Fabricating fuels

Coated fuel particles are collected after carbon-containing gases are flowed up through a heated"funnel" into which uncoated particles are poured, in a process called fluidized-bed chemical vapor deposition. Click here for additional photos.

Over the past decade, increased public pressure to provide more electricity, reduce air pollution, and slow the rate of global warming has led many Americans to revisit the potential of nuclear power to meet anticipated demands for more energy. The Department of Energy and others in the scientific community are interested in adapting the gas-cooled reactor for use both in producing hydrogen for fuel cells to power cars and buildings and in supplying electricity competitively.

Studies indicate that these advanced reactors--which would operate at a much higher temperature than the water-cooled reactors that produce 20% of our nation's electricity--could convert 43% of their fuel-core heat into electricity, much higher than the 31% efficiency of today's reactors.

Oak Ridge National Laboratory is contributing to this effort by improving how coated-particle fuels are made. According to David Williams of ORNL's Nuclear Science and Technology Division (NSTD), these nuclear fuel particles--some 4 billion in each reactor--must be able to withstand the reactor's high temperature and neutron radiation from heat-generat-ing fission reactions.

"Regulations allow a fuel failure rate of only 1 in 100,000 particles," he says, explaining that maintaining the integrity of the particle coating is the key to avoiding fuel failure. "To ensure protection of workers and the public, only a small fraction of fuel failure is tolerated, to minimize the amount of radioactivity released to the coolant."

Perfect Spheres

The fuel being perfected for gas-cooled reactors would provide an additional level of safety. The meltdown-proof fuel particles would function as miniature reactors with their own containment. Each unit of fuel is a dark gray uranium bead shrouded in black carbon-containing coatings that trap and retain radioactive fission products, preventing their escape into the environment. The coated fuel par-ticles--as small as ballpoint pen balls-- would be compacted into fuel sticks embedded in a graphite block, which would moderate the neutrons and enable passage of the coolant, helium gas.

Since the early 1960s ORNL researchers have made important strides in developing and testing coated-particle fuel for high-temperature, gas-cooled reactors. Funding levels for this research that dropped after the 1979 accident at the Three Mile Island nuclear power plant have risen again. Drawing on four decades of experience, ORNL's modern researchers are taking advantage of state-of-the-art instrumentation, microwave technology, and computer modeling to create and produce efficiently the best possible coated fuel particles for advanced gas-cooled reactors.

In the late 1970s, ORNL's Milt Lloyd learned about a type of sol-gel process called internal gelation from its inventor, M. E. A. Hermans of the Netherlands. Lloydbrought back this knowledge to ORNL's sol-gel group. Today, chemist Jack Collins has become ORNL's expert on using the hexamethylenetetramine (HMTA) internal gelation process, which begins with black pellets of depleted uranium dioxide and ends with perfect black uranium dioxide beads. Collins and Rodney Hunt, both of NSTD, will eventually produce beads enriched to 20% in uranium-235.

In the early 1980s, Lloyd, Collins, Paul Haas, and others were interested in using internal gelation only to make different sizes of fuel spheres. "We are tailoring the chemistry to make beads of one size with the right density and smooth surfaces so the coatings won't have structural flaws," Collins says of their efforts today. "The goal is to determine a fail-safe formula that consistently produces the desired kernel product and a process that can be scaled up by the chemical engineers."

In the bead-making process perfected by Collins, pellets of uranium oxide are mixed with nitric acid to make acid-defi-cient uranyl nitrate, which is cooled, combined with a chilled HMTA and urea solution, and dispatched to an injector system. This chilled stream (0 o C) is dispersed into perfect drops of uniform size, with the help of controlled vibration, and is caught in a veil of silicone oil. The temperature difference causes the droplets to precipitate into perfect solid spheres and to flow with the oil without coalescing. The oil maintains the spherical shape of individual droplets.

As they leave the column, the gel spheres travel through a few yards of plastic tubing. The beads are then collected in a stainless-steel, wire-mesh basket and washed with tricholoroethylene (TCE), to remove the surface layer of silicone oil, and with dilute ammonia solution, to remove unwanted reaction products from the spheres. The spheres are heat-treated to form a dense ceramic "kernel" that is used as the starting point for the coating process. NSTD staff plan to work with Terry White of the Fusion Energy Division to use a microwave furnace to heat the chilled broth to form the gel spheres. In this way, the hot silicon oil and TCE steps can be eliminated.

Coatings and Characterization

During irradiation, the nuclear kernel will undergo fission, causing it to swell and give off fission products that span the periodic table, including radioactive gases. Each uranium kernel will be coated at ORNL to form a tiny pressure vessel.

Rick Lowden is in charge of coating the nuclear fuel kernels, John Hunn heads the group characterizing the fuel particles' coatings, and Peter Pappano compacts the coated particles into graphite fuel sticks that are inserted into the large holes in a hexagonal-graphite fuel element block; the small holes encircling the large ones allow the helium coolant to flow. All three researchers are with ORNL's Metals and Ceramics Division.

"Each fuel particle is coated with four layers, starting with an inner carbon buffer layer,followed by a pyrolytic carbon coating, a silicon carbide layer, and an outer pyrolytic carbon layer," Lowden says. "Each layer has itsown function. The buffer layer consists of porous carbon derived from a gas mixture containing acetylene, typically used in cutting and welding torches. The mixture produces a very soft, porous coating that accommodates fission product recoils from the kernel surface, provides a space for the fission gas released from the kernel, and accommodates kernel swelling without transmitting a force to the outer layer. The second "sealant" layer, which is made of hard, dense carbon, helps trap the fission products inside and protects the fuel kernel from chlorine generated during the deposition of the next coating.

"Because the silicon carbide layer does not change much during irradiation and is impervious to gaseous fission products, it serves as the primary structural component of this miniature pressure vessel. The silicon carbide also protects the inner layers from an accidental introduction of air--carbon will burn up in oxygen. Additionally, because ceramics are brittle and could be susceptible to damage during handling and compaction, another hard carbon layer is added to the outside to protect the silicon carbide layer."

The coatings are produced using a fluidized-bed chemical vapor deposition process first investigated at ORNL some 40 years ago. Uncoated or partly coated kernels are poured into a funnel inside a heated furnace. Fluidizing and reactant gases are flowed from the bottom up, "stirring up the particles, like the Ping Pong balls in a lottery machine," Lowden says. Different gas mixtures are used to deposit carbon and silicon on the fuel kernels.

"We are improving this process by incorporating advanced process monitoring and control techniques developed in other industries," Lowden explains. "Full automation removes human error from the process."

ORNL is also taking advantage of the computer models developed to examine other applications of fluidized beds, such as combustors or chemical digesters. "The combination of a well-controlled and instrumented furnace and computer modeling will help us to improve the process and, it is hoped, make better fuel," Lowden says. "This approach should also simplify the scale-up of the process."

John Hunn's team is tasked with characterizing the kernels and coated particles. Intimate knowledge of the microstructure and properties of the kernel and coatings is paramount to understanding the relationships among processing, product, and irradiation performance. Like Lowden, Hunn is exploiting recent advancements in materials characterization techniques and exploring new methods, to more fully understand the behavior of the various components.

"We're also looking at developing a higher-temperature fuel or fuel for reactors with a different neutron spectrum in which, for example, silicon carbide is replaced with zirconium carbide or titanium nitride," says Lowden. Says NSTD's David Williams: "ORNL will help improve the quality of particle fuels for all gas reactors by devising methods that can automate fuel inspection and production and increase the understanding of why a particular process yields the best product."