Fusion power

Fusion reactions occur when two or more atomic nuclei come close enough for long enough that the nuclear force pulling them together exceeds the electrostatic force pushing them apart, fusing them into heavier nuclei. For nuclei heavier than iron-56, the reaction is endothermic, requiring an input of energy.[4] The heavy nuclei bigger than iron have many more protons resulting in a greater repulsive force. For nuclei lighter than iron-56, the reaction is exothermic, releasing energy when they fuse. Since hydrogen has a single proton in its nucleus, it requires the least effort to attain fusion, and yields the most net energy output. Also since it has one electron, hydrogen is the easiest fuel to fully ionize.

The repulsive electrostatic interaction between nuclei operates across larger distances than the strong force, which has a range of roughly one femtometer—the diameter of a proton or neutron. The fuel atoms must be supplied enough kinetic energy to approach one another closely enough for the strong force to overcome the electrostatic repulsion in order to initiate fusion. The "Coulomb barrier" is the quantity of kinetic energy required to move the fuel atoms near enough. Atoms can be heated to extremely high temperatures or accelerated in a particle accelerator to produce this energy.

An atom loses its electrons once it is heated past its ionization energy. An ion is the name for the resultant bare nucleus. The result of this ionization is plasma, which is a heated cloud of ions and free electrons that were formerly bound to them. Plasmas are electrically conducting and magnetically controlled because the charges are separated. This is used by several fusion devices to confine the hot particles.

The fusion reaction rate increases rapidly with temperature until it maximizes and then gradually drops off. The deuterium-tritium fusion rate peaks at a lower temperature (about 70 keV, or 800 million kelvin) and at a higher value than other reactions commonly considered for fusion energy.

A reaction's cross section, denoted σ, measures the probability that a fusion reaction will happen. This depends on the relative velocity of the two nuclei. Higher relative velocities generally increase the probability, but the probability begins to decrease again at very high energies.[5]

In a plasma, particle velocity can be characterized using a probability distribution. If the plasma is thermalized, the distribution looks like a Gaussian curve, or Maxwell–Boltzmann distribution. In this case, it is useful to use the average particle cross section over the velocity distribution. This is entered into the volumetric fusion rate:[6]

${\displaystyle P_{\text{fusion}}=n_{A}n_{B}\langle \sigma v_{A,B}\rangle E_{\text{fusion}}}$

where:

• ${\displaystyle P_{\text{fusion}}}$ is the energy made by fusion, per time and volume
• n is the number density of species A or B, of the particles in the volume
• ${\displaystyle \langle \sigma v_{A,B}\rangle }$ is the cross section of that reaction, average over all the velocities of the two species v
• ${\displaystyle E_{\text{fusion}}}$ is the energy released by that fusion reaction.

The Lawson criterion considers the energy balance between the energy produced in fusion reactions to the energy being lost to the environment. In order to generate usable energy, a system would have to produce more energy than it loses. Lawson assumed an energy balance, shown below.[6]

${\displaystyle P_{\text{out}}=\eta _{\text{capture}}\left(P_{\text{fusion}}-P_{\text{conduction}}-P_{\text{radiation}}\right)}$

where:

• ${\displaystyle P_{\text{out}}}$ is the net power from fusion
• ${\displaystyle \eta _{\text{capture}}}$ is the efficiency of capturing the output of the fusion
• ${\displaystyle P_{\text{fusion}}}$ is the rate of energy generated by the fusion reactions
• ${\displaystyle P_{\text{conduction}}}$ is the conduction losses as energetic mass leaves the plasma
• ${\displaystyle P_{\text{radiation}}}$ is the radiation losses as energy leaves as light.

The rate of fusion, and thus P_fusion, depends on the temperature and density of the plasma. The plasma loses energy through conduction and radiation.[6] Conduction occurs when ions, electrons, or neutrals impact other substances, typically a surface of the device, and transfer a portion of their kinetic energy to the other atoms. The rate of conduction is also based on the temperature and density. Radiation is energy that leaves the cloud as light. Radiation also increases with temperature as well as the mass of the ions. Fusion power systems must operate in a region where the rate of fusion is higher than the losses.

