpiles under international safeguards. This would provide a small but not sufficient measure of restraint. Perhaps more importantly, tritium production for weapons should be halted as it is inconsistent with nonproliferation and disarmament goals. (Commercial requirements are far smaller than weapons and can be met from current stockpiles and by-products from Canadian heavy water reactors).¹⁶Certainly, the program in the United States to develop a new tritium production source should be halted since it is unnecessary. Current tritium supplies are more than adequate to meet US stockpile needs if further efforts towards reducing the number of nuclear weapons are made.¹⁷

C. CTBT and ICF

Article I of the Comprehensive Test Ban Treaty states:

1. Each State Party undertakes not to carry out any nuclear weapon test explosion or any other nuclear explosion, and to prohibit and prevent any such nuclear explosion at any place under its jurisdiction or control.

2. Each State Party undertakes, furthermore, to refrain from causing, encouraging, or in any way participating in the carrying out of any nuclear weapon test explosion or any other nuclear explosion.¹⁸

The United States government, both in previous judgments and in its submission of the treaty to the US Senate for ratification, has stated that ICF and other similar experiments are not covered by the treaty's ban on explosions. The rationale is that these are not nuclear weapons explosions. The United States is not alone in this interpretation. Germany also regards experiments in controlled thermonuclear fusion to be exempt from the treaty. However, the design implications of stockpile stewardship programs, including NIF and similar programs, has caused widespread concern and is one reason India was not a signatory to the CTBT in 1996 (though it may now sign in the context of its own nuclear tests and planned stockpile stewardship program).

The main US statement in regard to pure fusion was made in the context of an interpretation of the Non-Proliferation Treaty. In 1975, in response to Swiss concerns about laser fusion research, the United States declared: "Such contained explosions area not 'other nuclear explosive devices' in the sense of the NPT and research in this area is allowed under Article IV.1."¹⁹A more detailed statement followed and was quoted in the transmittal of the CTBT to the U.S. Senate for its advice and consent:

The thrust of the NPT was to bar non-nuclear weapon states from acquiring "nuclear explosive devices." The US statement was only to the effect that laser facilities are not such "devices" and their operation by non-nuclear weapon states is therefore permissible under Article IV, allowing them to conduct research into "peaceful" uses of nuclear energy.

The CTBT negotiations have created a different record and a different set of restraints. First, the CTBT does not concern "nuclear explosive devices." Rather it bans any "nuclear explosion," including "peaceful nuclear explosions," by any state, and is intended to constrain weapons development. It also requires signatories to "prevent" nuclear explosions from occurring in their jurisdiction. Second, the negotiations involved extensive discussion of allowing some fission explosions. Specifically, the US position initially was that the CTBT should allow for hydronuclear testing which would yield up to four pounds of nuclear explosive energy. Eventually, a treaty was signed that excluded all nuclear explosions, including low yield hydronuclear tests. The US stated that sub-critical tests would be permitted under the CTBT because they did not achieve self-sustaining nuclear reactions and did not involve nuclear explosions.

This negotiating record indicates that fusion nuclear explosions equivalent to fission hydronuclear tests would be prohibited under the CTBT. Even planning for such explosions would appear to be prohibited, since the treaty requires parties "to refrain from causing, encouraging, or in any way participating in the carrying out of any nuclear weapon test explosion or any other nuclear explosion." Yet, facilities such as NIF would, if they work as their designers hope, cause nuclear explosions that are considerably larger than four-pound hydronuclear tests.

In light of the possible illegality of facilities such as NIF and LMJ, greater discussion over the interpretation of Article I is necessary. A number of factors could go into the interpretation of Article I, such as results of future CTBT Review conferences and, possibly, an opinion by the International Court of Justice. In any event, this is a question that should be answered before further work proceeds on NIF, MTF and other projects designed to achieve thermonuclear ignition.

At the heart of interpreting Article I is the definition of a nuclear explosion. Determining a workable definition is actually quite complex and somewhat arbitrary because an explosion is an interplay between total amount of energy released, energy density, and the time in which the energy is released. The time factor is perhaps the easiest. While there is no single definition of reaction time suitable for all explosions, we adopt the approach suggested by Richard Garwin and use one millisecond for the purposes of this discussion of nuclear weapons.²¹This is long enough to cover experiments most likely to assist in pure fusion weapons development. It would also automatically exclude chain reactions in fission reactors as well as fusion research based on steady-state magnetic confinement (e.g., tokamaks which are doughnut shaped facilities that use magnetic fields to confine the plasma for relatively long periods of time) from the definition of nuclear explosions. These exclusions are necessary since steady state nuclear reactions are clearly not prohibited by the CTBT, which is confined to banning nuclear explosions.

It is more difficult to pin down an exact number for the total amount of energy released and energy density that would characterize an explosion. Nuclear explosions have been defined in various ways. However, these definitions have only been made for fission explosions. In addition, limits have been proposed for fusion research under the CTBT. However, these limitations have not been based upon a technical definition of fusion explosions.

