There are 104 commercial nuclear power reactors in the United States, which supply about 20 percent of the nation's electricity. These are light water reactors (LWR) fueled with low-enriched uranium (LEU), containing initially about 5 percent of the fissile isotope uranium 235. Each nuclear plant receives about 25 tons of LEU fuel annually, in the form of long pencil-thin rods of uranium oxide ceramic enclosed in thin metal "cladding", that are bundled together (in bunches of 300) to form fuel elements. Each year, nearly the same amount of spent fuel is removed from each reactor, but it's now intensely hot, both thermally and radiologically. In fact, even after five years of cooling in the "swimming pool" associated with each reactor, a fuel element would soon glow red-hot in the atmosphere because of the continuing radioactive decay of the products of nuclear fission. At this point, spent-fuel elements can be loaded into dry casks and stored at reactor sites on outdoor concrete pads with two casks added each year per reactor.
The country's long-term disposal plan has been that all spent fuel from U.S. reactors, plus some high-level waste from defense programs, would be entombed at the Yucca Mountain mined geologic repository in Nevada. But the senators from that state have opposed its opening, and President Barack Obama may have effectively killed the project by slashing funding and convening a blue-ribbon panel to find an alternative.
Some commercial interests argue that such spent nuclear fuel should be reprocessed (or "recycled," which is the industry's current term) into fresh fuel. They claim that this will greatly reduce the need for mined uranium and for underground repositories, and is, in any case, desirable--just as is all recycling of material such as paper, glass, aluminum, and steel. In reality, however, recycling spent nuclear fuel from U.S. reactors wouldn't solve any problems and would add additional cost and hazard.
In my congressional testimony, speeches, and published articles, I have provided technical details and abundant references to explain my opposition to reprocessing of LWR fuel back into fresh fuel as is practiced in France and a few other countries. France has decades of experience in technically successful reprocessing of its LWR spent fuel. It currently obtains one usable mixed-oxide (MOX) fuel element, a mixture of plutonium and uranium oxides, from every seven LWR spent fuel elements. But aside from the almost 1-percent plutonium in the spent fuel and the 94-percent uranium 238, the other 5 percent of mass (called "fission products") is removed and melted together with glass into a vitrified product that is encased in welded stainless steel canisters. These are stored at the French reprocessing plant at La Hague, awaiting the availability some decades hence of a mined geologic repository. In fact, for all the U.S. delays and roadblocks, it's far ahead of France in planning for a permanent repository.
In truth, reprocessing doesn't eliminate or even significantly reduce the need for a repository, as demonstrated by the authoritative presentations of Idaho National Laboratory Associate Director Phillip J. Finck, who has worked in the French program and is now in charge of a major portion of the U.S. government nuclear energy research program. According to Finck, Yucca Mountain could accommodate only about 10 percent more spent MOX fuel and vitrified fission products as produced at La Hague than it could normal spent fuel. This is because after four years in a reactor MOX fuel is much hotter than normal spent fuel, and so fewer spent MOX fuel elements can be accommodated in the same space as ordinary and cooler spent fuel (also called UOX, for its uranium-oxide content).
What's the long-term plan then? Well, there is no plan, but there are proposals. One of the more foolish was the Global Nuclear Energy Partnership (GNEP), advanced by the Bush administration and abandoned by Obama. It planned on building a U.S. reprocessing plant to recycle spent UOX into MOX fuel not for light water reactors, but for a new generation of fast reactors that would, if they were built, burn up the plutonium and the so-called higher actinides such as neptunium, americium, and curium in spent fuel. This would have had the advantage of reducing the long-persisting decay heat from spent MOX, which would make it simpler to store it in an underground repository.
But because the fuel for these specialized fast reactors would have to be reprocessed many times over, the cost would have made the GNEP approach uneconomical as a waste-treatment option. GNEP proponents also never specified what form they would store the hottest and most hazardous fission products--strontium 90 and cesium 137--each of which has a roughly 30-year half-life and would need to be stored for 300 years.
Reprocessing of LWR fuel also fails to save uranium, a common argument in favor of recycle. Although 1 percent of the fuel is plutonium and can be burned as MOX; recycling all LWR fuel, including reuse of uranium, would save at most 20 percent of the necessary supply of raw uranium ore. Analysis shows this isn't worth doing unless the cost of natural uranium rose to something like $750-$1,000 per kilogram. Its current price, however, is much lower, on the order of $70 per kilogram. Even at a price of $750 per kilogram, reprocessing would only be marginally preferable.
Although the French experience with reprocessing has been technically successful, if costly, the British experience has been a failure. Britain built a government-operated plant at Sellafield, but it hasn't performed reliably and the physical fabrication of MOX for sale to Japan and other customers has been stymied by Sellafield's failure to produce MOX fuel pellets to specification. Furthermore, Britain has no plans to use MOX in its own reactors. France plans to hold its spent MOX fuel until a currently unplanned fleet of fast "breeder" reactors are built that, in principle, would produce (or "breed") more plutonium than they burn. At that time, the MOX could be reprocessed to make fuel that would have 10-20 percent plutonium, rather than the typical 5-7 percent plutonium content necessary for use in LWRs. Unfortunately, there is no successful commercial experience in the world with plutonium-fueled fast reactors. France operated a 250-megawatt-electric fast reactor named Phénix, and, briefly, a commercial scale 1,250-megawatt-electric version named Superphénix, but the latter operated only a tiny fraction of the time and has been dismantled.
My own view is that if plutonium-fueled fast reactors could be demonstrated to be less costly than the typical LWR "burner" reactor, and if it could be demonstrated to be just as safe, then recycling of nuclear fuel would make sense. One would still need mined geologic repositories for the spent fission products and other high-level wastes, however.
To help, I have proposed a world breeder reactor laboratory that would focus on improving computer simulations of fast-reactor designs and safety analysis; so that one could more confidently simulate and analyze not only normal operation but also potential accidents that can't be experimentally investigated. After perhaps 20 years, a thorough cost and safety analysis of fast reactors compared with LWRs might be possible. And if a convincing demonstration by simulation were achieved, then it might be time to go ahead and build a prototype reactor to confirm the simulation.
Whether or not fast reactors eventually are economical and safer than LWRs, more LWRs should be built if they can compete economically with fossil fuel-burning plants at costs that include an appropriate carbon tax. The spent LWR fuel should be stored at the reactor sites or in centralized storage locations in standard dry casks. This will suffice for 100 years or more, while the world addresses more sensibly the provision of suitable, mined long-term geologic repositories. If fast reactors become safer than LWRs, and economically competitive with them as well, the world will be happy not to have wasted its plutonium by burning it once in an LWR rather than essentially forever in a fast reactor.