Fusion trapping (left) against temperature (bottom) for various fusion approaches as of 2021, assuming DT fuel[7]

The Lawson criterion argues that a machine holding a thermalized and quasi-neutral plasma has to generate enough energy to overcome its energy losses. The amount of energy released in a given volume is a function of the temperature, and thus the reaction rate on a per-particle basis, the density of particles within that volume, and finally the confinement time, the length of time that energy stays within the volume.[6][8] This is known as the "triple product": the plasma density, temperature, and confinement time.[9]

In magnetic confinement, the density is low, on the order of a "good vacuum". For instance, in the ITER device the fuel density is about 1.0 × 10¹⁹ m⁻³, which is about one-millionth atmospheric density.[10] This means that the temperature and/or confinement time must increase. Fusion-relevant temperatures have been achieved using a variety of heating methods that were developed in the early 1970s. In modern machines, as of 2019, the major remaining issue was the confinement time. Plasmas in strong magnetic fields are subject to a number of inherent instabilities, which must be suppressed to reach useful durations. One way to do this is to simply make the reactor volume larger, which reduces the rate of leakage due to classical diffusion. This is why ITER is so large.

In contrast, inertial confinement systems approach useful triple product values via higher density, and have short confinement intervals. In NIF, the initial frozen hydrogen fuel load has a density less than water that is increased to about 100 times the density of lead. In these conditions, the rate of fusion is so high that the fuel fuses in the microseconds it takes for the heat generated by the reactions to blow the fuel apart. Although NIF is also large, this is a function of its "driver" design, not inherent to the fusion process.

Multiple approaches have been proposed to capture the energy that fusion produces. The simplest is to heat a fluid. The commonly targeted D-T reaction releases much of its energy as fast-moving neutrons. Electrically neutral, the neutron is unaffected by the confinement scheme. In most designs, it is captured in a thick "blanket" of lithium surrounding the reactor core. When struck by a high-energy neutron, the blanket heats up. It is then actively cooled with a working fluid that drives a turbine to produce power.

Another design proposed to use the neutrons to breed fission fuel in a blanket of nuclear waste, a concept known as a fission-fusion hybrid. In these systems, the power output is enhanced by the fission events, and power is extracted using systems like those in conventional fission reactors.[11]

Designs that use other fuels, notably the proton-boron aneutronic fusion reaction, release much more of their energy in the form of charged particles. In these cases, power extraction systems based on the movement of these charges are possible. Direct energy conversion was developed at Lawrence Livermore National Laboratory (LLNL) in the 1980s as a method to maintain a voltage directly using fusion reaction products. This has demonstrated energy capture efficiency of 48 percent.[12]

