The 21st century may one day be known more than anything else for the period when human beings transitioned from fossil fuels to clean renewable energy. Such a transition, as we know even now, is crucial to sustaining our physical environment and to supporting the growth of human civilization.
Debates have been raging around the role that nuclear fusion might play in this transition. Last February, the JET reactor in the United Kingdom broke the record for the amount of energy produced per pulse; last year, the experimental advanced superconducting tokamak (EAST) in China broke the record for highest plasma temperature achieved in a tokamak. Such developments stir up enormous excitement about a potential epoch-making breakthrough in fusion technology. Media reports and press releases on individual developments in fusion, however, often fail to provide a bird’s eye view of the field, exaggerating progress or selling fusion as a magical, “unlimited” source of energy not bound by the engineering or economic limitations of other forms of energy.
The truth is there are multiple approaches to fusion simultaneously being pursued, and they all have advantages and disadvantages. To get a more accurate idea of where we are in a technological field so crucial to the future of humanity, it helps to review some of the scientific and economic fundamentals at play before making any predictions about the future.
Fusion reactions are thermonuclear reactions, which means they occur at the level of the nucleus—unlike chemical reactions, which occur at the level of electrons—and are triggered by thermal conditions. As readers might recall from elementary school, the nucleus of any atom consists of protons, which are positively charged electrostatically, and neutrons, which are electrically neutral. Although positive protons repel each other, the nucleus is held together by a force known as the “strong nuclear force”—one of the four fundamental forces of nature—which acts in short range. Electrostatic repulsion due to positively charged protons acts in a relatively longer range than the strong nuclear force and extends beyond the nucleus. The nucleus is positively charged electrostatically as a result of the charges of the protons.
When two nuclei collide with each other, they can repel each other or they can fuse together (among other possible interactions). As repulsive electrostatic forces act in the longer range, fusion occurs if the colliding positively charged nuclei are fast enough to overcome the repulsive barriers and penetrate to the range of the attractive strong nuclear force interaction, and if they try enough times until they fuse. The number of fusion reactions within a given time period for particular reactants multiply substantially with increased temperature and increased packing (or density), as the nuclei are faster and more likely to collide when they’re closer together. It is possible, although not a likely outcome at the level of a single interaction, that a slow collision below the barrier energy can result in fusion, as the reactions are quantum mechanical in nature.
Electrical energy production from fusion relies in principle on heat and radiation released from fusion of lighter nuclei, such as hydrogen isotopes: deuterium and tritium. This energy from fusion reactions comes from changes in the combined mass of the products compared to the reactants in accordance with Albert Einstein’s E = mc² equation, and can be predicted from the nuclear binding energy curve. There are different types of fusion reactions in nature, and not just those of hydrogen isotopes. Light or heavy nuclei can fuse or combine given suitable conditions, but not all will release energy. It is possible to tell whether a fusion reaction will release energy or not from the nuclear binding energy curve. Energy-releasing fusion reactions are typically those involving light nuclei, particularly those that are lighter than iron-56. Lighter nuclei are favored because they tend to release more energy per unit mass and are relatively easier to fuse due to weaker repulsion.
Here, “energy-releasing” does not take into consideration the cost of overcoming the electrostatic repulsion of the positively charged protons to initiate the reaction which, despite being a small fraction of the reward (or output), remains difficult to overcome. The reason it is difficult to overcome is because, effectively, only a small fraction of the energy given to reactants in a system actually ends up causing fusion. This is in large part due to scattering collisions that occur more frequently than fusion reactions, especially at lower energies, and thereby hinder simple approaches to fusion, such as accelerator-based fusion.
Scattering collisions are also a problem in complex systems that heat plasma (the fusion fuel), as much of the energy given to heat the plasma is lost due to electromagnetic radiation, which follows these scattering collisions due to changes in velocity or energy of the scattering nuclei. Fortunately, these losses tend to scale slower with increasing temperature than energy gained from fusion, which in principle allows for self-sustained plasma at higher temperatures—assuming all other conditions for nuclei density and confinement time are satisfied. The break-even point is commonly referred to as the “static ignition temperature” of the plasma. At that temperature, the energy produced from the reactions equals the radiation losses associated with the temperature conditions and the reactants, allowing the reactions to cover the losses instead of relying on an external source to maintain the temperature of the plasma.
Self-sustained plasma is the first step to fusion energy. It remains unrealized because more energy is lost on each reaction than what is gained using current technology. Viable plasma requires high-fuel nuclei density and temperature for a long enough time before it becomes self-sustainable. Temperatures required for viable fusion plasma are on the order of 150 million Kelvin (nearly 270 million degrees Fahrenheit), but it is not the only requirement, as sufficient nuclei density and minimum confinement time must also be achieved before the plasma can produce enough power to sustain itself.
