The Future of Nuclear Power, Part IV: Progress Toward Fusion Power

Note added on December 13, 2022: In a widely anticipated public announcement today, the National Ignition Facility (NIF) at Lawrence Livermore Laboratory reported that in a run on December 5 they have achieved ignition for the first time ever in a nuclear fusion experiment. As described in detail in the post below, NIF uses intense power pulses from 192 high-power laser beams to induce implosion on a tiny pellet containing a mixture of deuterium and tritium (see Figs. VI.3-6 below), thereby attaining the extreme temperatures and densities needed to generate inertial confinement fusion for a small fraction of a second. The new breakthrough is that from continued improvements, especially to target design, the experiment exceeded the breakeven point: in a pulse where the lasers input 2.05 megaJoules of energy to the target pellet, the resulting fusion generated 3.15 megaJoules of energy output. This achievement represents a crucial milestone in fusion research, but it still leaves a long way to go — probably decades — to attain commercially viable fusion reactors. For reference, the NIF facility needs about 300 megaJoules of electrical energy to operate their lasers; the breakeven milestone refers only to fusion energy output compared to laser pulse input to the target pellet. Also as explained in the post below, most commercial attempts to produce fusion reactors currently are based on the alternative approach of magnetic confinement fusion.

July 30, 2022

VI. The Status and outlook for fusion power

Earth intercepts a tiny fraction of the power emitted by the Sun: the ratio of the Earth’s cross-sectional area to the surface area of a sphere of radius equal to the Earth’s orbit radius, or less than one billionth. Even less heats the Earth, since some of that power is absorbed in the atmosphere and more of it is reflected from the atmosphere and Earth’s surface. Our short-term replacement of fossil fuel-generated power relies on that power from the Sun in the form of either solar power or wind power. The Sun’s energy causes winds on Earth because the air heated by the Sun over parts of Earth’s surface rises, being replaced by cooler air moving in from other parts of the globe, causing wind. Scientists have long had the feeling we could eventually do better than solar and wind power by trying to produce energy on Earth in the same way as the Sun does, by thermonuclear fusion of light nuclei.

In order to produce thermonuclear fusion energy, one must find ways to raise the temperature and density of a plasma containing the fuel to very high values. The hotter and denser one can make a plasma of fusing nuclei in the laboratory, the higher the rate of fusion will be. The power produced by fusion reactions, per unit volume of the plasma, is given by:

Pfusion = nAnB <σvAB> Efusion ,                  (Equation VI.1)

where: nA(B) are the number densities of the two fusing nuclei in the plasma (that number for protons in the Sun’s core is 9 x 1031 per cubic meter); <σvAB> is the fusion reaction rate determined by the cross section σ) for the reaction and averaged over the relative velocities vAB of the fusing nuclei within the plasma; and Efusion is the energy released in each fusion reaction that occurs. To generate significant amounts of net energy (power times time), one wants to maximize Pfusion, make it greater than power input and power losses due to radiation emitted by the hot plasma and ions leaving the plasma, and sustain the plasma conditions leading to successful fusion for as long as possible. In addition, of course, one must efficiently capture the released energy.

Figure VI.1 shows the fusion reaction rate <σvAB> for a few different choices of fusing nuclei, as a function of the plasma temperature maintained. The temperature is specified on the bottom scale in Kelvin (the temperature in Kelvin is equal to that in degrees Celsius plus 273 degrees) and on the top scale in the mean resulting ion kinetic energy. The reaction 2H + 3H → 4He + n, with the neutron carrying off most (14.1 MeV) of the energy release of 17.6 MeV, appears to be the most attractive option (blue curve in the figure). It is referred to as D-T fusion to represent the fusing deuterium (2H) and tritium (3H) nuclei and is, indeed, the option chosen in most, though not all, ongoing fusion research.

Figure VI.1. The reaction rate for three different light-nucleus fusion reactions plotted as functions of the plasma temperature. Credit: Dstrozzi

In comparison to the 15 million degree interior temperature of the Sun, Fig. VI.1 indicates that one can increase the rate of D-T fusion by more than a factor of 1000 by increasing the plasma temperature by an order of magnitude. That increase is essential to have any hope of producing more energy than one inputs. So most attempts at fusion power production in the laboratory aim for plasma temperatures of 100 – 150 million °C! The Sun gets by at lower temperature because the density of protons is so high; in a typical fusion reactor, the number densities in the plasma are more on the order of 1020 ions/m3 (about one millionth of atmospheric density near Earth’s surface), rather than the nearly 1032 protons/m3 in the Sun’s interior. At such densities and 100 million degree temperature, D-T fusion is capable of producing a few megawatts of power for every cubic meter of fuel.

There are not containers on Earth that can withstand temperatures of 100 million °C, so that implies that the plasmas in fusion reactors have to be confined in such a way that they do not touch walls. And the plasmas have to remain stable for long enough periods to yield more energy than input. These are the major technical challenges faced by attempts to produce fusion power on Earth. None of the attempts has yet succeeded in producing net energy, but recent results for two different methods of confining the plasma (inertial confinement and magnetic confinement, explained below) have come close. If the challenges are met in the future, fusion offers enormous advantages over fission reactors: abundant fuel, little radioactive waste (just relatively short-lived activated materials from neutron impingement), and no possibility of runaway or meltdown: if the input power is cut off, the plasma fades and fusion is discontinued, unlike a self-sustaining fission chain reaction.

Inertial confinement:

Although most ongoing fusion research focuses on magnetic confinement, we deal first with inertial confinement because that is the method that was adapted for hydrogen bombs in the 1950s. According to Equation VI.1, the energy that can result from fusion depends on the density and temperature achieved in the plasma and on the time that the plasma can be confined. Inertial confinement aims to achieve very high temperature and density, but over very short times, by causing a mixture of the fusing nuclei to implode. The confinement is inertial in the sense that fusion, and hopefully ignition of the fuel, occurs faster than the fuel can disperse after the implosion, due to the inertia of the particles. The implosion is triggered by an enormous energy pulse impinging on a small, typically peppercorn-sized, pellet containing a frozen mixture of the fusion fuel.

