[NCI Logo] space  

Can Terrorists Build Nuclear Weapons?

Carson Mark, Theodore Taylor, Eugene Eyster, William Maraman, Jacob Wechsler

J. Carson Mark is a member of the Nuclear Regulatory Commission's Ad visory Committee on Reactor Safeguards and of the Foreign Weapons Eval uation Group of the U.S. Air Force. He is a former division leader of Los Alamos National Laboratories' Theoretical Division and serves as a consultant to Los Alamos and a number of governmental agencies.

Theodore Taylor is chairman of the board of Nova, Inc., which specializes in solar energy applications. He is a nuclear physicist who once designed the United States' smallest and largest atomic (fission) bombs. He also designed nuclear research reactors. He has served as deputy director (Scientific) of the Defense Atomic Support Agency and as an independent consultant to the U.S. Atomic Energy Commission. He is coauthor (with Mason Willrich) of Nuclear Theft: Risks and Safeguards and is the subject of John McPhee's The Curve of Binding Energy.

Eugene Eyster is a former leader of Los Alamos National Laboratories' WX Division, which is responsible for the explosive components of nuclear weap ons. A specialist in chemical explosives, he participated in the Manhattan Project.

William Maraman, a specialist in chemical and metallurgical processing of plutonium and uranium, is director of TRU Engineering Co., which does consulting work on transuranic elements. He was at Los Alamos National Laboratories for thirtyseven years, where he was leader of the Plutonium, Chemistry and Metallurgy Group and of the Material Sciences Division.

Jacob Wechsler is a physicist specializing in nuclear explosives. He was a member of the Manhattan Project and was leader of Los Alamos National Laboratories' WX Division, which is responsible for the explosive components of nuclear weapons.

General Observations

Two options for nuclear devices to be built by terrorists are considered here: that of using the earliest design principles in a so-called crude design and that of using more advanced principles in a so-called sophisticated design.

A crude design is one in which either of the methods successfully dem onstrated in 1945---the gun type and the implosion type---is applied. In the gun type, a subcritical piece of fissile material (the projectile) is fired rapidly into another subcritical piece (the target) such that the final assembly is supercritical without a change in the density of the material. In the implosion type, a near-critical piece of fissile material is compressed by a converging shock wave resulting from the detonation of a surrounding layer of high explosive and becomes supercritical because of its increase in density.

A small, sophisticated design is one with a diameter of about 1 or 2 feet and a weight of one hundred to a few hundred pounds, so that it is readily transportable (for example, in the trunk of a standard car). Its size and weight may be compared with that of a crude design, which would be on the order of a ton or more and require a larger vehicle. It would also be possible, in about the same size and weight as a crude model but using a more sophis ticated design, to build a device requiring a smaller amount of fissile material to achieve similar effects.

For a finished implosion device using a crude design, terrorists would need something like a critical mass of uranium (U) or plutonium (Pu) or, possibly, UO2 (uranium oxide) or Pu02 (plutonium oxide). For a gun type device, substantially more than a critical mass of uranium is needed, and plutonium cannot be used. It may be assumed that the terrorists would have acquired (or plan to acquire) such an amount either in the form of oxide powder (such as might be found in a fuel fabrication plant), in the form of finished fuel elements for a reactor---whether power, research, or breeder--- or as spent fuel.

For a small, sophisticated design, the terrorists may need a similar amount of fissile material since practically all the presumed reductions in size and weight have to be taken from the assembly mechanism, and, with a less powerful assembly, not only will it be important to have the active material in its most effective form, but its amount will have to be sufficient to achieve supercriticality. Alternatively, a smaller amount could be used in a sophis ticated design with a more powerful and heavier assembly mechanism.

Conceivably oxide powder might be used as is, although terrorists might choose to go through the chemical operation of reducing it to metal. Such a process would take a number of days and would require specialized equip ment and techniques, but these could certainly be within the reach of a dedicated technical team.

