In This Issue
Competitive Materials and Solutions
December 1, 1998 Volume 28 Issue 4
The Bridge, Volume 28, Number 4 - Winter 1998

Materials Challenges for Fusion Energy

Tuesday, December 1, 1998

Author: Steven J. Zinkle

Materials R&D will play a major role in determining fusion's future viability as an energy-production method.

Fusion - the forceful joining of light atoms - is the process that operates our sun. It is a nuclear process, and heat is one of the by-products. Of course, under normal circumstances, atoms don't want to come together. In the sun, gravity is the force that causes the fusion reaction. Here on Earth, there are two different approaches for achieving the same result. One uses magnetic fields to compress the reacting particles together; the other uses laser or ion beams. These techniques are called magnetic and inertial confinement, respectively.

There are about 10 different feasible fusion reactions. The deuterium-tritium (DT) reaction is the one that has received the most attention, in part because it is the easiest to make work, and also because it produces quite a bit of energy. Deuterium is readily available in water as a naturally occurring isotope of hydrogen. Tritium has to be produced from lithium, but there is a relatively abundant supply of this element.

One of the most difficult challenges of the fusion reaction is dealing with neutron radiation. Neutrons produced during fusion can travel tens of centimeters out into the containment structure, causing damage to the constituent materials. I'll talk more about this in a minute.

The big advantage of fusion compared with fossil-fuel-based energy production is its relatively small fuel requirements. For the same amount of energy, fusion requires about six orders of magnitude (~106) less fuel compared with chemical energy sources (coal, oil, etc.). A convenient way to think about this is to consider that an ordinary cup of tap water contains the energy equivalent of a full tank of gasoline in an automobile. That is, the approximately one drop of heavy water in that cup could, through fusion, provide as much energy as 20 gallons of gas.

Plasma physicists often point out that essentially all of our universe is made up of plasma, (i.e., the interstellar plasma). It is a relatively cold plasma, about 10,000?C. By contrast, the heat of a fusion reaction is in the range of hundreds of millions of degrees. The atomic densities required for fusion, particularly for magnetic fusion, are relatively low compared with the density of air. In other words, the power density (watts per unit volume) in a fusion reactor would be comparable to fossil fuel power plants, even though the fusion plasma operates at a much higher temperature. Achieving higher power densities in fusion reactors requires the development of as-yet-unknown materials or confinement concepts, due to the high heat fluxes which would occur at the plasma-facing components of the reactor. The most popular current technique to create the fusion reaction relies on magnetic confinement using a donut-shaped, or toroidal, device to contain the plasma.

So, how close are we to a demonstration of fusion as a viable energy-production method? The standing joke among those working in the field is that it's always 25 years or more to reach the demonstration condition. This is one of the most frustrating things about working in fusion energy: The 25-year horizon seems always to be creeping away as we move forward in time. But, if you really look at what the plasma physicists have done, it is quite remarkable. Going back to the infancy of the fusion energy program in the 1950s, there has been a nearly constant rate of advancement.

The first goal has been to reach what's called the break-even condition, which occurs when the amount of energy used to heat up the plasma is equal to the amount of energy that is produced by the fusion reaction. The latest machines have recently crossed this threshold for a DT fuel mixture. That's an important step, but it's obviously not good enough, because there's no net energy gain. The more important goal for commercialization is ignition, or the creation of a self-propagating reaction due to plasma heating from the charged particles produced by the fusion reaction. In this state, there is no need for external power. The ignition condition for magnetic fusion reactors occurs when the fusion power is about 10 times the input power.

The "Triple Product"
One way to measure progress toward these goals is to calculate the "triple product," which is the product of the density of the ionized electrons (ne), the length of their thermal confinement, and the plasma ion temperature (Ti). On average over the past 30 years, the doubling time of the triple product has been about every 3 years. It slowed a bit in the past decade, to about every 5 years. This rate of performance improvement is not quite as impressive as that in the semiconductor industry, but it is still quite remarkable. The triple product must increase by about another factor of 10 in order to reach the regime where a commercial fusion power plant would be expected to operate.

Fusion energy research is now at a crossroads. Through advances in the plasma physics, we can say with some confidence that the concept is feasible; plasmas can be made with sufficient density and temperature to achieve the break-even condition, both with inertial confinement and magnetic confinement. The focus is now shifting to developing a commercial-scale machine (Figure 1), and engineering is starting to play a more central role in this regard. Unresolved questions include whether the toroidal shape is the best reactor design and whether inertial confinement is the optimal way to contain the reaction. Materials technology will have a major impact in influencing what path will be taken.