In a hydrogen bomb that trigger is provided by an initial explosion of a fission (atomic) bomb, which generates an intense bath of X-rays. Those X-rays very rapidly heat the outer layer of the vessel containing the hydrogen fuel. The enormous heating pulse blows off the outer layer, and the reaction (according to Isaac Newton’s Third Law of motion and forces) to those sudden, powerful, outward forces are inward forces that compress and heat the fuel inside to reach conditions at which the fusion fuel ignites. Concepts over the years for achieving controlled inertial confinement fusion have called for replacing that atomic bomb explosion by the sudden bombardment of the fuel pellet with various types of beams carrying large amounts of energy, from lasers to heavy ions.

Figure VI.2. Schematic of the intended stages of inertial confinement fusion using lasers. The blue arrows represent radiation; orange is blowoff; yellow is inwardly transported thermal energy.
1.   Laser beams or laser-produced X-rays rapidly heat the surface of the fusion target, forming a surrounding plasma envelope.
2.   Fuel is compressed by the rocket-like blowoff of the hot surface material.
3.   During the final part of the capsule implosion, the fuel core reaches 20 times the density of lead and ignites at 100,000,000 ˚C.
4.   Thermonuclear burn spreads rapidly through the compressed fuel, yielding many times the input energy

By far the most successful inertial confinement approach to fusion power has been achieved at the National Ignition Facility (NIF) at Lawrence Livermore Laboratory in California, using an array of 192 high-power laser beams to produce the energy pulse. The basic concept is illustrated in Fig. VI.2. A more realistic schematic of the NIF geometry is shown in Fig. VI.3.

Figure VI.3. False-color illustration of laser beams irradiating a NIF hohlraum with internal walls shaped like a rugby ball and holding an aluminum fuel capsule, one of several new target designs being explored on the NIF system. The laser beams enter the hohlraum through laser entrance holes and strike the inside of the hohlraum to generate X rays. Ignition is achieved when a self-sustaining fusion reaction produces more energy than the laser energy delivered to the target. Credit: Jacob Long

The lasers at NIF irradiate a small container known as a hohlraum whose actual size is indicated in the left-hand photo in Fig. VI.4. When the laser beams hit the inner, rugby-ball-shaped walls of the hohlraum, they generate a large bath of X-rays, which heat a tiny microcapsule (see right-hand photo in Fig. VI.4) in which the D-T fuel is stored. When the outer layer of the microcapsule is blown off by the heat, the reaction forces set offa rocket-like implosion that compresses and heats partially frozen hydrogen isotopes inside the capsule to conditions of pressure and temperature found only in the cores of stars and giant planets and in exploding nuclear weapons. The speed of the implosion—more than 400 kilometers per second—allows the fusion reactions to take place before the fuel can disassemble; the fuel is trapped by its own inertia (hence the term inertial confinement fusion).”

The 192 lasers at NIF deliver about 500 trillion watts of power (in comparison, a typical laser pointer delivers less than 5 thousandths of a watt) to the hohlraum, but only over a few nanoseconds, delivering a total energy of 1.8 million Joules. The arrangement of the laser ports around the target chamber is shown in Fig. VI.5, along with the inside of the chamber containing the hohlraum and its microcapsule at the center. The cost of building the NIF facility was over $3 billion U.S. The generation of all that laser power and the process of inducing the burning plasma are illustrated in a video animation, from which a still image is shown in Fig. VI.6.

Figure VI.6. This is a still image with an artist’s conception of the laser beam paths to the target hohlraum inside the NIF target chamber, once they have initiated the implosion of the target fuel. It is taken from this video animation.

The attainment of extreme densities in NIF fusion is aided by a shock wave that results from inward pressure on all sides of the fuel mixture, producing a very hot and dense inner hot spot within the fuel mixture. The temperature within that hot spot can reach 100 million °C and the density in that hot spot can reach values more than five times the central density in the Sun. So fusion reactions begin to occur within the hot spot, producing helium-4 nuclei (also known as alpha particles). Within that dense hot spot, those alphas come to rest within a distance of less than 0.02 millimeters, depositing the energy they carry from the fusion release within the hot spot, to heat it further. The neutrons produced in the D-T fusion reactions still exit the target, carrying most of the produced energy for potential use in a power plant.

According to the Livermore website: “Fusion ignition occurs when the heating power from alpha particles produced by fusion reactions in the hot spot at the center of the target capsule overcomes the cooling effects of x-ray losses, electron conduction, and implosion expansion, causing a self-heating feedback loop in the fusion fuel and an explosive amplification of energy output. NIF’s goal is to produce as much or more energy from fusion than the amount of laser energy delivered to the target. This will be accomplished by creating a “burning plasma,” in which a burn wave of fusion reactions propagates into the cold fuel surrounding the hot spot. In this process, known as alpha heating, the alpha particles spread throughout the cold fuel, depositing their energy, stimulating additional fusion reactions, and greatly increasing the yield.” The entire process from irradiation to burning plasma takes about 20 nanoseconds. The burning plasma is basically a short-lived “microstar.”

NIF has produced a burning plasma, but not yet break-even ignition, where energy produced begins to exceed energy input. But after many years of setbacks and improvements, they recently (August 2021) attained the breakthrough indicated in Fig. VI.7, producing 1.3 million Joules of output energy, an order of magnitude higher than they had attained previously.  They did not yet achieve break-even, since the output energy was about 70% of the total laser input energy. While not yet break-even, this breakthrough is one of the developments that has led to renewed enthusiasm for future prospects of fusion energy.