Fuel elements of any type will have to be subjected to chemical pro cessing to separate the fissile material they may contain from the inert clad ding material or other diluents. This process would also require specialized equipment, a supply of appropriate reagents, well-developed techniques spe cific to the materials handled, and at least a few days to conduct the operation. Spent fuel from power reactors would contain some plutonium but at such low concentrations that it would have to be separated from the other materials in the fuel. It would also contain enough radioactive fission fragments that the chemical separation process would have to be carried out by remote operation, a very complicated undertaking requiring months to set up and check out, as well as many days for the processing itself. The fresh fuel for almost all power reactors would be of no use, since the uranium enrichment is too low to provide an explosive chain reaction.

The terrorists would need something like a critical mass of the material they propose to use. For a particular fissile material, the amount that con stitutes a critical mass can vary widely depending on its density, the char acteristics (thickness and material) of the reflector employed, and the nature and fractional quantity of any inert diluents present (such as the oxygen in uranium oxide, the uranium 238 in partially enriched uranium 235, or chem ical impurities).

For comparison purposes, it is convenient to note the critical masses with no reflector present (the "bare crit") of a few representative materials at some standard density. For this discussion, the following examples of bare critical masses have been chosen:

10 Kilograms (kg) of Pu 239, alpha-phase metal (density = 19.86 grams per cubic centimeter [gm/cc]).

52 kg of 94% U-235 (6% U-238) metal (density = 18.7 gm/cc).

approximately 110 kg of U02 (94% U-235) at full crystal density (density = 1I gm/cc). approximately 35 kg of Pu02 at full crystal density (density = 11.4 gm/cc).

In all cases (others as well as these), the mass required for a bare crit varies inversely as the square of the density. Thus, the bare crit of delta-phase plutonium metal (density = 15.6 gm/cc ) is about 16 kg. Similarly, at densities the square root of two times larger than those above, the bare crit masses would be one-half those indicated. If any reflector is present, the mass re quired to constitute a critical assembly would be smaller than those above. With a reflector several inches thick, made of any of several fairly readily available materials (such as uranium, iron, or graphite, for example), the critical mass would be about half the bare crit. Thicker reflectors would further reduce the mass but would be more awkward without providing much more of a reduction. Although beryllium is particularly effective in this respect---providing critical masses as low as one-third the bare crit---it is not readily available in the form needed and is not considered further.) It is consequently assumed here that a mass of half the bare crit is what terrorists would require to complete a near-critical (crude) assembly.

With respect to the effects of dilution by isotopes of heavy elements, only the two most obvious cases need be considered. One is that of reactor- grade plutonium. This material is not uniquely specified, since the fractional amount of the Pu-240 depends on the level of exposure of the fuel in the reactor before it is discharged. However, at burn-up levels somewhat higher than present practice, the bare crit of plutonium would be only some 25-35 percent higher than that for pure Pu-239. Because of spontaneous fission, the effect of the Pu-240 on the neutron source in the material is thus likely to be more important than its effect on the critical mass. Nevertheless, nuclear weapons can be made with reactor-grade plutonium.

The other obvious dilution case is that of uranium at enrichments lower than 94 percent. Here the effect on critical mass, and consequently on the amount of material that must be acquired and moved by the assembly system, is quite appreciable. For example, the bare crit of 50 percent enriched ura nium is about 160 kg (~3 times that of 94 percent material) and for 20 percent material about 800 kg ( ~15 times that for 94 percent). Similar factors will apply for uranium oxide as a function of enrichment. In this same con nection, it may be noted that the mixed oxide fuel once considered for the Clinch River Breeder Reactor (~22 percent plutonium oxide plus ~78 per cent uranium oxide) would correspond to uranium at an enrichment of somewhat less than 40 percent and have a critical mass a little more than four times larger than 94 percent uranium oxide.

