would need just an explosive device—and a very simple device at that—to carry out a single action.

Second, the explosion itself might be the most effective factor in achieving the terrorists’ objectives, rather than the nuclear blast yield. Moreover, an enormous number of victims could have a negative effect on that part of the international community that adopts a positive or neutral attitude toward terrorists.

Therefore, low-yield nuclear explosive devices might be rather attractive for terrorists, barring any serious technical barriers to their construction. We shall see later that under some conditions this problem may have a solution.

Let us consider two approaches to lowering the yield of a nuclear explosion. The first is based on extremely high compression of a fissile material. It is well known that its critical mass is inversely proportional to the square of its density. For example, plutonium density in modern weapons designs is three to four times higher as a result of implosion.² At higher compressions, there is no limit on the minimum amount of fissile material required to construct a nuclear explosive. One can imagine micronuclear explosives with yields in the ton range, requiring fissile materials on the order of hundreds or even tens of grams. But what can actually be achieved along this line of development is limited only by available implosion technologies. Thus, it does not seem that this straightforward approach could be used by terrorists, because it requires a very high degree of technical expertise.

The other approach is connected to the so-called fizzle effect, which really is a preinitiation of a nuclear chain reaction in a fissile material in a supercritical state (due to the occurrence of “accidental” neutrons). As a result, the yield of the explosion is reduced in comparison with its nominal value. It should be noted that all types of nuclear weapons have a nonzero fizzle probability. One can categorize all types of nuclear weapons as either fast (implosive systems) or slow (gun-type assemblies) depending on the “waiting time” between the start of criticality and the moment of optimal condition. The fizzle effect is more probable in slow systems and for fissile materials with a high level of neutron self-emission (due mostly to the process of spontaneous fission). Therefore, nuclear terrorists could be very interested in a gun-type nuclear device with reactor-grade or weapons-grade plutonium.

Estimates of the fizzle yield were made by Dr. Carson Mark, former Theoretical Division Leader of Los Alamos National Laboratory.³ He considered “as a purely hypothetical example” a weapons-grade plutonium assembly of the implosion type used at Trinity (the first American nuclear test, July 16, 1945), with the nominal yield of 20 kilotons. The fizzle yield in this case might be 0.5 kiloton. A similar assembly in a gun-type system would produce a fizzle yield of some 10-20 tons.

The fizzle phenomenon is of a statistical nature where the main parameter would be the moment of neutron occurrence during the waiting period. The fizzle could be managed to some extent, but management of this kind requires