Figure VI.7. Plot of NIF fusion energy production results from 2012 to 2021. The high output on the right (1.3 megaJoules) indicates a fusion burn propagating through the fuel. The laser input energy is roughly the same in all of these shots, so the much higher output in 2021 represents a breakthrough in gain, though still short of break-even.

Magnetic confinement:

Because positively charged nuclei and negatively charged electrons are freed to move independently within a plasma, the individual particles are subject to magnetic forces. In the presence of a magnetic field, a moving charged particle will circle around the magnetic field lines, like the particles in a cyclotron. Therefore, appropriate shaping of magnetic field lines can be used, in principle, to confine charges in the plasma so they remain within a volume that does not intercept chamber walls. Furthermore, the plasma itself can conduct electricity, thus acting as a high-current electromagnet of its own. The magnetic field resulting from such a plasma current can combine with that produced by external electromagnets to craft an optimal field configuration for confining the plasma. This magnetic confinement approach to controlled thermonuclear fusion was first conceptualized in the 1950s by Russian physicists Igor Tamm and Andrei Sakharov (who later was awarded the Nobel Peace Prize for his efforts as a Soviet dissident). But several generations of resulting machines were needed to iterate on the optimal magnetic field construction.

The most widely used approach is called a Tokamak, a Russian acronym for “toroidal chamber with magnetic coils.” The state-of-the-art Tokamak design is embodied in the giant international project known as ITER: the International Thermonuclear Experimental Reactor. By the mid-1980s it was clear that mitigating various design flaws and instabilities in earlier Tokamaks (more than 100 test Tokamaks were built worldwide) would require an enormous reactor, too expensive for any one country to fund. The international collaboration for ITER was born starting from an agreement between U.S. President Ronald Reagan and Soviet Premier Mikhail Gorbachev, and design work began in 1988. 34 years later, it is now under construction in Cadarache in the south of France (a little north of Marseille), with major funding from the U.S., China, EU, Russia, Japan, South Korea, and India, and with cooperation agreements with several other countries. The total project cost, originally estimated as U.S.$6.6 billion has ballooned to more than $22 billion.

The design of the ITER Tokomak is illustrated in Fig. VI.8, with a (tiny) blue human figure at the bottom to indicate the scale of the device. Figure VI.9 shows a cutaway view of the entire plant, including systems providing power, electronics, vacuum, cryogenics, etc., support for the central Tokamak.

Figure VI.8. A cutaway view of the ITER Tokamak design. When it operates, the D-T plasma will be confined within the doughnut-shaped chamber with the D-shaped cross section. A human figure is indicated in blue at the bottom to indicate the scale of the device.
Figure VI.9. A cross-sectional view of the entire ITER reactor plant, with rooms containing various support systems surrounding the Tokamak.

Figure VI.10 shows an aerial view of the Cadarache site during ITER construction in 2018, with the Tokamak assembly dwarfed in the center, and surrounding buildings providing various support, testing and office space. ITER is scheduled to begin operation in 2025, and to generate net energy by D-T fusion by 2035. Its goal is to produce 500 megawatts (MW) of power for an input power of 50 MW, demonstrating viability. 500 MW for a price tag of $22 billion is not economically viable yet – that’s why there is an E in ITER. But the progress on its construction is another source of optimism for the future of fusion power.

Figure VI.10. Aerial view of the Cadarache site for ITER during construction in 2018.  The Tokamak structure is seen at the center.

A conceptual drawing of the magnets and confined plasma involved in the ITER Tokamak design is shown in Fig. VI.11 to illustrate how the plasma gets confined. The ensemble of toroidal field coils (blue rings) surrounding the doughnut-shaped (toroidal) vacuum chamber in Fig. VI.11 establishes a large magnetic field that circles around the center of the ring chamber (blue arrow labeled “toroidal magnetic field”). If this were the only source of magnetic field, particles in the plasma would gradually drift toward the inner radius of the torus because the ring magnets are closer together there and the magnetic field stronger in magnitude. So, additional coils are added to circulate current within the doughnut hole (central solenoid).  The central solenoid for ITER is one of the world’s largest superconducting magnets, producing a field strength of 13 Tesla, equivalent to 280,000 times the Earth’s magnetic field that allows compasses to operate. That magnetic field is strong enough to “hoist an aircraft carrier six feet out of the water.”

Figure VI.11. Conceptual drawing of the various magnet coils used to create the helical magnetic field (twisting black lines) that confines the D-T plasma (pink shaded “doughnut”) in the ITER Tokamak. The very large electric current (labeled green arrow) carried by the plasma itself plays a critical role in shaping the resultant field.

The central solenoid and the plasma itself form a transformer: by driving current pulses through the inner coils, a circumferential pulsed current (labeled green arrow in Fig. VI.11) of several million amperes is induced within the plasma itself. That plasma current sets up its own poloidal magnetic field (green arrows circulating from inside to outside the doughnut) that gets added to the toroidal magnetic field (blue arrow within the doughnut), to produce a net helical magnetic field indicated by thin black lines. The helical field lines corkscrew around the circumference of the doughnut. In the presence of these fields, the charged particles in the plasma move along helical paths inside the pink shaded doughnut, making many turns from the inner to the outer radius and back, inside the doughnut, for each complete orbit around the circumference of the doughnut. Those turns help to minimize losses of particles from the plasma and some plasma instabilities. The outer poloidal coils in Fig. VI.11 are used to generate a magnetic field that shapes and positions the plasma as desired.