As a final general observation, for a crude design, terrorists would need something like 5 or 6 kg of plutonium or 25 kg of very highly enriched uranium (and more for a gun-type device), even if they planned to use metal. They would have to acquire more material than is to go into the device, since with metal considerably more material is required to work with than will appear in the finished pieces. The amounts they would need can be compared with the formula quantities identified in federal regulations for the protection of nuclear materials: 5 kg U-235, or 2 kg plutonium. Sites at which more than a formula quantity is present are required to take measures to cope with a determined, violent assault by a dedicated, well-trained, and well-armed group with the ability to operate as two or more teams. Trans port vehicles that carry more than a formula quantity must be accompanied by armed escort teams and have secure communications with their base. Transport vehicles carrying smaller amounts are not so heavily guarded, but there are provisions intended to ensure that in the aggregate no more than a formula quantity is on the road at one time. For terrorists having to acquire at least several formula quantities, there are formidable barriers to overcome.

Crude Designs

Crude designs are discussed primarily in the context of the problems facing a terrorist group. Schematic drawings of fission explosive devices of the earliest types showing in a qualitative way the principles used in achieving the first fission explosions are widely available. However, the detailed design drawings and specifications that are essential before it is possible to plan the fabrication of actual parts are not available. The preparation of these drawings requires a large number of man-hours and the direct participation of indi viduals thoroughly informed in several quite distinct areas: the physical, chemical, and metallurgical properties of the various materials to be used, as well as the characteristics affecting their fabrication; neutronic properties; radiation effects, both nuclear and biological; technology concerning high explosives and/or chemical propellants; some hydrodynamics; electrical cir cuitry; and others.

It is exceedingly unlikely that any single individual, even after years of assiduous preparation, could equip himself to proceed confidently in each part of this diverse range of necessary knowledge and skills, so that it may be assumed that a team would have to be involved. The number of specialists required would depend on the background and experience of those enlisted, but their number could scarcely be fewer than three or four and might well have to be more. The members of the team would have to be chosen not only on the basis of their technical knowledge, experience, and skills but also on their willingness to apply their talents to such a project, although their susceptibility to coercion or considerations of personal gain could be factors. In any event, the necessary attributes would be quite distinct from the paramilitary capability most often supposed to typify terrorists.

Assuming the existence of a subnational group equipped for the activist role of acquiring the necessary fissile material and the technical role of making effective use of it, the question arises as to the time they might need to get ready. The period would depend on a number of factors, such as the form and nature of the material acquired and the form in which the terrorists proposed to use it; the most important factor would be the extent of the preparation and practice that the group had carried out before the actual acquisition of the material. To minimize the time interval between acquisition and readiness, the whole team would be required to prepare for a consid erable number of weeks (or, more probably, months) prior to acquisition. With respect to uranium, most of the necessary preparation and practice could be worked through using natural uranium as a stand-in.

The time intervals might range from a modest number of hours, on the supposition that enriched uranium oxide powder could be used as is, to a number of days in the event that uranium oxide powder or highly enriched (unirradiated) uranium reactor fuel elements were to be converted to ura nium metal. The time could be much longer if the specifications of the device had to be revised after the material was in hand. For plutonium, the time intervals would be longer because of the greatly increased hazards involved (and the absolute need of foreseeing, preparing for, and observing all the necessary precautions). in addition, although uranium could be used as a stand-in for plutonium in practice efforts, there would be no opportunity to try out some of the processes required for handling plutonium until a suf ficient supply was available.

To achieve a minimum turnaround time, the terrorists would, before acquisition, have to decide whether to use the material as is or to convert it to metal. They would have to make the decision in part in order to proceed with the design considerations, in part because the amounts needed would be different in the two cases, and in part to obtain and set up any required equipment.

For the first option---using oxides without conversion to metal---the terrorists would need accurate information in advance concerning the phys ical state, isotopic composition, and chemical constituents of the material to be used. Although they would save time by avoiding the need for chemical processing, one disadvantage (among others) is the requirement for more fissile material than would be needed were metal to be used. This larger amount of fissile (and associated) material would require a larger weight in the assembly mechanism to bring the material into an explosive configuration.