There are numerous materials involved in a fusion reactor, ranging from electrical insulators to superconducting magnets. For DT fuels, the tritium is produced from lithium in what is called a blanket region. The blanket region has to be cooled, with the extracted heat going to the power-conversion system (e.g., a steam cycle). And, of course, there needs to be some structure surrounding all of this.

The environment for fusion is a rather unique performance arena and poses a number of challenges. A DT reactor, for example, will experience unprecedented neutron radiation damage, on the order of 2 to 10 times greater than what core internal structures of existing fission reactors must cope with. In order for fusion to be economically competitive and acceptable to the public, the reactor structure must be made of low-activation materials (i.e., materials that will not become highly radioactive). There will also be quite high particle- and heat-flux conditions, and high operating temperatures will cause thermal creep. Unfortunately, in a lot of cases, materials developed for other technologies can't be straightforwardly applied to fusion energy.

I will focus now on three specific areas where materials impact fusion reactor design: the plasma-facing region, where there is high heat flux and particles are impacting the metal structure; the plasma-diagnostic, heating, and magnet systems; and the structure of the blanket and first-wall region surrounding the plasma, which is the heart of the heat-extraction system.

Heat fluxes for fusion energy cover a wide spectrum. At the extreme end, when there is a disruption condition, the fluxes can be very high (about 103-104 MW/ m2). Fortunately, these occur for only a millisecond or so. Under steady-state conditions, the heat fluxes (about 1-10 MW/m2) are comparable to what you might find in a rocket nozzle application, or a bit higher than what occurs in a fission reactor.

Because the cost for fusion energy is going to be largely driven by the size of the plant, the reactor should be made as compact as possible. However, as you start decreasing the size, the heat flux on the first wall increases. Straightforward analyses reveal the maximum steady-state heat flux to be about 1-10 MW/m2 for a reasonable wall thickness (about 5 mm) in a fusion reactor, depending on the material.

Most available structural materials are limited by having rather low thermal stress resistance. In some special cases, copper alloys for instance, you can achieve high heat fluxes at room temperature, but copper alloys don't have high-temperature capability. For example, copper-nickel-beryllium alloy (Cu-2 percent, Ni-0.3 percent) is in many respects an outstanding material. It can be cast and aged, has a room temperature strength about twice that of steel, and has 70 percent of the thermal conductivity of pure copper. But, if you go up in temperature, above 300?C, it becomes very brittle. So, it is not a feasible material for use at high temperatures. Virtually all copper alloys suffer from either low strength or low-fracture toughness at elevated temperatures.

Carbon-carbon composites have been used in several plasma physics machines, as well as in numerous nonfusion high-heat-flux applications. But, in situations where neutron damage starts to become more important, they are unsuitable because the radiation damage quickly degrades the thermal conductivity.

Another technique that is being looked at is to put a fast-flowing liquid coolant on the first surface. This would be a very efficient way to take away the heat, but there are a number of technological challenges. For example, it's hard to see how this could be applied to a toroidal geometry. This approach is being seriously considered in inertial confinement schemes, where there is a more suitable reactor shape.

Another of the key factors for plasma-facing materials is that neutral particles coming out of the plasma can sputter the surface atoms. In addition to erosion concerns, these sputtered atoms contaminate the plasma, since they absorb heat from the plasma fuel but do not fuse. The plasma power loss is proportional to NiZi, where Ni is the number of sputtered impurity atoms and Zi is the corresponding atomic number.

If you look at the sputtering behavior of various materials at fusion-relevant conditions (10-1,000 eV hydrogen ion energies), stainless steel is one of the worst possible plasma-facing materials. It has a high sputtering yield (probability of a surface atom being sputtered into the plasma by an incident hydrogen ion) and a high atomic number, making it impossible to use in the first wall of a fusion reactor due to unacceptable plasma power losses. The most attractive plasma-facing candidates at the present time are either beryllium or carbon, because of their low atomic number, or something like tungsten, because of its low sputtering yield.

Reactor Ports a Success Story
One example of a success story is in the materials used to make the ports through which the energy beams are sent to heat up the plasma. The windows are needed in order to avoid atmospheric contamination of the plasma. One heating technique, electron cyclotron heating, uses high-frequency (about 140 GHz) radio waves. Up until about 5 years ago, engineers were limited in terms of the power density they could get through this window, since there was no available material that could withstand a high-power radio beam.

