Spent nuclear reactor fuel for the most common reactor (the light water reactor) typically contains about 94 percent uranium, five percent highly radioactive fission products, (a few of which are very long-lived), just under one percent plutonium, and small proportions of other transuranic elements such as neptunium and americium.

Long-lived fission products, such as iodine-129 and long-lived heavy elements such as plutonium-239 can be converted into short-lived radionuclides by bombarding them with neutrons in various kinds of nuclear reactors. This process is called transmutation. (Note: transmutation by fission is responsible for the generation of nuclear energy in the first place.) Proponents claim that transmutation might enable a high-level waste management program to avoid the construction of a geologic repository.

In order to use nuclear reactors (whether of existing varieties or new kinds) to transmute long-lived radionuclides into short-lived ones, it is first necessary to separate out the long-lived radionuclides. Hence, successful waste management by this approach requires both separation technologies and transmutation technologies. The report considers both technological and economic aspects of these technologies.

The National Research Council study was partly motivated by the need for an independent evaluation of claims made by the US Department of Energy and its contractors, General Electric and Argonne National Laboratory, that the advanced liquid metal reactor and the associated reprocessing technology, called electrometallurgical processing, could drastically reduce the amount of long-lived radioactive waste needing management and disposal. The study examines whether any combination of separation and transmutation technologies could result in the conversion of sufficient amounts of (essentially all?) long-lived radionuclides into relatively short-lived radionuclides that can be stored until they have decayed to very low levels of radioactivity. The study concluded that current technologies could not accomplish this purpose because sufficient amounts of long-lived radionuclides would remain in all cases to require the construction of a repository. Moreover, it would take hundreds of years to reduce the radioactivity of transuranics by a factor of ten, and thousands of years to reduce it by a factor of one hundred.

As to technologies still under development, such as a sub-critical reactor connected to an accelerator neutron source, proposed by Los Alamos National Laboratory,² these would take a long time to develop and it remains speculative whether they could be commercialized. Even if they were, it is "improbable" that the very high degree of separation necessary to guarantee transmutation of essentially all long-lived radionuclides could be achieved.

The study provides up-to-date cost estimates for building and operating a new reprocessing plant in the United States, based in European experience. The study states that reprocessing costs for actual plants (THORP in Britain and UP3 in France) are reported to be $600 to $1,400 per kilogram of heavy metal.

Costs of a new US reprocessing plant would depend on whether the plant is government owned, utility owned, or commercial (see Table below). These costs would be different because governments would have the lowest costs of capital, no taxes, no insurance requirements, and no requirements for profit, while at the other extreme, a purely private company would have the highest charges in all these categories. Utilities, which are regulated monopolies, would have costs between those of government and private industry in most of these areas.

These costs apply to reprocessing the spent fuel that results from an initial loading of fresh uranium fuel. If the reactor is fueled with recycled uranium or plutonium then the costs of reprocessing would be higher. The processes that would need to be added to reduce process losses of radioactive materials to very low levels would add further costs.

Finally, the study rejects the claim by Argonne National Laboratory that electrometallurgical processing costs would be about $350 per kilogram of heavy metal, or only about one-sixth the private facility costs of a PUREX plant. Past experience indicates that initial cost estimates are likely to grow as the technology is developed. The report cites an independent cost estimate that electrometallurgical processing would be 57 percent more expensive than the PUREX process.

In sum, the study concludes that no separation or transmutation technologies could help avoid repository programs, which will remain essential to managing wastes from existing reactors.

1. Available from the National Academy Press, Washington, D.C., $79.95 plus postage. The summary of the report can be read on the World Wide Web at: www.nas.edu.

2. Protons in the accelerators hit a lead target at high speed and break apart lead nuclei, generating neutrons in the process.

3. See p. 431 of the report. The costs per kWhe are calculated by IEER from the reprocessing costs in the study by assuming that fuel is irradiated to a level of 40,000 MWDth/MTHM (megawatt-days thermal per metric ton of heavy metal) and that the efficiency of the nuclear power plant is 33%.