The heating of the plasma is provided initially by microwaves generated by high-power, high-frequency oscillators outside the torus. But once the plasma is sufficiently hot (at least 100 million °C) and dense, the charged particles within it undergo frequent collisions during their motion around the doughnut. Those collisions form an ohmic resistance for the plasma current, like that in a wire, generating heat that raises the plasma temperature further and, hence, increases the fusion rate. For D-T fusion, the resulting alpha particles (carrying 3.5 MeV of the released energy) deposit most of their energy rapidly in the plasma, where they are also trapped for some time, and they enhance the collisional self-heating even more, depositing a fraction of the fusion energy to add to the internal heating. The maximum temperature that could be reached in a Tokamak via the plasma resistive heating alone is 20-30 million °C, so the external microwave heating remains essential to keep the fusion rate high.

If the fusion rate becomes high enough to generate just as much energy as one puts in via the external plasma heating, you’ve reached the breakeven condition. So far, no Tokamaks have even reached, let alone surpassed, breakeven. The record set at previous Tokamaks was achieved in 2020 at the Joint European Torus (JET) in Oxfordshire, UK, which generated 16 megawatts of output power for 24 megawatts of input heating power, a fractional power return similar to that attained with inertial confinement at NIF in 2021.

However, the plasma density and power output per unit volume achieved in magnetic confinement are much smaller than attained momentarily in NIF inertial confinement. Hence, magnetic confinement fusion must be maintained over much larger plasma volumes and time periods much longer than the tens of nanoseconds of burn in an inertial confinement micropellet, in order to achieve comparable energy production. In December 2021, JET produced a total energy of 59 megaJoules, sustaining a production rate of 11 megawatts for more than 5 seconds; this was twice their previous (1997) record. Although still well short of breakeven, the output was still enough to add to the renewed enthusiasm for the future of fusion power. To do better requires improvements in some or all of density, temperature, volume, and confinement time of the plasma.

JET could not sustain the plasma for much more than 5 seconds because its copper electromagnets get too hot. ITER uses superconducting magnets to alleviate this limit. Fresh deuterium and tritium fuel must be continually injected into the plasma because particles do get lost from the plasma via collisions (or when confinement ceases). These lost particles generally hit and heat the inner torus walls, which must be water-cooled. But ITER also has ten times the toroidal vessel volume than JET, lessening problems associated with particle interactions with the chamber walls. The fuel is injected, by the way, as neutral atoms, which then get ionized via collisions with the ions already present in the plasma. That ionization also contributes to the plasma heating.

How will output energy be captured in a fusion reactor? Most of the energy produced in D-T fusion is carried off by the emitted neutron. In ITER, which is an experiment rather than a power plant, those neutrons will be absorbed in shields surrounding the Tokamak. The neutrons can penetrate the chamber walls. Good materials for absorbing neutrons are hydrogen-rich and may include boron, which has a high capture probability for low-energy neutrons. So, “concrete and polyethylene doped with boron make inexpensive neutron shielding materials.” But in a real fusion reactor one wants to milk the kinetic energy carried by those neutrons to produce electric power. This will most likely be done by absorbing the neutrons in a lithium-based blanket lining the inner walls of the plasma chamber. The heated blanket will then heat water to generate steam and to run turbines for power generation, much as in a fission reactor. Alternative fusion approaches that attempt to convert the kinetic energy of fusion products to electric energy more directly will be discussed further below.

Lithium is an effective material for a neutron-absorbing blanket because it helps, in principle, to solve a serious issue with fusion fuel availability. Deuterium is a readily available fuel, since it comprises 0.015% of naturally occurring hydrogen, and hydrogen is super-abundant in Earth’s oceans. A liter of water contains about 1022 (i.e., 10 billion trillion) deuterium atoms. Tritium is another matter because it decays radioactively (with a half-life of 12.3 years). It can be produced from deuterium in a fission reactor or a particle accelerator. But it also can be replenished within an operating Tokamak via collisions of neutron fusion products with a lithium blanket, via the reactions:

n + 6Li → 3H + 4He   or   n + 7Li → 3H + 4He + n.

Even if production and reuse of tritium in a Tokamak blanket turns out to be sufficient to replenish the fuel continuously – and this has yet to be demonstrated – there is a serious issue with providing the initial amounts of tritium needed to get fusion going in the first place. The worldwide stockpile of tritium at the moment is only about 25 kilograms, produced as a waste product in heavy water fission reactors, where the deuterium in the water absorbs neutrons from fission reactions to yield tritium. The tritium is currently harvested in 19 Canada Deuterium Uranium (CANDU) nuclear fission reactors, mostly in Canada and South Korea, each producing about 0.5 kilograms a year. But roughly half of these reactors are due to be retired within a decade, while ITER startup will begin to exhaust the supply, along with tritium radioactive decay. Without new sources, ITER projections shown in Fig. VI.12 foresee a tritium crisis arising by about 2050. Thus, relying on D-T fusion to contribute meaningfully to worldwide energy needs in the second half of this century will require serious side efforts on tritium production.

Figure VI.12. ITER projections for worldwide supply of tritium during the first half of the 21st century, as CANDU heavy water reactors are retired and ITER operations burn tritium.

Two other significant engineering problems ITER and future Tokamaks need to solve concern neutron shielding and steady operation. According to the Wikipedia article on fusion power: “The neutron flux expected in a commercial D-T fusion reactor is about 100 times that of fission power reactors, posing problems for material design. After a series of D-T tests at JET, the vacuum vessel was sufficiently radioactive that it required remote handling for the year following the tests.” The neutrons produced in fission reactors are moderated to significantly lower energies than those produced in D-T fusion, and they tend to get absorbed within the fuel rods by initiating further fission reactions or within the control rods.

Reliability issues for Tokamaks center on the need for pulses of electric current in the central solenoid to induce the plasma current and its contribution to the confining magnetic fields in Fig. VI.11. As pointed out in a recent Science magazine article, “In between pulses the magnetic field ebbs, interrupting tokamak operations—and power delivery. The repetitive starts and stops of the reactor’s powerful magnetic fields also generate mechanical stresses that could eventually tear the machine apart.” Successful operation of ITER to produce net power reliably for significant periods of time by 2035 will go a long way toward demonstrating fusion’s viability.

