Analysis of Risks of Diversion of Plutonium or Highly Enriched Uranium

Abstract

The production of a nuclear explosive device requires a range of specialized skills and facilities and a sufficient quantity of concentrated fissionable material. Significant amounts of such materials exist in hundred of locations, in a variety of forms including mixtures with nonfissionable materials and with radioisotopes. Nuclear weapons start with pure metallic materials of predominantly a single isotope (uranium-235 or plutonium-239). Chemically or isotopically diluted materials either require larger amounts of fissile isotope or else cannot be made to produce an explosive reaction without extensive processing involving sophisticated techniques and facilities. Obstacles to diversion of fissionable materials include combinations of physical security, guards and defense forces, accountancy, diluents, limited amounts of material in any one location, and timely detection and forceful reaction to attempts at diversion or theft.

A generalized risk assessment model is developed, incorporating the main factors that influence the probability of successful diversion. A roughly inverse relationship is involved between the amount of material potentially seized or diverted and the levels, skills, facilities, and resources required to make a workable nuclear device. Absolute probabilities are indeterminate. Experience indicates an upper bound of about 1.6×10⁻⁶/year/weapon-equivalent amount of material. Inequality relationships can be used to develop relative rankings and semiquantitative magnitudes of diversion risks.

Small university research reactors appear to contribute less than 0.1% of the total diversion risk of U.S. materials. Limiting the amounts of material at any one university location is practical so that thefts from several--perhaps half a dozen--reactors would be required to make a single explosive device plausible.

A related, nonnuclear risk appears relatively likely, arising from the likelihood of unwarranted public overreactions to a variety of actions from vandalism to attempted thefts. These reactions are likely, given the present levels of discrimination on this topic.

Cost/benefit relationships are also derived. These exhibit disparities of several orders of magnitude in the ratios of commitments made or planned to the size of risk exposure reductions attainable.