American Scientist

Tritium on Ice: The Dangerous New Alliance of Nuclear Weapons and Nuclear Power. Kenneth D. Bergeron. xii + 230 pp. The MIT Press, 2002. $24.95.

The nuclear stockpile of the United States consists entirely of thermonuclear weapons that contain two distinct stages: a fission explosion (the so-called primary) driving a larger combined fission and fusion explosion (the so-called secondary). The primary is "boosted"—that is, a small container of tritium (a heavy radioactive isotope of hydrogen), generally mixed with other gases, is placed inside the plutonium pit of the weapon. Although nuclear reactions induced in this tritium when the bomb is set off produce only a negligible amount of fusion energy, the neutrons they release greatly enhance the fission yield of the primary. Because tritium has a half-life of approximately 12 years, and because there is need for a significant "performance margin" for the primary to have sufficient yield to ignite the secondary, the tritium must be replenished at frequent intervals.

Until recently, tritium and plutonium for nuclear weapons were produced in government-owned nuclear reactors dedicated to that military purpose. These aging facilities have been shut down, both because of growing concerns about their safety and because today the United States has a great excess of plutonium for weapons. But a decision had to be made regarding how to resupply the tritium.

The following options existed: (1) Do nothing, on the assumption that the radioactive decay of tritium, by rendering weapons unusable, will diminish the stockpile at a pace equal to or slower than that required by arms control agreements. (2) Buy tritium from other countries—in particular, Canada or Russia. (3) Construct a reactor of either an old or new type, or finish building one that is incomplete, and dedicate it solely to producing tritium for nuclear weapons. (4) Construct an accelerator to produce neutrons, which in turn can be used to breed tritium from targets containing materials such as helium-3 or lithium-6. (5) Pay for "irradiation services" from an existing commercial nuclear reactor by introducing lithium-bearing rods into it, from which tritium can then be extracted.

In December of 1998, Secretary of Energy Bill Richardson announced that he had chosen to pursue this last approach, while also continuing research and development on construction of an accelerator. Accordingly, a contract was negotiated with the government-owned Tennessee Valley Authority (TVA) to introduce lithium-bearing rods into two of TVA's commercial reactors.

Tritium on Ice, by Kenneth D. Bergeron, is a detailed analysis and a severe criticism of Richardson's decision, on several grounds: First, Bergeron maintains that it breaches the long-established "firewall" between civilian production of nuclear power and military production of nuclear weapons. Second, he questions the safety of the reactors that would produce tritium under the proposed arrangement. On this issue he has great expertise, having worked for Sandia National Laboratories for about 25 years on safety issues related to both military and civilian nuclear reactors. Finally, he describes in detail the bureaucratic process leading to Richardson's decision and finds it defective.

Legally, the Atomic Energy Act directs that "special nuclear materials" shall not be produced in civilian nuclear power plants. Plutonium and highly enriched uranium and other heavy fissionable isotopes fall into this category, but tritium does not. An attempt by U.S. Rep. Edward J. Markey of Massachusetts to pass legislation to include tritium under that definition was unsuccessful. Thus no law bars Richardson's decision, but Bergeron maintains that the spirit of the existing law has been broken.

Although this is indeed true, it is questionable how much the decision heightens the risk of proliferation of nuclear weapons. In the past the barrier between civilian and military uses has at times been penetrated, but generally in the direction of military reactors providing civilian benefits (such as medical isotopes and limited amounts of electric power). In this case, the breach is in the opposite direction. The United States, as a nuclear-weapons state, is entitled under the terms of the Treaty on Nonproliferation of Nuclear Weapons to produce nuclear arms. The barrier between the civilian and military nuclear fuel cycles was erected as a signal to non–nuclear-weapons states to encourage them to pursue nonmilitary nuclear power under safeguards designed to prevent nuclear materials produced in civilian power plants from being used for military purposes.

Bergeron draws too sharp a distinction between the option Richardson selected and the other methods of tritium production in terms of their influence on the proliferation of nuclear weapons. Specifically, he endorses the dubious notion that the production of tritium with an accelerator would not increase the risk of proliferation. In fact, such a high-intensity accelerator could also be used to breed plutonium. Indeed, starting in 1948 the University of California Radiation Laboratory pursued a project to do just that—to breed plutonium from depleted uranium using a high-intensity accelerator beam. Technically speaking, all nuclear fuel cycles, even proposed fusion reactors, constitute a "proliferation risk," because they produce neutrons, which can be used to breed plutonium or tritium. Therefore devices of any type that produce large numbers of neutrons require some level of safeguards to prevent such diversion; the various fuel cycles differ only in terms of the burden they impose on the safeguards process. Thus the effect of Richardson's decision on the risk of proliferation is not as black-and-white as Bergeron maintains.