Commercial fusion power companies:

The other reason for renewed optimism that fusion power may take off in the not-too-distant future is the current rush of private equity funding for commercial fusion enterprises. As reported in a November 2021 Nature article: “There are now more than 30 private fusion firms globally…the 18 firms that have declared their funding say they have attracted more than US$2.4 billion in total, almost entirely from private investments.” That investment total continues to grow today. The competition between industry and government-funded laboratories was a real spur to rapid advancements at the end of the 20th century in the case of the Human Genome Project, where Craig Venter’s commercial company Celera Genomics automated gene sequencing to accelerate the research. There is reason for hope that industry-government competition and cooperation may provide a similar boost to fusion power timelines.

Six of the biggest players in the commercial fusion world are: TAE Technologies, based in Foothill Ranch, California, which has raised $1.2 billion from investors including Google, Chevron and the New York City investment bank Goldman Sachs; Helion Energy, in Everett, Washington, with $578 million in private equity investment; Commonwealth Fusion Systems (CFS) in Cambridge, Massachusetts, spun off from research at the Plasma Science and Fusion Center of MIT ($1.8 billion invested from group including Bill Gates, Google and George Soros); Tokamak Energy ($200 million), in Culham, UK (the site of the JET research Tokamak); General Fusion (GF, also $200 million), based in Burnaby, Canada; and First Light Fusion, spun off from research at Oxford University in the UK. The prospects of friendly competition and cooperation with government-sponsored, larger efforts is promising enough that “the UK government and the US Department of Energy are also investing in firms such as Tokamak Energy, CFS and GF.”

One advantage of so much commercial competition is that competing companies will work on developing alternative technologies and we’ll all get to see which ones develop most rapidly and most reliably. This is somewhat analogous to the many pharmaceutical companies competing to develop COVID vaccines quickly in 2020, leading to a faster path to reliable vaccines than most people imagined possible. The commercial companies are interested in developing modest-sized fusion power plants – they cannot compete with government funding in building facilities as huge as ITER or NIF. But they are taking an array of different technological approaches, as indicated in the Nature article. We will illustrate a few of these alternative approaches, addressing specific concerns of Tokamak-based magnetic confinement fusion, below.

Figure VI.13. Illustration of a smaller, nearly spherical Tokamak facilitated by high-power magnets wound with high-temperature superconducting ribbon.

Tokamak Energy and Commonwealth Fusion Systems are aiming to produce much more compact mini-Tokamaks (see Fig. VI.13) with a shape more spherical than the huge ITER device. For example, the spherical Tokamak designed by Tokamak Energy will be 3.5 meters in diameter, compared to ITER’s 30 meters, including the surrounding cooling systems. The main technical advance that allows the enormous size reduction is the use of high-temperature superconducting ribbons to construct the magnets, which should produce stronger fields and remove the need for a liquid helium cryo-plant in comparison with the more conventional superconducting magnets at ITER. In fact, CFS already beat ITER to the punch in producing the strongest man-made magnets (20 Tesla, compared to ITER’s 13 Tesla) on Earth, just before the huge ITER inner solenoid was delivered to Cadarache. The short video in Fig. VI.14 shows how the CFS magnet, designed at MIT, is assembled and will fit into the compact demonstration model spherical Tokamak the company calls SPARC.

Figure VI.14. A video animation illustrating how the high-temperature superconducting toroidal magnets developed by Commonwealth Fusion Systems will fit together in the assembly of their designed mini-Tokamak SPARC.

Several companies are developing novel, non-Tokamak, approaches to magnetic confinement, with a focus on addressing the tritium fuel shortage.  General Fusion is attempting a magnetized target reactor, still using D-T fusion, but in the very different configuration shown in Fig. VI.15. Here, the plasma resides within a cavity generated inside a liquid metal target by spinning it rapidly with a centrifuge. Then, in a series of pulses, once per second or so, the elaborate piston system pictured pumps more liquid metal into the chamber, compressing the cavity and the plasma inside over a few tens of milliseconds, during which conditions for fusion could be attained. By including liquid lithium in the liquid metal material, more tritium would be produced in place via neutron bombardment every time fusion reactions proceed, improving on the Tokamak design, where the lithium blanket lies further from the fusing fuel.

Figure VI.15. The magnetized target approach being developed by General Fusion to confine D-T plasma within a cavity inside spinning liquid metal, compressed periodically by a large array of pistons.

TAE Technologies is attempting an even more exotic method for magnetic confinement fusion, eliminating the fuel availability problem by replacing D-T fusion with the reaction:

1H + 11B → 4He + 4He + 4He .

This reaction produces only about half as much energy as D-T fusion (8.68 vs. 17.6 MeV), but it produces no neutrons, hence, no significant activation concern outside the device, and uses naturally abundant, stable hydrogen and boron, rather than radioactive tritium as fuel. A major drawback is that the higher electric repulsion involved with the boron nuclei, of atomic number 5 rather than 1, requires reaching a temperature of 1 billion degrees (!), rather than 100 M degrees for D-T.

Figure VI.16. The TAE Technologies design for proton-boron fusion.

In the TAE Technologies design, pictured in Fig. VI.16, hot, supersonic pulses of hydrogen plasma will be injected from either side, to collide inside the central magnetic solenoid. The collision produces a hot plasma spinning about its own axis, which sets up its own self-confining magnetic field. Boron atoms will be injected tangentially from an array of injectors surrounding the central solenoid, in such a way as to keep the plasma rotating – just as you might keep a child’s merry-go-round spinning with tangential pushes. Collisions with the protons in the plasma will ionize the boron atoms and start the fusion process. The current TAE prototype has so far sustained plasma temperatures of 75 million degrees Celsius, but there’s quite a long way to go to demonstrate the ability to make this technology work and reach a billion degrees!