As to the second option---converting the materials to metal---a smaller amount of fissile material could be used. However, more time would be needed and quite specialized equipment and techniques---whether merely to reduce an oxide to the metal or to separate the fissile material from the cladding layers in which it is pressed or sintered in the nuclear fuel elements of a research reactor, for example. The necessary chemical operations, as well as the methods of casting and machining the nuclear materials, can be (and have been) described in a straightforward manner, but their conduct is most unlikely to proceed smoothly unless in the hands of someone with experience in the particular techniques involved, and even then substantial problems could arise.

The time factor enters the picture in a quite different way. In the event of timely detection of a theft of a significant amount of fissile material--- whether well suited for use in an explosive device or not---all relevant branches of a country's security forces would immediately mount an intensive response. In addition to all the usual intelligence methods, the most sensitive technical detection equipment available would be at their disposal. As long as thirty-five years ago, airborne radiation detectors proved effective in prospecting for uranium ore. Great improvements in such equipment have been realized since. A terrorist group would therefore have to proceed deliberately and with caution to have a good chance of avoiding any mishap in handling the material, while at the same time proceeding with all possible speed to reduce their chance of detection.

In sum, several conclusions concerning crude devices based on early design principles can be made.

I . Such a device could be constructed by a group not previously engaged in designing or building nuclear weapons, providing a number of requirements were adequately met.

2. Successful execution would require the efforts of a team having knowledge and skills additional to those usually associated with a group engaged in hijacking a transport or conducting a raid on a plant.

3. To achieve rapid turnaround (that is, the device would be ready within a day or so after obtaining the material), careful preparations extending over a considerable period would have to have been carried out, and the materials acquired would have to be in the form prepared for.

4. The amounts of fissile material necessary would tend to be large--- certainly several, and possibly ten times, the so-called formula quantities.

5. The weight of the complete device would also be large---not as large as the first atomic weapons (~10,000 pounds), since these required aero dynamic cases to enable them to be handled as bombs, but probably more than a ton.

6. The conceivable option of using oxide powder (whether of uranium or plutonium) directly, with no postacquisition processing or fabrication, would seem to be the simplest and most rapid way to make a bomb. However, the amount of material required would be considerably greater than if metal were used. Even at full cyrstal density, the amounts are large enough to appear troublesome: ~55 kg (half bare crit) for 94 percent uranium oxide and ~17.5 kg for plutonium oxide. However, the density of the powder as acquired is nowhere close to crystal density. To approach crystal density would require a large and special press, and the attempt to acquire such apparatus would constitute the sort of public event that might blow the cover of a clandestine operation. Besides, the time required for processing with such a press would preclude a rapid turnaround. Even to achieve densities a little above half of crystal would require some pressing apparatus (not as conspicuous as a large press and conceivably obtainable quietly), but time would again be required to process material quantities of perhaps three or four times those above. The densities available in powder without pressing are not well determined but are quite low, probably in the range of 3 to 4 gm/Cc, although possibly lower.

Within the confines of the crude design category---that of a device guar anteed to work without the need for extensive theoretical or experimental demonstration---an implosion device could be constructed with reactor- Brade plutonium or highly enriched uranium in metal or possibly even oxide form. The option of using low-density powder directly in a gun-type assembly should probably be excluded on the basis of the large material requirements.

There remains the possibility of using a rather large amount of oxide powder (tens of kilograms or possibly more) at low density in an implosion- type assembly and simply counting on the applied pressure to increase the density sufficiently to achieve a nuclear explosion. Some sort of workable device could certainly be achieved in that way. However, obtaining a per suasive determination of the actual densities that would be realized in a porous material under shock pressure (and hence of the precise amount of material required) would be a very difficult theoretical (and experimental) problem for a terrorist team. In fact, solving this problem does not belong in the crude design category. Still, a workable device could be built without the need for extensive theoretical or experimental demonstration.