• Tokamak: the most well-developed and well-funded approach. This method drives hot plasma around in a magnetically confined torus, with an internal current. When completed, ITER will become the world's largest tokamak. As of September 2018 an estimated 226 experimental tokamaks were either planned, decommissioned or operating (50) worldwide.[18]
• Spherical tokamak: also known as spherical torus. A variation on the tokamak with a spherical shape.
• Stellarator: Twisted rings of hot plasma. The stellarator attempts to create a natural twisted plasma path, using external magnets. Stellarators were developed by Lyman Spitzer in 1950 and evolved into four designs: Torsatron, Heliotron, Heliac and Helias. One example is Wendelstein 7-X, a German device. It is the world's largest stellarator.[19]
• Internal rings: Stellarators create a twisted plasma using external magnets, while tokamaks do so using a current induced in the plasma. Several classes of designs provide this twist using conductors inside the plasma. Early calculations showed that collisions between the plasma and the supports for the conductors would remove energy faster than fusion reactions could replace it. Modern variations, including the Levitated Dipole Experiment (LDX), use a solid superconducting torus that is magnetically levitated inside the reactor chamber.[20]
• Magnetic mirror: Developed by Richard F. Post and teams at LLNL in the 1960s.[21] Magnetic mirrors reflect plasma back and forth in a line. Variations included the Tandem Mirror, magnetic bottle and the biconic cusp.[22] A series of mirror machines were built by the US government in the 1970s and 1980s, principally at LLNL.[23] However, calculations in the 1970s estimated it was unlikely these would ever be commercially useful.
• Bumpy torus: A number of magnetic mirrors are arranged end-to-end in a toroidal ring. Any fuel ions that leak out of one are confined in a neighboring mirror, permitting the plasma pressure to be raised arbitrarily high without loss. An experimental facility, the ELMO Bumpy Torus or EBT was built and tested at Oak Ridge National Laboratory (ORNL) in the 1970s.
• Field-reversed configuration: This device traps plasma in a self-organized quasi-stable structure; where the particle motion makes an internal magnetic field which then traps itself.[24]
• Spheromak: Similar to a field-reversed configuration, a semi-stable plasma structure made by using the plasmas' self-generated magnetic field. A spheromak has both toroidal and poloidal fields, while a field-reversed configuration has no toroidal field.[25]
• Dynomak is a spheromak that is formed and sustained using continuous magnetic flux injection.[26][27][28]
• Reversed field pinch: Here the plasma moves inside a ring. It has an internal magnetic field. Moving out from the center of this ring, the magnetic field reverses direction.

Plot of NIF results from 2012 to 2021

• Indirect drive: Lasers heat a structure known as a Hohlraum that becomes so hot it begins to radiate x-ray light. These x-rays heat a fuel pellet, causing it to collapse inward to compress the fuel. The largest system using this method is the National Ignition Facility, followed closely by Laser Mégajoule.[29]
• Direct drive: Lasers directly heat the fuel pellet. Notable direct drive experiments have been conducted at the Laboratory for Laser Energetics (LLE) and the GEKKO XII facilities. Good implosions require fuel pellets with close to a perfect shape in order to generate a symmetrical inward shock wave that produces the high-density plasma.
• Fast ignition: This method uses two laser blasts. The first blast compresses the fusion fuel, while the second ignites it. As of 2019 this technique had lost favor for energy production.[30]
• Magneto-inertial fusion or Magnetized Liner Inertial Fusion: This combines a laser pulse with a magnetic pinch. The pinch community refers to it as magnetized liner inertial fusion while the ICF community refers to it as magneto-inertial fusion.[31]
• Ion Beams: Ion beams replace laser beams to heat the fuel.[32] The main difference is that the beam has momentum due to mass, whereas lasers do not. As of 2019 it appears unlikely that ion beams can be sufficiently focused spatially and in time.
• Z-machine: Sends an electric current through thin tungsten wires, heating them sufficiently to generate x-rays. Like the indirect drive approach, these x-rays then compress a fuel capsule.

• Z-pinch: A current travels in the z-direction through the plasma. The current generates a magnetic field that compresses the plasma. Pinches were the first method for human-made controlled fusion.[33][34] The z-pinch has inherent instabilities that limit its compression and heating to values too low for practical fusion. The largest such machine, the UK's ZETA, was the last major experiment of the sort. The problems in z-pinch led to the tokamak design. The dense plasma focus is a possibly superior variation.
• Theta-pinch: A current circles around the outside of a plasma column, in the theta direction. This induces a magnetic field running down the center of the plasma, as opposed to around it. The early theta-pinch device Scylla was the first to conclusively demonstrate fusion, but later work demonstrated it had inherent limits that made it uninteresting for power production.
• Sheared Flow Stabilized Z-Pinch: Research at the University of Washington under Uri Shumlak investigated the use of sheared-flow stabilization to smooth out the instabilities of Z-pinch reactors. This involves accelerating neutral gas along the axis of the pinch. Experimental machines included the FuZE and Zap Flow Z-Pinch experimental reactors.[35] In 2017, British technology investor and entrepreneur Benj Conway, together with physicists Brian Nelson and Uri Shumlak, co-founded Zap Energy to attempt to commercialize the technology for power production.[36][37][38]
• Screw Pinch: This method combines a theta and z-pinch for improved stabilization.[39]