Figure VI.17. Helion’s proposed plasma accelerator for D-3He fusion and direct production of electricity from fusion power.

Helion’s plasma accelerator, pictured in Fig. VI.17, is also designed to eliminate the need for tritium and the production of neutrons, but by using D-3He fusion (see the red curve in Fig. V.1):

2H + 3He → 4He + p,

which releases 18.4 MeV, mostly carried by the resulting proton. Deuterium and 3He plasma pulses are heated and magnetically compressed at each end of the apparatus and then accelerated toward the center at speeds of about 450 kilometers per second. When the two plasma bunches collide in the center, the resulting hot plasma is further compressed by a powerful central magnetic field, until it reaches 100 million °C, where fusion begins at appreciable rates. The resulting alpha-particle heating not only keeps the fusion going, but also causes the plasma to expand. The expanding plasma interacts with the central magnetic field to induce an electric current that can then provide electric power directly, without having to produce steam and activate a turbine. Though much of the technology still needs to be demonstrated, the major drawback of the Helion approach is that 3He, though stable, is scarce. It will have to be produced in accelerator or reactor facilities via fusion of two deuterons, in sufficient quantities for fusion reactors of this design to become economically viable. But because it produces no neutrons, which penetrate material, it has much less activation concerns than more conventional approaches.

Finally, there is one company, First Light Fusion, which is basing its design on inertial confinement, rather than magnetic confinement. They propose to produce the shock wave that will compress a D-T fuel cell not with an array of high-power laser beams, as at NIF, but rather by bombarding the target cell with high-speed (50 kilometers per second) projectiles fired from an electromagnetic gun. The critical target design, which they claim will “multiply the velocity and create convergence,” is a trade secret, so details are not readily available.

Driven by the competition among designs and teams, and in attempts to continue attracting private investment, each of these companies claims that they can have a commercially viable fusion power reactor demonstrated by the early- to mid-2030s. Some of the projected timelines are indicated in Fig. VI.18 in a milestone chart copied from the Nature article. The aim of most of these companies is to produce modest-sized power plants yielding 100 to several hundred megawatts of net power, at eventual construction costs per plant from several hundred million to a billion U.S. dollars.

Most of the commercial projections in Fig. VI.18 seem highly optimistic. In all probability, each design will encounter unforeseen technical challenges that require some years to surmount, and some approaches will prove unviable, as has been the half-century history of fusion power to date. However, it does now seem plausible that fusion power plants can generate a significant fraction of the world’s energy needs sometime within the second half of this century. This is not rapid enough to address the ongoing climate crisis. So, using the Sun’s energy in solar and wind power plants will continue to play a critical role over the next few decades, while producing “mini-Suns” in power plants on Earth may provide the long-term solution for abundant, clean, carbon-free, and largely radioactive-waste-free, power.

VII. nuclear power outlook

The outlook for nuclear power has waxed and waned over the past half-century. Initially viewed as the energy source of the future, fission reactors have suffered immensely in public perception as a result of several well-publicized reactor accidents and public fears of radioactivity and its potential long-term health impacts. The three major accidents we reviewed in Part II of this post revealed problems in conventional power reactors associated with design flaws, operator error, and failure to provide safeguards for rare natural disasters. Including these three major accidents, there have been 28 worldwide fission reactor accidents and incidents between 1952 and 2011 that caused multiple fatalities and/or more than US$100 million in property damage. Just about half of these events have occurred in U.S. power plants.

In the wake of the devastating Fukushima Daiichi disaster of 2011, a number of countries suspended licensing of any further fission power plants and are looking to phase out existing ones. Meanwhile, harnessing thermonuclear fusion for energy production on Earth has, up until now, been a half-century-long R&D program, and we are still likely decades away from seeing commercially viable fusion power plants.

However, the looming crisis of climate change caused by the burning of fossil fuels is leading to a re-evaluation of nuclear power in the new light as a carbon-free energy source. The majority of fission power plants around the world have operated safely and reliably for decades, providing about 10% of the world’s electricity needs (see Fig. VII.1). In France, nuclear power has accounted for more than 70% of the country’s electricity production, keeping electricity prices among the lowest in Europe, but their reactors are beginning to show signs of wear.  

Figure VII.1. Cumulative years of operation of fission power reactors worldwide, through 2019. The three accidents treated in Part II are indicated in the years they occurred.

In the U.S., 92 remaining operating fission power plants at present reliably provide 19-20% of the country’s electricity and about 55% of its carbon-free electricity. However, almost all of that capacity comes from reactors constructed between 1967 and 1990; there are only 2 new fission power plants currently under construction in the U.S. The U.S. Nuclear Regulatory Commission has traditionally licensed such reactor plants to operate for only 60 years, including one 20-year renewal extension. It is likely that a majority of those old fission power plants, showing wear and tear similar to problems cropping up in France, will be decommissioned by 2050. (If those decommissioned reactors are to be returned to “green field” sites, the costs will be enormous to dispose safely of residual radioactivity within reactor vessels and containment structures. NRC licenses require reactor owners to “establish or maintain a financial mechanism — such as a trust fund or a guarantee from its parent company — to ensure there will be sufficient money to pay for the ultimate decommissioning of the facility.“) Thus, along with considering how to replace the capacity of fossil-fuel plants over the coming few decades, we must also consider how to replace much of the currently world-leading carbon-free electricity production.

That realization has led to the current focus on, and investment in, the development of small modular fission reactors (SMR) and thermonuclear fusion reactors. The main impetus for SMR development has been economic, namely, avoiding the enormous upfront capital investment and drawn-out licensing and construction periods required for a more conventional, gigawatt-scale fission power plant, often with no guarantee that it will operate in the face of strong public opposition. One of us (SV) often used to eat at a charming restaurant sited in the shadow of an abandoned nuclear power plant in Shoreham, Long Island, which never operated, in the face of enormous public outcry based on radiation fears and the paucity of escape routes. Nonetheless, that abandoned plant continues to cost Suffolk County residents significantly inflated electricity prices to compensate for its construction costs.