The amount of low-density oxide powder required for a small, crude, implosion-type device is far larger than previously suggested by Theodore Taylor; his view has changed only as to the feasibility of a small, crude device such as terrorists might attempt to build with a single small container of plutonium oxide powder seized from a fuel fabrication plant. We agree, however, that a crude implosion device could be constructed with reactor-grade plutonium or highly enriched uranium in metal or possibly even in oxide form.

7. Devices employing metal in a crude design could certainly be con structed so as to have nominal yields in the 10 kiloton range---witness the devices used in 1945. By nominal yield is meant the yield realized if the neutron chain starts after the assembly is complete and the fissile material is at or near its most supercritical configuration: projectile fully seated in the target for the gun-type device or all the material compressed in the implosion device. In all such systems, there is an interval between the moment when the fissile material first becomes critical (projectile still on its way to its destination, or only a small part of the material compressed) and the time it reaches its intended state. During this interval, the degree of supercriticality is building up toward its final value. If a chain reaction were initiated by neutrons from some source during this period, the yield realized would be smaller---possibly a great deal smaller---than the nominal yield. Such an event is referred to as preinitiation (or sometimespredetonation). Obviously, the longer is this interval or the greater is the neutron source in the active material, the larger is the probability of experiencing a preinitiation. The neutron source in even the best plutonium available (lowest Pu-240 content) is so large and the time interval for a gun-type assembly with available pro jectile velocities (~1000 ft./sec.) is so long that predetonation early in this time interval is essentially guaranteed. For this reason, plutonium cannot be used effectively in a gun-type assembly. The neutron source in enriched uranium is several thousand times smaller than in the plutonium referred to, so that uranium can be used in a gun-type assembly (with available projectile velocities) and have a tolerable preinitiation probability. For this to be true, it is necessary to have rather pure uranium metal, since even small amounts of some chemical impurities can add appreciably to the neutron source. The source in uranium oxide, for example, may be ten or so times larger than in pure metal; the source in reactor-grade plutonium may be ten or more times larger than in weapons-grade plutonium. However, reactor-grade plutonium can be used for making nuclear weapons.

If the assembly velocities (of the projectile or material driven by an implosion) are quite low, the earliest possible preinitiation could lead to an energy release (equivalent weight of high explosive) not many times larger than the weight of the device. If the velocities are quite high (so that the degree of supercriticality increases appreciably during the very short time it takes the neutron chain to build up), the lowest preinitiation yield may still be in the 100 ton range, even in a crude design. Reductions in the weight of the assembly-driving mechanism (whether gun-firing apparatus or amount of high explosive) will, other things being equal, tend to result in lower assembly velocities. The considerations outlined will put some limits on what may be decided to be desirable in connection with a crude design.

8. There are a number of obvious potential hazards in any such operation, among them those arising in the handling of a high explosive; the possibility of inadvertently inducing a critical configuration of the fissile material at some stage in the procedure; and the chemical toxicity or radiological hazards inherent in the materials used. Failure to foresee all the needs on these points could bring the operation to a close; however, all the problems posed can be dealt with successfully provided appropriate provisions have been made.

9. There are a number of other matters that will require thoughtful planning, as well as care and skill in execution. Among these are the need to initiate the chain reaction at a suitable time and for some reliable means to detonate the high explosive when and as intended.

10. Some problems that have required a great deal of attention in the nuclear-weapons program would not seem important to terrorists. One of these would be the requirement (necessary in connection with a weapons stockpile) that devices have precisely known yields that are highly repro ducible. Terrorists would not be in a position to know even the nominal yield of their device with any precision. They would not have to meet the extremely tight specifications and tolerances usual in the weapons business, although quite demanding requirements on these points would still be nec essary. Similarly, in connection with a stockpile of weapons, much attention has been given to one-point safety: the assurance that no nuclear yield would be realized in the event of an unplanned detonation of the high explosive, such as might occur in the case of an accident or fire. To ensure the safety of bystanders, this requirement has been deemed important in the context of a large number of devices widely deployed and subject to movement from place to place by a variety of transport modes and by a series of handling teams. Terrorists would not be concerned with this problem, although they would still have a great interest in the safe handling of their device.