• Fusor: An electric field heats ions to fusion conditions. The machine typically uses two spherical cages, a cathode inside the anode, inside a vacuum. These machines are not considered a viable approach to net power because of their high conduction and radiation losses.[40] They are simple enough to build that amateurs have fused atoms using them.[41]
• Polywell: Attempts to combine magnetic confinement with electrostatic fields, to avoid the conduction losses generated by the cage.[42]

• Magnetized target fusion: Confines hot plasma using a magnetic field and squeezes it using inertia. Examples include LANL FRX-L machine,[43] General Fusion (piston compression with liquid metal liner), HyperJet Fusion (plasma jet compression with plasma liner).[44][45]
• Uncontrolled: Fusion has been initiated by man, using uncontrolled fission explosions to stimulate fusion. Early proposals for fusion power included using bombs to initiate reactions. See Project PACER.
• Beam fusion: A beam of high energy particles fired at another beam or target can initiate fusion. This was used in the 1970s and 1980s to study the cross sections of fusion reactions.[5] However beam systems cannot be used for power because keeping a beam coherent takes more energy than comes from fusion.
• Muon-catalyzed fusion: This approach replaces electrons in diatomic molecules of isotopes of hydrogen with muons—more massive particles with the same electric charge. Their greater mass compresses the nuclei enough such that the strong interaction can cause fusion.[46] As of 2007 producing muons required more energy than can be obtained from muon-catalyzed fusion.[47]

A deep reinforcement learning system has been used to control a tokamak-based reactor. The AI was able to manipulate the magnetic coils to manage the plasma. The system was able to continuously adjust to maintain appropriate behavior (more complex than step-based systems). In 2014, Google began working with California-based fusion company TAE Technologies to control the Joint European Torus (JET) to predict plasma behavior.[49] DeepMind has also developed a control scheme with TCV.[50]

• Electrostatic heating: an electric field can do work on charged ions or electrons, heating them.[51]
• Neutral beam injection: hydrogen is ionized and accelerated by an electric field to form a charged beam that is shone through a source of neutral hydrogen gas towards the plasma which itself is ionized and contained by a magnetic field. Some of the intermediate hydrogen gas is accelerated towards the plasma by collisions with the charged beam while remaining neutral: this neutral beam is thus unaffected by the magnetic field and so reaches the plasma. Once inside the plasma the neutral beam transmits energy to the plasma by collisions which ionize it and allow it to be contained by the magnetic field, thereby both heating and refueling the reactor in one operation. The remainder of the charged beam is diverted by magnetic fields onto cooled beam dumps.[52]
• Radio frequency heating: a radio wave causes the plasma to oscillate (i.e., microwave oven). This is also known as electron cyclotron resonance heating, using for example gyrotrons, or dielectric heating.[53]
• Magnetic reconnection: when plasma gets dense, its electromagnetic properties can change, which can lead to magnetic reconnection. Reconnection helps fusion because it instantly dumps energy into a plasma, heating it quickly. Up to 45% of the magnetic field energy can heat the ions.[54][55]
• Magnetic oscillations: varying electric currents can be supplied to magnetic coils that heat plasma confined within a magnetic wall.[56]
• Antiproton annihilation: antiprotons injected into a mass of fusion fuel can induce thermonuclear reactions. This possibility as a method of spacecraft propulsion, known as antimatter-catalyzed nuclear pulse propulsion, was investigated at Pennsylvania State University in connection with the proposed AIMStar project.