The many worldwide companies now seeking to market SMRs forecast substantial cost savings from mass production, but the public will have to be convinced of the safe operation of selected first-generation SMRs before the industry will ever make it to the mass production stage. The SMR designs we have described in Part III of this post do offer significant improvements in safety over conventional reactors, on paper. One improvement is the use of passive primary cooling systems integrated within a single reactor pressure vessel, eliminating the need for pumps subject to failure and power loss, and for explicit operator interventions to prevent thermal runaway. Some designs furthermore control the fission rate naturally in passive safety systems relying on the temperature-dependence of the fuel density. An additional improvement is the move to multiple, smaller reactors, so that a potential meltdown in any one of them involves a radioactive load far smaller than for any of the Fukushima Daiichi reactors.  However, a severe natural disaster akin to the Fukushima tsunami could still affect an entire bank of SMRs, such as envisioned by NuScale.

Molten salt modular reactors offer further advantages by operation without pressurization, without the potential release of combustible hydrogen, with fission rate control offered naturally by the temperature-dependence of fuel properties, and with natural storage of generated heat in the molten salt for supplementing more variable solar and wind power. Furthermore, by operating at much higher temperatures than water reactors, they can generate electricity from heat considerably more efficiently or use the heat directly for various applications. One particularly problematic aspect of at least the Terrestrial Energy molten salt reactor design is the need to replace reactors every 7 years, due to deterioration of the graphite moderators.

But the history of fission reactors makes it clear that safety has to be demonstrated over years of practical operation, not just on paper. Reactor designers have often not foreseen, or chosen for economic reasons not to address, some potential, rare hazard conditions. New types of hazards are introduced by the possibilities of nuclear terrorism and cyber warfare. The Russian invasion of Ukraine has shown that even military attacks on nuclear power plants are within the realm of possibility. And the radioactive waste issue persists, perhaps even gets exacerbated, for SMRs. While SMR designs often address the known safety issues that have shown up in traditional fission reactors, one is reminded of Jeff Goldblum’s response to Richard Attenborough in The Lost World: Jurassic Park, when Attenborough’s character explains: “Don’t worry. I’m not making the same mistakes again.” Goldblum responds: “No, you’re making all new ones.”

Reactor companies seeking to keep their construction and operation costs competitive with other energy sectors are frequently reluctant to provide safety manpower or hardware beyond the minimum needed to satisfy regulatory agency requirements. And the regulatory agencies are not always effective in resisting industry pushback. Industry needs to thread the needle between excessive costs that customers will be unwilling to pay and sufficiently robust, safe reactor operation to not jeopardize public and political support. On the construction side, the SMR industry aims for a “sweet spot” cost below roughly U.S. $3,500 per kilowatt of electric power generated (or equivalently, $3.5 billion per gigawatt electric, GWe).

With regulatory approval in hand or close, we foresee initial sales and installation of water SMR power plants – particularly the NuScale integrated pressurized water SMR and the GE-Hitachi boiling water version, both described in Part III – around the world within the next decade. Installations of molten salt SMR plants may take another decade beyond that to gain approval and be implemented. But it is likely to require a decade of safe operation of those initial power plants before public suspicions can be overcome and SMR production can take off. If things go well, SMR power plants could play a significant role in replacing phased-out fossil fuel energy production by 2050.

Energy generation from thermonuclear fusion plants is a longer-term proposition. Fusion is an inherently safer carbon-free energy source than fission because fusion reactions in a power plant are not self-sustainable, not subject to thermal runaway, and do not produce serious long-term radioactive emissions or waste problems. But many engineering problems remain to overcome. The first essential goal is to demonstrate that energy output can reliably exceed energy input needed to raise fusion fuel plasmas to sufficient temperature and density for fusion reactions to proceed at useful rates. We expect that goal to be demonstrated no later than 2035 via ITER operations, and possibly sooner at one or more of the commercial projects now under way. But then come numerous additional engineering challenges to demonstrate long-term reactor reliability and fuel (especially, tritium) availability, and to get costs down to a level where fusion is economically competitive with fission plants. For perspective, ITER construction costs have exceeded that $3.5 billion/GWe competitiveness goal by more than an order of magnitude.

If these basic engineering challenges are met, fusion power plants during the second half of this century may be based on one or more of the innovative, smaller-scale designs currently being pursued by the commercial fusion power companies reviewed above. Among these, the ones that seem most promising to us for first implementation are the ones based on mini-Tokamaks, which combine the long learning curve climbed to date through Tokamak R&D with the now demonstrated technology of powerful electromagnets wound with high-temperature superconducting ribbon materials. The alternative designs described above are novel and intriguing, but will likely run up against their own long engineering learning curves.

We thus view strong prospects for nuclear power in some form to play a major role in worldwide power production during the second half of the 21st century, along with renewables. In order to prepare the public for such developments, it is important now for the nuclear industry, politicians and scientists to begin educating citizens about nuclear energy that they need not fear, while acknowledging that past fears were justified. After all, solar and wind power are really just nuclear fusion energy production at a distance, since they use energy from the Sun. By the middle of the current century, a century’s worth of science and engineering may have taught us how to harness the Sun’s energy production here on Earth.