The plasma must be confined in order to allow it to reach ignition, control its expansion, reduce its heat loss, and reduce damage to reactor materials by preventing uncontrolled contact. Many approaches have been tested and used for plasma confinement such as inertial, magnetic, magneto-inertial, and electric field confinement. All available confinement techniques face different forms of plasma instabilities (imbalances) due to the fluid nature of the plasma and interaction with electrodynamics, which result in turbulence that causes rapid energy loss and cooling down of the plasma—particularly upon contact with reactor materials. Controlling these instabilities has been one of the most challenging problems in fusion research.
It is no secret that inertial confinement fusion has been employed in certain types of nuclear weapons—such information was publicly released during the 1953 Atoms for Peace program, which aimed to promote peaceful applications of nuclear energy. There have been numerous attempts to emulate the mechanisms involved to control the density, confinement time, and temperature conditions outside of weapons and use that to generate energy and neutrons for different applications.
Much of the leading work on inertial confinement fusion is done at the National Ignition Facility (NIF) in Livermore, California. The NIF in part supports the verification of nuclear weapon reliability through improving the scientific understanding and computational models of fusion—given the ban on nuclear weapon tests—as well as the peaceful application of fusion energy. In August of 2021, the NIF came close to plasma ignition using intense lasers directed at a gold-coated depleted uranium cavity “hohlraum,” which releases intense X-rays that symmetrically ablate a diamond pellet shell containing deuterium-tritium fuel, causing it to compress to high densities. Compression to high densities reduces the time required for plasma to reach ignition, which, if reached, would allow the reactions to produce more energy than the radiation losses associated with the temperature of the plasma—thereby maximizing the fraction of the fuel that is burned and potentially allowing for electricity production.
Although inertial confinement fusion may be the closest form of fusion to ignition—and it would not be surprising if the NIF achieved ignition this year or next—it’s unclear if it’s the closest to producing net electric energy. Energy capture and conversion is a major challenge for inertial confinement fusion. Inertial confinement fusion would also be a pulsed source rather than steady state, which complicates integration to the energy grid, especially given its intensity.
Magnetic confinement, on the other hand, is widely regarded as the primary approach to confinement in large steady-state fusion systems. It relies on multiple magnetic fields, typically arranged in a toroidal configuration, for plasma confinement. Use of the toroidal shape prevents the plasma from escaping from the poloidal direction.
The most common toroidal configuration resembles a doughnut and is called a tokamak. It contains a central solenoid that employs changing magnetic fields to induce intense electric currents in order to heat up the plasma (“ohmic heating”), initially to the temperatures needed for fusion and to support stable operation, and has toroidal and poloidal magnetic fields to confine the plasma at equilibrium. There are around 30 experimental tokamaks in the United States, Europe, Japan, India, and China that are currently operational. Most of these experimental tokamaks are not intended for demonstration of viable self-sustained fusion, but are intended for plasma physics research aimed at understanding how to better confine and heat the plasma, control instabilities, remove the exhaust, and verify material compatibility under intense radiation conditions.
The main challenges to confinement in tokamaks are instabilities due to ohmic heating such as the kink instability, and plasma drift due to force imbalances that result from magnetic field and pressure spatial gradients in different directions. (The most common are Rayleigh-Taylor-like instabilities). The kink instability limits the power of the ohmic heating that could be used to bring the plasma to ignition temperatures. Notably, recent advances have allowed for use of other methods to heat the plasma—limiting the dependence on ohmic heating—such as neutral beam injection and high-frequency electromagnetic waves. Decades of fusion experiments have also allowed for improved understanding of plasma instabilities and improved stabilization of Rayleigh-Taylor-like instabilities and others.
A stellarator is another magnetic confinement configuration that is also commonly used. There are currently 14 operational stellarators in the world, with the largest ones in Japan and Germany. They employ much more complex geometry than tokamaks, and developed historically as a solution to the Rayleigh-Taylor-like instabilities problem in tokamaks. Stellarators also require less injected power than tokamaks. However, stellarator technology faces several challenges, particularly with trapped particles—such as the super-banana orbit problem—that increase energy losses, which is why interest in general has shifted to tokamaks (which also happen to be easier to build).
While there are many concepts for purely magnetic toroidal confinement, tokamak systems are clearly the winners in this category at the present time. ITER, the most important international fusion project of the current decade, will employ a tokamak system to demonstrate self-sustained plasma in late 2025, and will run at full power by 2035 in southern France.