11. Throughout the discussion, it has been supposed that the terrorists were home grown. It is conceivable that such an operation could be spon sored by another country, in which case some of the motivation, technical experts, and muscle men might be brought in from outside. This difference would not change the problems that would have to be addressed or the operations required, but it could increase the assurance that important points are not overlooked. It might also provide the basis for considering a sophis ticated design rather than a crude type.

More Sophisticated Devices

Most of the schematic drawings that are available relate to the earliest, most straightforward designs and indicate in principle how to achieve a fission explosion, without, however, providing the details of construction. Since 1945, notable reductions in size and weight, as well as increases in yield, have been realized. Schematic drawings of an entirely qualitative sort are also available that indicate the nature of some of the principles involved in these improvements.

Merely on the basis of the fact that sophisticated devices are known to be feasible, it cannot be asserted that by stealing only a small amount of fissile material a terrorist would be able to produce a device with a reliable multikiloton yield in such a small size and weight as to be easy to transport and conceal. Such an assertion ignores at least a significant fraction of the problems that weapons laboratories have had to face and resolve over the past forty years. It is relevant to recall that today's impressively tidy weapons came about only at the end of a long series of tests that provided the basis for proceeding further. For some of these steps, full-scale nuclear tests were essential. In retrospect, not every incremental step taken would now seem necessary. Indeed, knowing only that much smaller and lighter weapons are feasible, it is possible at least to imagine going straight from the state of understanding in 1945 to a project to build a greatly improved device. The mere fact of knowing it is possible, even without knowing exactly how, would focus terrorists' attention and efforts.

The fundamental question, however, would still remain: that of whether the object designed and built would or would not actually behave as pre dicted. Even with their tremendous experience, the weapons laboratories find on occasion that their efforts are flawed. Admittedly, weapons designers are now striving to impose refinements on an already highly refined product, but they have had to digest surprises and disappointments at many points along the way.

For persons new to this business, as it may be supposed a terrorist group is, there is a great deal to learn before they could entertain any confidence that some small, sophisticated device they might build would perform as desired. To build the device would require a long course of study and a long course of hydrodynamic experimentation. To achieve the size and weight of a modern weapon while maintaining performance and confidence in perfor mance would require one or more full-scale nuclear tests, although consid erable progress in that direction could be made on the basis of nonnuclear experiments.

In connection with an effort to reduce overall size and weight as far as possible, it would be necessary to use fissile material in its most effectiveform, plutonium metal. Moreover, while reducing the weight of the assembly mechanism, which implies reducing the amount of energy available to bring the fissile material into a supercritical configuration, it would not be possible at the same time to reduce the amount of fissile material employed very much. In this case, the amount of fissile material required in the finished pieces would be significantly larger than the formula quantity. Alternatively, in an implosion device without a reduction in weight and size, it would be possible to reduce the amount of nuclear materials required by using more effective implosion designs than that associated with the crude design.

In either case---a small or a large sophisticated device---the design and building would require a base or installation at which experiments could be carried out over many months, results could be assessed, and, as necessary, the effects of corrections or improvements could be observed in follow-on experiments. Similar considerations would apply with respect to the chem ical, fabrication, and other aspects of the program.

The production of sophisticated devices therefore should not be consid ered to be a possible activity for a fly-by-night terrorist group. It is, however, conceivable in the context of a nationally supported program able to provide the necessary resources and facilities and an established working place over the time required. It could be further imagined that under the sponsorship of some malevolent regime, a team schooled and prepared in such a setting could be dispatched anywhere to acquire material and produce a device. In such a case, although the needs of the preparation program might have been met, the terrorists would still have to obtain and set up the equipment needed for the reduction to metal and its subsequent handling and to spend the time necessary to go through those operations.

In summary, the main concern with respect to terrorists should be fo cused on those in a position to build, and bring with them, their own devices, as well as on those able to steal an operable weapon.




Return to Nuclear Terrorism Page NCI Home Page




mail@nci.org