Bergeron's second criticism relates to reactor safety. The TVA reactors in question are of the pressurized light-water type—one of the few types whose outer containment building is not deemed sufficiently robust under existing regulations to contain the steam that would be released if the cooling system failed. Therefore the containment building incorporates buckets containing about 1,000 tons of ice to sufficiently reduce the total steam pressure. Hence the book's catchy title: Tritium on Ice.

Bergeron points out that there would be additional risk in the case of a "direct heating" event—that is, a core meltdown in the reactor (such as occurred at Three Mile Island). Here the molten core would directly interact with the concrete of the containment structure, releasing gases that would not be absorbed by the ice. This possibility has been a contentious safety issue for a considerable period of time.

After a long series of studies, the Nuclear Regulatory Commission has certified pressurized light-water reactors of this design to be sufficiently safe in commercial practice. However, Bergeron argues that the margin of safety would be further diminished by the introduction of lithium-bearing rods, because the extra neutrons that would be absorbed by those rods require additional neutron reactivity in the reactor; that is, the fission reactions have to proceed at a somewhat higher rate to maintain electrical power output. The Nuclear Regulatory Commission is still studying this matter. Bergeron's discussion of the subject includes a very valuable and readable history of the evolution of reactor safety standards for the U.S. nuclear power industry.

Bergeron also attacks Richardson's political motives. He finds evidence of extensive lobbying by TVA, which wanted to produce tritium. In his view, TVA managers feared that the government was going to decrease its commitment to civilian nuclear power and therefore sought a stronger, more legitimate claim on federal support as a supplier of an essential military commodity. They wanted to obtain federal support for completion of an unfinished reactor plant rather than a contract for irradiation services, but as a second choice they accepted the latter. The nuclear power industry elsewhere in the United States showed little interest in supplying tritium.

As an investigative writer, Bergeron notes that vacillation and inconsistencies in governmental process preceded the decision. Here he tends to characterize the proceedings as conspiratorial rather than as merely erratic or expedient. Of course the main reason for Richardson's decision was financial. Constructing a dedicated reactor or accelerator for tritium production would have required large up-front capital investments, whereas Richardson's decision implies a relatively small initial outlay, with future costs depending on the level of demand, which is not easy to predict.

Richardson's other options were limited. Purchasing tritium abroad was (and remains) politically and legally infeasible. Canadian law does not permit sale for military purposes of the very large tritium inventories stemming from the nuclear fuel cycle in CANDU (Canadian deuterium uranium) reactors. And depending on the Russians to keep U.S. bombs working would hardly be acceptable in the United States even if the Russians were willing to accept such a deal.

In summary, Tritium on Ice contains an interesting and illuminating, although subjective, analysis of Richardson's decision. I criticize the book, however, for greatly exaggerating the impact of that decision on the proliferation of nuclear weapons. Criticism of the United States for not meeting its obligations to stem the proliferation of nuclear weapons should be based on the country's failure to shrink its stockpile at a rate consistent with the commitment under the Nuclear Nonproliferation Treaty to pursue nuclear arms reductions in good faith and work toward eventual elimination of these weapons. In my opinion, there is no rational need for the world's dominant nation in conventional (that is, nonnuclear) armaments to maintain some 10,000 nuclear weapons.

Bergeron holds that the Moscow Treaty on Strategic Offensive Reductions, signed May 24, 2002, reduces the nuclear weapons stockpiles sufficiently to make it unnecessary to pursue tritium procurement any time soon. I wish this were true. Actually, the Moscow Treaty, although it slowly reduces the category of "operationally deployed strategic nuclear weapons," does not require a significant decrease in the total size of the "enduring stockpile." (Similarly, the "reduce and hedge" policy of the Clinton Administration provided for large residual reserves.) Thus current U.S. policy on nuclear weapons excludes the option of doing nothing for the time being to produce more tritium. The failure to significantly shrink the total stockpile is the principal defect of that policy in supporting the cause of nuclear nonproliferation. Although the course Richardson chose can be criticized for the many reasons cited in Tritium on Ice, the influence of his decision on the overall U.S. nonproliferation policy is minor compared with the fundamental failure of our nation to reduce its nuclear arms.