Members of the project include China, the European Union, India, Japan, Korea, Russia, and the United States, with participation from the U.K. The same Chinese facility that broke world records for highest plasma temperature in a tokamak, with temperatures exceeding 160 million degrees Celsius for 20 seconds, more recently achieved the world’s longest lasting fusion (although not self-sustainable) at 70 million degrees for 17 minutes. ITER will initially run in 400-to-600-second pulses at 150 million degrees, and could potentially be extended to steady-state operation in the future, although that is not the primary design intent of the project. If everything goes according to plan in ITER, we are only three years away from self-sustainable plasma in a tokamak system.
Advances in science and particularly energy are no longer made by a few isolated geniuses working on thought problems.
Recent advances in tokamak technology are worthy of our enthusiasm, particularly at a scientific level. But viable fusion energy systems will be very expensive due to the complexity of the technology and required materials, and in part due to its novelty and limited industrial supply of parts. Although we may be a few years away from self-sustainable plasma, and perhaps another two decades away from tokamak plants that produce net energy and electricity, assuming adequate funding is provided, the technology will still need to go through an intense cost-reduction phase in order to compete with today’s nuclear fission systems. It is unclear if this can happen in this century, given that fission systems are substantially less complex than fusion systems (particularly tokamaks) and are easier to construct.
ITER, which intends to demonstrate self-sustained plasma without net energy production at the plant level or electricity generation, will cost at least $22 billion, and will weigh a massive 23,000 tons. The estimated initial costs of ITER technology are at least 10 times that of nuclear fission, as energy analysts Robert Hirsch and Roger Bezdek have pointed out—not accounting for running costs. A kilogram of tritium produced using today’s technology costs about $30 million, while a kilogram of extracted, converted, enriched, and fabricated uranium fuel in the form of low-enriched uranium dioxide costs less than $2,000.
Readers who have consumed news reports on fusion technology without a technical background might have the impression that fusion is an “unlimited” source of energy, and therefore that the high fuel cost must be worth it. But a deuterium-tritium reaction only produces 3.8 times the energy per unit mass of a fission reaction in uranium, and energy production in fission reactors is only practically limited by structural materials. For running costs to be competitive with fission, assuming similar fuel utilization could be achieved, fusion fuel costs would need to be lowered from $30 million per kilogram to $7,600 per kilogram.
Fusion scientists and engineers are working hard on these challenges. One of the goals of ITER is to demonstrate tritium production using different blanket designs for tritium breeding from lithium to allow the reactor to produce the expensive component of its fuel. There are also potential alternatives to the deuterium-tritium combination, such as the more abundant and more easily producible deuterium-helium-3, which nevertheless suffers from higher energy losses due to increased radiation, and thereby requires ignition temperatures that are difficult to reach in a tokamak. But even if high fuel cost challenges are addressed, the cost of tokamak construction alone is still likely to make fusion economically uncompetitive with other energy sources.
Many fusion energy startups are pursuing simpler approaches to plasma confinement as a cheaper alternative to tokamaks at the expense of reduced power production. Approaches pursued include inertial confinement fusion combined with magnetized target (General Fusion), accelerator-based fusion combined with magnetic compression (Helion), and magnetic confinement by induced current stabilized by shear flows (Zap Energy). Companies such as Helion energy have ambitious timelines and aim to demonstrate net electricity production by 2024. If successful, it would be a significant milestone in fusion energy.
In all likelihood, we will witness self-sustained fusion in the next few years. The question is who will make it there first: the NIF, ITER, or startups? My bet is on the NIF, although I don’t expect inertial confinement fusion to produce net electricity any time soon. ITER will be a giant step toward steady-state fusion, and will perhaps be the most important physics experiment and engineering project of this decade. But it has a serious cost problem that limits potential for deployment in the real world.
Advances in science and particularly energy are no longer made by a few isolated geniuses working on thought problems. Current advances follow from the synergy between scientific research and economic systems that make efficient investments to facilitate experiments, theory development, and computational simulations, in order to enable discovery and development of complex systems. Timelines for technology deployment therefore depend not only on scientific and technological maturity, but on economic competitiveness with other energy sources. While fusion may be demonstrated soon, it may or may not be competitive enough to influence the energy landscape in this century, as other, cheaper forms of clean energy production (like nuclear fission) are available with abundant fuel. When it comes, however, affordable self-sustained viable fusion will be one of the most important moments in human history—a greater achievement, perhaps, than the moon landing.
Khaled Talaat is a postdoctoral scholar in nuclear engineering at the University of New Mexico. He has conducted research on multiple subjects including aerosols, radiological protection, and Generation IV lead-cooled fast reactors.