The diagnostics of a fusion scientific reactor are extremely complex and varied.[57] The diagnostics required for a fusion power reactor will be various but less complicated than those of a scientific reactor as by the time of commercialization, many real-time feedback and control diagnostics will have been perfected. However, the operating environment of a commercial fusion reactor will be harsher for diagnostic systems than in a scientific reactor because continuous operations may involve higher plasma temperatures and higher levels of neutron irradiation. In many proposed approaches, commercialization will require the additional ability to measure and separate diverter gases, for example helium and impurities, and to monitor fuel breeding, for instance the state of a tritium breeding liquid lithium liner.[58] The following are some basic techniques.

• Flux loop: a loop of wire is inserted into the magnetic field. As the field passes through the loop, a current is made. The current measures the total magnetic flux through that loop. This has been used on the National Compact Stellarator Experiment,[59] the polywell,[60] and the LDX machines. A Langmuir probe, a metal object placed in a plasma, can be employed. A potential is applied to it, giving it a voltage against the surrounding plasma. The metal collects charged particles, drawing a current. As the voltage changes, the current changes. This makes an IV Curve. The IV-curve can be used to determine the local plasma density, potential and temperature.[61]
• Thomson scattering: "Light scatters" from plasma can be used to reconstruct plasma behavior, including density and temperature. It is common in Inertial confinement fusion,[62] Tokamaks,[63] and fusors. In ICF systems, firing a second beam into a gold foil adjacent to the target makes x-rays that traverse the plasma. In tokamaks, this can be done using mirrors and detectors to reflect light.
• Neutron detectors: Several types of neutron detectors can record the rate at which neutrons are produced.[64][65]
• X-ray detectors Visible, IR, UV, and X-rays are emitted anytime a particle changes velocity.[66] If the reason is deflection by a magnetic field, the radiation is cyclotron radiation at low speeds and synchrotron radiation at high speeds. If the reason is deflection by another particle, plasma radiates X-rays, known as Bremsstrahlung radiation.[67]

Neutron blankets absorb neutrons, which heats the blanket. Power can be extracted from the blanket in various ways:

• Steam turbines can be driven by heat transferred into a working fluid that turns into steam, driving electric generators.[68]
• Neutron blankets: These neutrons can regenerate spent fission fuel.[69] Tritium can be produced using a breeder blanket of liquid lithium or a helium cooled pebble bed made of lithium-bearing ceramic pebbles.[70]
• Direct conversion: The kinetic energy of a particle can be converted into voltage.[21] It was first suggested by Richard F. Post in conjunction with magnetic mirrors, in the late 1960s. It has been proposed for Field-Reversed Configurations as well as Dense Plasma Focus devices. The process converts a large fraction of the random energy of the fusion products into directed motion. The particles are then collected on electrodes at various large electrical potentials. This method has demonstrated an experimental efficiency of 48 percent.[71]
• Traveling-wave tubes pass charged helium atoms at several megavolts and just coming off the fusion reaction through a tube with a coil of wire around the outside. This passing charge at high voltage pulls electricity through the wire.

Parameter space occupied by inertial fusion energy and magnetic fusion energy devices as of the mid-1990s. The regime allowing thermonuclear ignition with high gain lies near the upper right corner of the plot.

Confinement refers to all the conditions necessary to keep a plasma dense and hot long enough to undergo fusion. General principles:

• Equilibrium: The forces acting on the plasma must be balanced. One exception is inertial confinement, where the fusion must occur faster than the dispersal time.
• Stability: The plasma must be constructed so that disturbances will not lead to the plasma dispersing.
• Transport or conduction: The loss of material must be sufficiently slow.[6] The plasma carries energy off with it, so rapid loss of material will disrupt fusion. Material can be lost by transport into different regions or conduction through a solid or liquid.

To produce self-sustaining fusion, part of the energy released by the reaction must be used to heat new reactants and maintain the conditions for fusion.