We can begin our exploration of the definition of fusion explosions by reference to better established fission explosion guidelines. The following are two definitions of fission nuclear explosions based upon two different physical criteria:

• Criticality Definition: As we have noted above, the US has used the threshold of criticality (the achievement of self-sustaining fission reactions) to define nuclear explosions of fissile materials. Under this definition the sub-critical experiments involving high explosives and fissile materials conducted at the Nevada Test Site are deemed to be allowable under the CTBT. While the sub-critical experiments pose their own problems in that they allow for continuation of weapons design, they provide a convenient starting point for determining a definition of fusion explosions.

• Specific Energy Release Definition: A 1987 Los Alamos report on the testing moratorium of 1958-1961 states that "a nuclear explosion has never been defined officially, but we consider a reasonable definition to be a specific fission energy release that is comparable to or greater than that of high explosive itself, about one kilocalorie per gram."²²In other words, the release of nuclear energy in an explosive fashion is not really an explosive unless the energy released is greater than the energy used to initiate the explosion.

A technical definition of a fusion nuclear explosion is needed in order to determine what experiments meet the letter of the CTBT. We recognize that, as with fission explosions, any definition will be arbitrary to a certain degree. Our review of the issue leads to the conclusion that the best definition for fusion explosions should rely upon the concept of ignition. Ignition has been defined in two different ways:

1. The creation of a self-propagating burn wave in the fuel pellet. This is a concept analogous to the concept of criticality in fission explosions.²³

2. A gain of one. In other words, the fusion energy output of the fuel pellet is equal to the driver energy output. A gain of one is needed to demonstrate scientific feasibility of ignition.²⁴

Ignition as a concept is analogous to both the definitions of fission explosions discussed above. However, it is far more difficult to define because there is no unambiguous physical phenomenon to which we can tie the practical onset of a fusion explosion. A propagating burn wave (our first definition) can be achieved at gains different from and less than one. For instance, it is projected that burn propagation in NIF would occur at a gain of about 0.3.²⁵In a physical sense the first definition more directly corresponds to the concept of criticality. However, the precise gain at which the self-propagating burn wave is created is device dependent, and the onset of a burn wave would be difficult to measure and verify.

For the purposes of CTBT compliance, a minimally satisfactory definition of a fusion explosions would be a gain of one. Under this limit the energy released would always be less than the driver energy input into the fuel pellet. The conditions for establishing scientific feasibility would also not be achieved. The advantage of this proposal is that it is not limited to any particular technology or an arbitrary yield, but rather is based on a definition of explosions. This limit would therefore ban all ignition experiments. However, any definition of fusion explosions geared to ignition would still allow a considerable loophole for pure fusion weapon development even though the letter of the CTBT would be met. This is because a large number of neutrons per shot can be achieved at gains just under one - that is, just below the ignition threshold. Therefore, other limitations are likely to be required to prevent the development of pure fusion weapons.

Such limits on fusion research have been proposed. They are based either on total energy output, or on specific materials or devices used in the fusion research. The following two proposals have been made by nuclear weapons experts.

• The Garwin Limit: One proposal, by Richard Garwin, would limit neutron production to 10¹⁴ neutrons/shot. This corresponds to an explosion of 0.1 gram of high explosives. Since this limit has already been approached by MTF experiments (10¹³ neutrons) and by Russian high explosive research (10¹⁴ neutrons), this would effectively freeze these programs until such time as a review of fusion experiments has been completed.²⁶Similarly, experiments on facilities such as NIF would be limited, but not prohibited, by this proposal.

• The Kidder Proposal: Another proposal would ban tritium use in systems driven directly or indirectly by high explosives. The rationale behind the tritium portion of the ban is that while the deuterium plasma will undergo a number of fusion reactions, the higher threshold for D-D reactions would make it highly unlikely to achieve ignition or burn in machines designed for igniting plasmas containing both tritium and deuterium.²⁷High-explosive-driven components will most likely be an essential component of the miniaturization of pure fusion devices. However, such a ban would not impose any limits on laser-driven or ion-beam driven research or even the Sandia wire-array z-pinch - all of which can contribute to the development of pure fusion weapons.

The limits proposed by Garwin and Kidder are not sufficient to meet the letter of the treaty on their own. A ban on ignition is required for that. However, the Garwin and Kidder limits are helpful in setting limits in order to constrain the development of new weapons.

Application of all these limitations would allow for the continuation of most experiments at NOVA and similar facilities while halting the construction of new facilities such as NIF and Megajoule whose goals are to achieve ignition. This is not to say that such a course is without its dangers for nuclear disarmament and non-proliferation. But it would at least be compatible with the letter and spirit of the CTBT.