Recently, there have been some dramatic advances in chemical vapor deposition (CVD) diamond. The cost has dropped by more than an order of magnitude in the last 5 years and is expected to drop by another order of magnitude in the next 2 years with the construction of large-scale production plants. At the same time, the quality of the CVD diamond films has dramatically improved (in part due to the application of plasma technologies in the CVD process). Now, it is possible to purchase 46-inch-diameter free-standing diamond wafers that can be put in these beam ports.

Radiation can also have a negative effect on insulators, which are essential for monitoring and heating the plasma in a fusion reactor. If you expose any insulating material to an ionizing field, its electrical conductivity increases in a linear fashion. Conductivity increases of over 10 orders of magnitude have been observed at high dose rates. Fortunately, insulators such as aluminum oxide retain enough of their insulating properties to be useful for fusion applications. A few years ago, there was concern that nobody had ever taken one of these insulators and exposed it for a long time to an electric field during irradiation to simulate in-service conditions. Some electron-beam studies suggested insulators would suffer rapid electrical breakdown. If this were to happen, the fusion reactor would not operate.

We recently performed some experiments where we put insulators into a fission reactor and made in-situ measurements with a fusion-relevant electric field (200 V/mm) during irradiation over a period of several months. We found that there was no permanent degradation. The previously reported electrical breakdown is now considered to be an artifact of electron irradiation, and it doesn't occur under neutron-irradiation conditions. So, as far as the electrical insulators are concerned, we're in pretty good shape.

The structural materials used in a fusion reactor must address a number of concerns. They need to have a high thermal-stress capability. They have to be compatible with the coolant, which may be a liquid metal such as lithium or a gaseous coolant (He). They should be passively safe under accident conditions, environmentally friendly, and able to resist radiation damage. There are only a few structural materials that can adequately address all of these concerns. The three main candidates are ferritic martensitic steel, vanadium-based alloys, and SiC-fiber-reinforced SiC composites.

Many high performance structural materials used in other industries are not suitable for fusion applications. For example, titanium alloys used for aerospace applications have a very high solubility for hydrogen. This results in a lot of tritium leaving the plasma and entering the structure, creating an unsafe condition. (If there were an accident, the tritium could be released into the atmosphere.) Nickel-based superalloys are great materials in many respects, but when subjected to radiation, they have significant grain-boundary embrittlement problems.

Radiation can produce large changes in structural materials. At low temperatures (less than 0.3 Tm, where Tm is the melting temperature), the main concern is radiation hardening and embrittlement. As you go up in temperature, there is a phenomenon called radiation creep, which acts on top of thermal creep and can limit the amount of stress that can be put on the structure. Volumetric swelling is a significant concern for certain materials at intermediate temperatures (0.3-0.6 Tm). And, at very high temperatures (>0.45 Tm), there can be pronounced helium embrittlement at grain boundaries. So, the radiation environment in a fusion reactor is quite a bit more severe than it is for structural materials in existing fission reactors, and the challenges for materials scientists are also greater. Studies performed to date on the three main candidate structural materials have shown promising resistance to several types of radiation-induced degradation, but considerable work still needs to be done to determine their suitability for fusion reactors.

To summarize, on the physics side of things, producing energy by fusion looks very promising. We are at the break-even point. The investment that has been made to date is all from government sources. Utilities won't be interested in supporting this technology until it is really close to the point of being commercialized. The total federal expenditures on fusion research from 1950 to date is on the order of $10 billion, or less than 5 percent of what we spend each year on gasoline in the United States. The estimated value of fusion spin-off technologies is higher than the fusion research expenditures, with the plasma-processing-technologies market alone valued at over $200 billion per year (plasma processing of semiconductors and ceramics coatings on materials, waste processing, plasma electronics, etc.).

There are a number of questions that remain. What is the best pathway to a commercially attractive power plant? Should it be magnetic confinement or inertial confinement? Can fusion become cost competitive with coal and fission? Clearly, materials R&D will play a major role in determining the fate of fusion energy.

About the Author:Steven J. Zinkle is a senior research staff member at Oak Ridge National Laboratory, Oak Ridge, Tennessee. This paper is adapted from remarks he made 6 October during the symposium, Materials - The Opportunity, part of the 1998 NAE Annual Meeting.