queues

Some researchers see the answer to the safety problem in revolutionary reactor designs that promise to be more "inherently safe"—physically incapable of suffering a catastrophic meltdown. One such design, at least in theory, is the Pebble Bed Modular Reactor, being developed in China and South Africa. It's powered with balls of uranium-filled graphite rather than the typical fuel rods. If the cooling system were to fail, the reactor temperature stays well below the balls' melting point and then automatically cools down.

Westinghouse is working with the Department of Energy toward the possibility of getting a design certified by the NRC by 2017 or so. China currently has a small prototype pebble-bed reactor and plans to start construction this year on a 200-megawatt plant using the technology.

Most industry observers think the design is intriguing but faces big hurdles in this country because it uses a gas coolant, instead of water, and different fuel. The NRC would have to develop special processes for reviewing such a design because its expertise is in pressurized water or boiling water reactors.

Exelon Corp., which operates 17 commercial reactors in the U.S., was interested in the pebble-bed reactor in the late 1990s but is no longer involved. "There were technical problems such as fuel issues that made us decide we didn't want to proceed," says Amir Shahkarami, senior vice president of nuclear generation at Exelon.

While safety may be nuclear power's biggest PR problem, cost is what killed development a generation ago, ultimately determining that only half the plants licensed by the NRC got built. And nuclear plants generally face an unfortunate trade-off: making them more safe can make them more expensive.

Makers of Generation III models are addressing the cost issue in a number of ways. For one, they claim the reactors will remain in service more years, so construction costs will be spread over a longer operating life. Today's plants are being designed to last at least 60 years—longer than any other plants except hydroelectric dams. Existing nuclear plants were expected to be retired after 40 years, though roughly half have gotten 20-year license extensions.

The new plants are also designed to be much simpler and quicker to build, reducing financing costs by potentially hundreds of millions of dollars. For instance, there's the ABWR reactor, which has been built in Japan by GE-Hitachi and which NRG Energy Inc. hopes to build with Toshiba's help in South Texas. The reactor is built in modules, vastly speeding construction time. GE-Hitachi, a joint venture of General Electric Co. and Hitachi Ltd., says it has built the plant in 42 months in Japan, which is more than twice as fast as the Generation II reactors it built in the 1980s. The company compares construction methods to putting up a modular home versus constructing a stick-built house.

NRG hopes to build two ABWR reactors in Texas, next to its existing South Texas Project nuclear plant. Each plant will employ 190 modules, which NRG believes will cut field labor costs by 30%. Faster construction also will reduce the length of time it will have to rent a heavy crane at $400,000 a month.

Still, nuclear plants will remain very expensive. Recent estimates put Generation III plant costs at $4,000 to $6,700 per kilowatt of capacity, or $4.4 billion to $11 billion, for plants ranging from 1,100 megawatts to 1,600 megawatts in size. In comparison, a recent Massachusetts Institute of Technology study estimated the price of a coal plant at about $2,300 a kilowatt of capacity and a gas-fired plant at about $850 a kilowatt of capacity.

In fact, only a handful of U.S. utilities are big enough to build Generation III reactors alone, without being part of a consortium. As a result, some see nuclear power's future in small reactors that could be manufactured in factories instead of on site—and cost only $3,500 to $5,000 per kilowatt of capacity, or millions of dollars instead of billions.

Babcock & Wilcox, a unit of McDermott International, has designed a small 125-megawatt reactor that would be built at its U.S. factories and then delivered to power-plant sites by rail or barge. This would eliminate a bottleneck—and the associated higher costs—for ultra-heavy forgings that are required for large reactors. Small reactors could be built at a number of domestic heavy-manufacturing sites. The Lynchburg, Va., company has been building small reactors and other key components for Navy ships for decades, at plants in Indiana and Ohio.

Another plus of small reactors: They're designed to be refueled less frequently, reducing the number of refueling outages. Instead of every 18 months to two years, they could go four or five years, reaping a saving from having less down time. Another feature of some reactors is the ability to do more maintenance while plants are running, again reducing idle time.

Babcock & Wilcox hopes to apply for certification of a design for its mPower modular reactor in 2011. It's too early to seek orders, but it's working with Exelon and the Tennessee Valley Authority on a preliminary design, to make sure it would meet the needs of utilities. It's unlikely any could be built in the U.S. before the middle of the next decade.

Critics say there's not enough practical experience to know if any of the U.S. designs, big or small, will function as proponents say. Only one, the ABWR, has completed the review process at the NRC and completed the detailed design that would be used as the basis of actual construction.

What's more, critics say that the economics of small plants simply don't work: The licensing costs are so great for nuclear plants, somewhere between $50 million and $100 million per site, and security and construction costs are so high that the economics work only for big plants, with lots of output, so costs can be spread over many kilowatt-hours of electricity. Proponents hope factory-like construction techniques and a wider availability of suitable sites will help them overcome that drawback.

It's one of the most contentious issues surrounding nuclear power: Where do you put the spent fuel?

Tens of thousands of metric tons of nuclear waste—mainly spent fuel rods—are sitting at power-plant sites while the federal government struggles to come up with a site to store it all. No nation has yet built a permanent waste site, although the current situation can continue for some time: Even critics say storage methods in place now should allow fuel to be stored safety for 50 to 100 years, while permanent plans are worked out.

The big problem with controlling waste: Today's reactors capture only about 5% of the useful energy contained in uranium—which means lots of radioactive leftovers once the fuel is used. Some Generation III reactors promise to address this problem by squeezing more electricity out of their fuel, reducing the total amount of waste produced, but it's only by a relatively small amount. In short, it does nothing to solve the looming waste issue, though it does produce more megawatts of electricity in the short run.

Some Generation IV reactors, known as fast reactors, may offer a breakthrough in the future—because they're designed to burn previously used fuel.

GE-Hitachi, for example, is developing a fast reactor called Prism that would take spent fuel or weapons waste, sitting in storage today, and use nearly all of it as fuel, leaving little waste. What's left would also be less radioactive than current waste, and would need to be stored for hundreds of years instead of thousands of years, scientists say. Fast reactors are able to unlock energy in waste because they can burn plutonium, neptunium and other materials that Generation II and Generation III reactors leave behind.

GE-Hitachi estimates there's enough energy sitting in nuclear storage sites in the U.S. to completely meet the nation's energy needs for 70 years, if fast reactors were used to convert waste into electricity.

The company hopes to apply for NRC certification of its Prism design in 2011 and build a prototype reactor at an estimated cost of $3.2 billion within the next decade. The cost is enormous for a reactor that would be only 311 megawatts in size, amounting to $10,000 per kilowatt of capacity, but the company says costs for subsequent units should drop.

Critics point out that the U.S. tried to develop fast reactors in the past, but dropped its efforts because the technical hurdles and cost appeared too great. The NRDC, in a recent report, said that fast reactors would be "expensive to build, complex to operate, susceptible to prolonged shutdowns…and difficult and time-consuming to repair."

--Ms. Smith is Wall Street Journal staff reporter in San Francisco. She can be reached at rebecca.smith@wsj.com.