At the dawn of the nuclear era, it was generally envisioned that the first crop of commercial reactors would soon be replaced by more advanced fast breeder reactors, glorious machines that would create more new fuel than they burned. Integral to the breeder vision was the reprocessing of spent fuel to extract the unused uranium and plutonium, thereby conserving resources that researchers at the time believed (incorrectly) to be scarce. The United States and other countries poured billions of dollars and decades of R and D into developing a viable breeder—largely ignoring other reactor types—but precious little came of it. The technology proved just too costly and accident-prone. By the mid-1990s, many industry watchers assumed the idea was dead.

Photo: Argonne National Labortory

Breeder Obsession: The DOE's nuclear energy R and D program is fixated on spent fuel reprocessing and breeder reactors (like the decommissioned Experimental Breeder Reactor-II shown here), despite concerns about cost, safety, security, and feasibility.

It's not. Over the last three years, the U.S. Department of Energy (DOE) has quietly revived its support for breeder R and D. The stated aim of its Generation IV program is to develop advanced reactor designs to replace today's light-water reactors. But three of the six Gen IV designs are fast reactors; a fourth is a thermal reactor, which also would rely on reprocessing. The remaining two designs would be able to operate on reprocessed fuel, but wouldn't require it. Gen IV's companion program, the Advanced Fuel Cycle Initiative (AFCI), is meanwhile looking at new ways to reprocess spent fuel. The work is overseen by the Idaho National Engineering and Environmental Laboratory, in Idaho Falls, home to the DOE's breeder fuel testing and reprocessing research during the 1970s and 1980s.

Much more than just lab curiosities, these programs are seeking bona fide commercial-scale designs, ready to be deployed by the year 2030. In fiscal year 2004, the programs are funded at a total of US $83 million, and could receive much more in coming years if Congress passes the comprehensive energy bill in its current form.

What's wrong with this picture? Plenty. The chief concern with recycling spent fuel is the proliferation risk posed by the separated plutonium. A terrorist or errant government that manages to steal a few kilograms would skirt the toughest hurdle in making a bomb—getting the material itself. For that reason, in fact, the U.S. government halted civilian spent-fuel reprocessing in 1977. All commercial U.S. nuclear plants currently operate on the so-called once-through cycle, directly disposing of spent fuel.

Ignoring that history, the new DOE effort not only encourages reprocessing; it aims to share the technology with nine other countries that now participate in the Gen IV International Forum. By cultivating the considerable expertise needed to do reprocessing, "these programs would certainly assist non-nuclear weapons states in developing a capability to rapidly produce large quantities of nuclear weapon-usable materials," says Thomas Cochran, director of the nuclear program at the Natural Resources Defense Council, in Washington, D.C.

The economics of reprocessing don't add up, either. For one thing, as the physicists Richard L. Garwin and Georges Charpak point out in their book Megawatts and Megatons, the price of uranium would have to rise to $700 per kilogram, from its current $30/kg, to make reprocessing economical. But the supply of uranium, including that in seawater, is nearly inexhaustible. Garwin, though a die-hard supporter of nuclear energy, calls the current emphasis on reprocessing "premature and really dumb and counter-productive, in my considered view."

Meanwhile, British Nuclear Fuels Ltd., in Daresbury, UK, has announced that it will shut down its commercial reprocessing plant in 2010, citing a lack of contracts for its services. Even France, which is often cited as having the most successful reprocessing program, would have saved 164 billion francs (US $30 billion) over an assumed 45-year life of its reactors had it opted for direct disposal of spent fuel, according to a high-level French government report issued in 2000.

Doe's Nuclear Energy R and D Program

Goal: Develop and deploy by 2030 a new fleet of advanced nuclear power reactors

Why it's a Loser: Rather than focus on alternative reactor designs, the U.S. Department of Energy has resuscitated the problem-plagued breeder program. These reactors rely on reprocessing spent fuel, which would be hugely uneconomical, pose a proliferation risk, and may not even work when scaled up to commercial levels

Organization: Idaho National Engineering and Environmental Laboratory (INEEL), Argonne National Laboratory, and other U.S. Department of Energy and academic labs, plus counterparts in nine other countries

Center of Activity: INEEL, in Idaho Falls, Idaho

Number of People on the Project: At least 300

Budget: US $83 million in FY2004

To be sure, the reprocessing techniques being touted by Gen IV and AFCI differ somewhat from the World War II-era technique used in France, the UK, and elsewhere in that pure plutonium is never separated from the spent fuel. Instead, the plutonium remains mixed with small but highly radioactive amounts of neptunium, americium, and curium, collectively known as minor actinides. This feature has led some to call the modified reprocessing techniques proliferation-resistant.

Marvin Miller, a retired nuclear engineer at the Massachusetts Institute of Technology (MIT) in Cambridge, Mass., calls that characterization misleading. He looked at the risk posed by pyroprocessing, one of the AFCI techniques being considered, and found that though it's safer than current methods, it poses a greater proliferation risk than the once-through cycle. "If you have a pyroprocessing plant, how easy is it to jiggle with the knobs to get a cleaner plutonium product, and can that be detected?" Miller says. "Those are still open questions."

Miller also wonders whether the technology can really work at the commercial level. "It's been demonstrated that you can separate and recycle plutonium as nuclear fuel. But it's never been demonstrated that you can cleanly separate and recycle these minor actinides," he says.

Last spring, a study group convened by John M. Deutch and Ernest J. Moniz, professors of chemistry and of physics, respectively, at MIT, considered what it would take to make nuclear power a "significant option" for reducing greenhouse gases. They assumed worldwide nuclear generating capacity would have to grow almost threefold, to 1000 gigawatts, by the year 2050, and then looked at the best way to get there, factoring in cost, safety, waste, and proliferation.

Their conclusion? Stick with what you know. "For the next decades, government and industry in the U.S. and elsewhere should give priority to the deployment of the once-through fuel cycle, rather than the development of more expensive closed-fuel-cycle technology involving reprocessing and new advanced thermal or fast reactor technologies," their report, The Future of Nuclear Power, states. Lest there be any doubt, they add, "This recommendation implies a major re-ordering of priorities of the U.S. Department of Energy nuclear R and D programs."

Some of the Gen IV reactor designs "strain credulity," Moniz told IEEE Spectrum. Reprocessing spent fuel may reduce waste management problems in the very long run—say, a thousand years out, he said. "We nevertheless find that argument not compelling when traded off against [recycling's] near-term problems."

Whether electric utilities will ultimately agree to build and operate the new reactors is yet another matter. Absent enormous government subsidies, it's safe to say they won't: no nuclear plants have been ordered in the United States since 1978, no units are under construction, and plans for more than 100 reactors have been canceled.

"The utilities have a hard enough time running standard nuclear plants," Miller notes. "These new technologies are too sophisticated, they're uneconomical, and they don't have the proliferation-resistance of the once-through cycle....the U.S. is spending a lot of money on these technologies, but it all looks very dubious."