The IAEA Database on Nuclear Power Reactors,

Backgrounder on Radioactive Waste,

Disposition of High-Level Waste and Spent Nuclear Fuel (National Academies Press, 2001),

Nuclear Reprocessing: Dangerous, Dirty, and Expensive,

Storage and Disposal of Radioactive Waste,

A. Boyle, If There’s a Meltdown, Then What?, NBC News, March 15, 2011,

Wikipedia, Three Mile Island Accident

INES: The International Nuclear and Radiological Event Scale,

Wikipedia, Corium (nuclear reactor),

Jon F. Elliott, The Kemeny Report on the Accident at Three Mile Island, Ecology Law Quarterly, March 1980

Joseph Mangano, Three Mile Island: Health Study Meltdown, Bulletin of the Atomic Scientists 60, 30 (2004). 

J.W. Gofman and A.R. Tamplin, Poisoned Power: The Case Against Nuclear Power Plants Before and After Three Mile Island, Rodale Press, 1979. 

Wikipedia, Linear No-Threshold Model

Joel N. Shurkin, Edward Teller, ‘Father of the Hydrogen Bomb,’ is Dead at 95, Stanford Report Sept. 24, 2003

Wikipedia, The China Syndrome

The American Nuclear Society, The TMI-2 Cleanup: Challenging and Successful,

CDC, Radioisotope Brief: Strontium-90

Wikipedia, Xenon-135

Wikipedia, Iodine-131

Wikipedia, Caesium-137

G. Brumfiel, Satellite Photo Shows Russian Troops Were Stationed in Chernobyl’s Radioactive Zone, NPR, April 7, 2022,

U.S. Nuclear Regulatory Commission, Information for Radiation Workers  

International Atomic Energy Agency, Summary Report on the Post-Accident Review Meeting on the Chernobyl Incident, 1986 (INSAG-1):

International Atomic Energy Agency, INSAG-7: The Chernobyl Accident: Updating of INSAG-1

The Gorbachev Foundation, Mikhail Gorbachev, Turning Point at Chernobyl, April 17, 2006 

Wikipedia, Fukushima Nuclear Disaster

World Nuclear Association, Fukushima Daiichi Accident

World Nuclear Association, Fukushima: Background on Reactors,

Union of Concerned Scientists, Fukushima FAQs

IRSN France, Fukushima in 2016: The Social Consequences of the Accident,

Investigation Committee on the Accident at the Fukushima Nuclear Power Stations, Sept. 28, 2012

The Economist, The Fukushima Black Box, Jan 7, 2012

T. Yanaguchi, Bill for Japan’s Fukushima Cleanup to Double to $201 Billion: Source, Reuters, Nov. 29, 2016,

H. Pailliere and J. Donovan, Nuclear Power Ten Years After Fukushima, IAEA News, March 11, 2021,

Advanced Small Modular Reactors (SMRs), Office of Nuclear Energy, DOE,

Benefits of Small Modular Reactors, Office of Nuclear Energy, DOE

Small Nuclear Power Reactors, 

J. Liou, What are Small Modular Reactors (SMRs)? IAEA Office of Public Information and Communication,

D. Solan, et al., Economic and Employment Impacts of Small Modular Nuclear Reactors, The Energy Policy Institute, June 2010,

R. Fares, 3 Ways Small Modular Reactors Overcome Existing Barriers to Nuclear, Scientific American, May 9, 2016,

E. Engheim, Different Small Modular Reactor Designs,

World Nuclear Association, Small Nuclear Power Reactors

J. Doyle, et al., Highly Reliable Nuclear Power for Mission-Critical Applications,

The BWRX-300 Small Modular Reactor,

GE-Hitachi Kicks Off BWRX-300 SMR for Estonia, Neutron Bytes, Oct. 5, 2019,

GE-Hitachi Nuclear Energy Selected by Ontario Power Generation as Technology Partner for Darlington New Nuclear Project, GE Press Release, Dec. 2, 2021,

GE-Hitachi Nuclear Energy’s BWRX-300 SMR Selected for Saskatchewan, World Nuclear News, June 28, 2022,

GE-Hitachi, Decades of Innovation Helped GEH Create PRISM

Molten Salt Reactors,

M. Shwartz, Stanford-Led Research Finds Small Modular Reactors Will Exacerbate Challenges of Highly Radioactive Nuclear Waste, Stanford News, May 20, 2022,

L.M. Krall, A.M. Macfarlane, and R.C. Ewing, Nuclear Waste from Small Modular Reactors, Proceedings of the National Academy of Sciences 119, e2111833119 (2022),

M. Schneider, Fast Breeder Reactors in France, Science and Global Security 17, 36 (2009),

Pursuing Fusion Ignition, Lawrence Livermore National Laboratory,

Europe Beats Japan to ITER Prize, Physics World, June 28, 2005,

ITER Powers Ahead, CERN Courier, Nov. 3, 2021,

U. Irfan, Fusion Energy is a Reason to be Excited About the Future, Vox, Jan. 26, 2022,

JET: EUROfusion’s Flagship Device,

J. Amos, Major Breakthrough on Fusion Energy, BBC News, Feb. 9, 2021,

D. Clery, Out of Gas: A Shortage of Tritium Fuel May Leave Fusion Energy with an Empty Tank, Science 376, 1372 (2022),

D. Clery, Breakdowns Could Plague Fusion Power Plants, Science 376, 1376 (2022),

P. Ball, The Race to Fusion Energy, Nature 599, 362 (2021),

TAE Technologies Secures Funds to Build Next Fusion Machine, World Nuclear News, July 20, 2022,

D. Chandler, MIT-Designed Project Achieves Major Advance Toward Fusion Energy, MIT News, Sept. 8, 2021,

L. Alderman, French Nuclear Power Crisis Frustrates Europe’s Push to Quit Russian Energy, New York Times, June 18, 2022,

Nuclear Power in the U.S.A.,

Safety of Nuclear Power Reactors,

Backgrounder on Decommissioning Nuclear Power Plants,

E. Lyman, Small Isn’t Always Beautiful: Safety, Security, and Cost Concerns About Small Modular Reactors, Union of Concerned Scientists, Sept. 2013,