Given the immense consequences of the development of pure fusion weapons, an indefinite delay of planned MTF experiments and a moratorium on the construction and planning of large ICF projects designed to achieve ignition (NIF, Laser Mégajoule)²⁸is necessary. If the US halts NIF it would be in a position to persuade other countries, such as France, Japan, and Germany, to halt their ICF projects designed to achieve ignition until a review process is undertaken which is comprehensive, inclusive, and world-wide. It could do this by arguing for an interpretation of the CTBT ban to include ICF ignition. The potential near-term and medium-term risks of continuing research with such machines are too great not to pursue restrictions.

The following limitations should be placed on fusion experiments in order to meet both the letter and spirit of the CTBT:

• Ignition of the fusion fuel should be used as the definition of a fusion nuclear explosion, thus prohibiting all ignition experiments, and planning or construction of all facilities designed to achieve ignition should be halted. In theory , the construction of devices such as NIF could proceed if there were a prior verifiable commitment under the CTBT to confine research to deuterium and ordinary hydrogen fuels, with which NIF and similar projects could not achieve ignition. Of course this would make such machines essentially useless since their main purpose is to achieve ignition. Experiments that do not achieve ignition can be done on existing machines.

• The total fusion energy output should be limited to 1014 neutrons/shot as proposed by Richard Garwin. This would prevent attempts to gain weapons-related information by increasing the energy of the driver and fusion energy output while staying below ignition.

• The use of tritium should be banned in all systems that use high explosives, as proposed by Ray Kidder.

In the long term, facilities such as the National Ignition Facility and MTF facilities pose even greater threats to both the CTBT and the disarmament process. As discussed above, if ignition is demonstrated in the laboratory, the weapons labs and the DOE would likely exert considerable pressure to continue investigations and to engage in preliminary design activities for a new generation of nuclear weapons (even if it is just to keep the designers interested and occupied). Ignition would also boost political support and make large-scale funding of such activities more likely.

Even without the construction of actual weapons, these activities could put the CTBT in serious jeopardy from forces both internal and external to the United States. Internally, those same pressures, which could lead to the resumption of testing of current generation weapons, could also lead to the testing of new weapons (to replace older, less safe or less reliable weapons). Externally, the knowledge that the United States or other weapons states were engaging in new fusion weapons design activities could lead other states to view this as a reversal of their treaty commitments. Comparable pressures to develop pure fusion weapons would be likely to mount in several countries. This would have severe negative repercussions for both non-proliferation and complete nuclear disarmament. The time to stop this dangerous thermonuclear quest for explosive ignition is now, before its scientific feasibility is established.

1. Zerriffi and Makhijani 1996. See below for a discussion of some of the findings of this report.

2. NAS-NRC 1986, p. 2.

3. NAS-NRC 1986, p. 35.

4. Drell 1994, p. 55.

5. Drell 1994, p. 54.

6. DOE 1995.

7. Kidder 1995, p. 5.

8. Kidder 1995, p. 5

9. Martin Kalinowski and Lars Colschen calculate that eliminating the booster and hence the secondary from US warheads would reduce their yields from a typical level of several hundred kilotons to figures in the range of a few hundred tons to a few kilotons of TNT equivalent in all but one case. In the case of the W89, the removal of the tritium bottle would cause the warhead not to operate. The overall effect of removing tritium from all warheads is estimated to be a reduction of cumulative yield by roughly two orders of magnitude. Kalinowski and Colschen 1995, p. 191.

10. The Court stated that "[t]here exists an obligation to pursue in good faith and bring to a conclusion negotiations leading to nuclear disarmament in all its aspects under strict and effective international control." (emphasis added) The obligation of all states is to "achieve a precise result - nuclear disarmament in all its aspects - by adopting a particular course of conduct, namely, the pursuit of negotiations on the matter in good faith." Legality of the Threat or Use of Nuclear Weapons, General List No. 95 (Advisory Opinion of 8 July 1996), paras. 99, 105(2)(F). See also Burroughs 1997, pp. 2-3, 48-51.

11. Blair 1995.

12. Zerriffi and Makhijani 1996.

13. DOE 1996 "Green Book", p. VII-3.

14. DOE 1996 "Green Book," p. V-10.

15. See Makhijani and Saleska 1996.

16. See Kalinowski and Colschen 1995.

17. It is estimated that tritium stocks for the US stockpile could last until 2032 if the United States were to reduce its arsenal to 1,000 warheads (a level above the most recent National Academy of Sciences recommendation). At a level of 500 warheads tritium stocks could last a little beyond 2040. See Zerriffi 1996 for details on tritium requirements for US nuclear weapons under various scenarios.

18. U.S. Senate 1997, p. 124.

19. NPT/CONF/C.II/SR.5, 1975

20. U.S. Senate 1997, pp. 4-5.

21. Garwin 1997, p.9. Garwin proposes that one millisecond be used to separate the explosive regime from the steady-state regime.

22. Thorn and Westervelt 1987, p. 4.

23. Lindl 1995, p. 6.

24. NAS-NRC 1997, pp. 10-11.

25. NAS-NRC 1997, p. 11.

26. See Jones and von Hippel 1998.

27. See Jones and von Hippel 1998. 1

28. See Table 2