MIT Technology Review

Tuesday, March 12, 2013

Safer Nuclear Power, at Half the Price

Transatomic is developing a new kind of molten-salt reactor designed to overcome the major barriers to nuclear power.

Transatomic Power, an MIT spinoff, is developing a nuclear reactor that it estimates will cut the overall cost of a nuclear power plant in half. It's an updated molten-salt reactor, a type that's highly resistant to meltdowns. Molten-salt reactors were demonstrated in the 1960s at Oak Ridge National Lab, where one test reactor ran for six years, but the technology hasn't been used commercially.

The new reactor design, which so far exists only on paper, produces 20 times as much power for its size as Oak Ridge's technology. That means relatively small, yet powerful, reactors could be built less expensively in factories and shipped by rail instead of being built on site like conventional ones. Transatomic also modified the original molten-salt design to allow it to run on nuclear waste.

High costs, together with concerns about safety and waste disposal, have largely stalled construction of new nuclear plants in the United States and elsewhere (though construction continues in some countries, including China). Japan and Germany even shut down existing plants after the Fukushima accident two years ago (see "Japan's Economic Troubles Spur a Return to Nuclear" and "Small Nukes Get Boost"). Several companies are trying to address the cost issue by developing small modular reactors that can be built in factories. But these are typically limited to producing 200 megawatts of power, whereas conventional reactors produce more than 1,000 megawatts.

Transatomic says it can split the difference, building a 500-megawatt power plant that achieves some of the cost savings associated with the smaller reactor designs. It estimates that it can build a plant based on such a reactor for $1.7 billion, roughly half the cost per megawatt of current plants. The company has raised $1 million in seed funding, including some from Ray Rothrock, a partner at the VC firm Venrock. Although its cofounders, Mark Massie and Leslie Dewan, are still PhD candidates at MIT, the design has attracted some top advisors, including Regis Matzie, the former CTO of the major nuclear power plant supplier Westinghouse Electric, and Richard Lester, the head of the nuclear engineering department at MIT.

The new reactor is expected to save money not only because it can be built in a factory rather than on site but also because it adds safety features—which could reduce the amount of steel and concrete needed to guard against accidents—and because it runs at atmospheric pressure rather than the high pressures required in conventional reactors.

A conventional nuclear power plant is cooled by water, which boils at a temperature far below the 2,000 °C at the core of a fuel pellet. Even after the reactor is shut down, it must be continuously cooled by pumping in water. The inability to do that is what caused the problems at Fukushima: hydrogen explosions, releases of radiation, and finally meltdown.

Using molten salt as the coolant solves some of these problems. The salt, which is mixed in with the fuel, has a boiling point significantly higher than the temperature of the fuel. The reactor has a built-in thermostat—if it starts to heat up, the salt expands, spreading out the fuel and slowing the reactions. That gives the mixture a chance to cool off. In the event of a power outage, a stopper at the bottom of the reactor melts and the fuel and salt flow into a holding tank, where the fuel spreads out enough for the reactions to stop. The salt then cools and solidifies, encapsulating the radioactive materials. "It's walk-away safe," says Dewan, the company's chief science officer. "If you lose electricity, even if there are no operators on site to pull levers, it will coast to a stop."

The new design improves on the original molten-salt reactor by changing the internal geometry and using different materials. Transatomic is keeping many of the design details to itself, but one change involves eliminating the graphite that made up 90 percent of the volume of the Oak Ridge reactor. The company has also modified conditions in the reactor to produce faster neutrons, which makes it possible to burn most of the material that is ordinarily discarded as waste. A conventional reactor produces about 20 metric tons of high-level waste a year, and that material needs to be stored for 100,000 years. The 500-megawatt Transatomic reactor will produce only four kilograms of such waste a year, along with 250 kilograms of waste that has to be stored for a few hundred years.

Bringing the new reactor to market will be challenging. Although the basic idea of a molten-salt reactor has been demonstrated, the Nuclear Regulatory Commission's certification process is set up around light-water reactors. The company will need the NRC to establish new regulations, especially since the commission must sign off on the idea of using less steel and concrete if the design's safety features are to lead to real savings.

NRC spokesman Scott Burnell says that the commission is aware of Transatomic's concept but that designs haven't been submitted for review yet. He says that for the next few years, the NRC will be focused on certifying more conventional designs for small modular reactors. He says the certification process for Transatomic will take at least five years once the company submits a detailed design, with additional review needed specifically for issues related to fuel and waste management.

A detailed engineering design itself may be years away. The company's next step is raising $5 million to run five experiments to help validate the basic design. Russ Wilcox, Transatomic's CEO and the former CEO of E Ink, estimates that it will take eight years to build a prototype reactor—at a cost of $200 million. He says that's less time than it took investors to get a return on E Ink, which was acquired for $450 million 13 years after the first investments in the company.

Even though it could take well over a decade for investors to get a return, venture funding isn't out of the question, Ray Rothrock says. But he says the company will face many challenges. "The technology doesn't bother me in the least," he says. "I have confidence in the people. I wish someone would build this thing, because I think it would work. It's all the other factors that make it daunting."

The company's biggest challenge might come from China, which is investing $350 million over five years to develop molten-salt reactors of its own. It plans to build a two-megawatt test reactor by 2020.