Magnetic mirror effect. If a particle follows the field line and enters a region of higher field strength, the particles can be reflected. Several devices apply this effect. The most famous was the magnetic mirror machines, a series of devices built at LLNL from the 1960s to the 1980s.[72] Other examples include magnetic bottles and Biconic cusp.[73] Because the mirror machines were straight, they had some advantages over ring-shaped designs. The mirrors were easier to construct and maintain and direct conversion energy capture was easier to implement.[12] Poor confinement led this approach to be abandoned, except in the polywell design.[74]

Magnetic loops bend the field lines back on themselves, either in circles or more commonly in nested toroidal surfaces. The most highly developed systems of this type are the tokamak, the stellarator, and the reversed field pinch. Compact toroids, especially the field-reversed configuration and the spheromak, attempt to combine the advantages of toroidal magnetic surfaces with those of a simply connected (non-toroidal) machine, resulting in a mechanically simpler and smaller confinement area.

Inertial confinement is the use of rapid implosion to heat and confine plasma. A shell surrounding the fuel is imploded using a direct laser blast (direct drive), a secondary x-ray blast (indirect drive), or heavy beams. The fuel must be compressed to about 30 times solid density with energetic beams. Direct drive can in principle be efficient, but insufficient uniformity has prevented success.[75]^:19–20 Indirect drive uses beams to heat a shell, driving the shell to radiate x-rays, which then implode the pellet. The beams are commonly laser beams, but ion and electron beams have been investigated.[75]^:182–193

Electrostatic confinement fusion devices use electrostatic fields. The best known is the fusor. This device has a cathode inside an anode wire cage. Positive ions fly towards the negative inner cage, and are heated by the electric field in the process. If they miss the inner cage they can collide and fuse. Ions typically hit the cathode, however, creating prohibitory high conduction losses. Fusion rates in fusors are low because of competing physical effects, such as energy loss in the form of light radiation.[76] Designs have been proposed to avoid the problems associated with the cage, by generating the field using a non-neutral cloud. These include a plasma oscillating device,[77] a magnetically shielded-grid,[78] a penning trap, the polywell,[79] and the F1 cathode driver concept.[80]

Diagram of the D-T reaction

The easiest nuclear reaction, at the lowest energy, is D+T:

²
₁D
+ ³
₁T
→ ⁴
₂He
(3.5 MeV) + ¹
₀n
(14.1 MeV)

This reaction is common in research, industrial and military applications, usually as a neutron source. Deuterium is a naturally occurring isotope of hydrogen and is commonly available. The large mass ratio of the hydrogen isotopes makes their separation easy compared to the uranium enrichment process. Tritium is a natural isotope of hydrogen, but because it has a short half-life of 12.32 years, it is hard to find, store, produce, and is expensive. Consequently, the deuterium-tritium fuel cycle requires the breeding of tritium from lithium using one of the following reactions:

¹
₀n
+ ⁶
₃Li
→ ³
₁T
+ ⁴
₂He

¹
₀n
+ ⁷
₃Li
→ ³
₁T
+ ⁴
₂He
+ ¹
₀n

The reactant neutron is supplied by the D-T fusion reaction shown above, and the one that has the greatest energy yield. The reaction with ⁶Li is exothermic, providing a small energy gain for the reactor. The reaction with ⁷Li is endothermic, but does not consume the neutron. Neutron multiplication reactions are required to replace the neutrons lost to absorption by other elements. Leading candidate neutron multiplication materials are beryllium and lead, but the ⁷Li reaction helps to keep the neutron population high. Natural lithium is mainly ⁷Li, which has a low tritium production cross section compared to ⁶Li so most reactor designs use breeding blankets with enriched ⁶Li.

Drawbacks commonly attributed to D-T fusion power include:

• The supply of neutrons results in neutron activation of the reactor materials.[82]^:242
• 80% of the resultant energy is carried off by neutrons, which limits the use of direct energy conversion.[83]
• It requires the radioisotope tritium. Tritium may leak from reactors. Some estimates suggest that this would represent a substantial environmental radioactivity release.[84]

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.[85]

In a production setting, the neutrons would react with lithium in the breeding blanket composed of lithium ceramic pebbles or liquid lithium, yielding tritium. The energy of the neutrons ends up in the lithium, which would then be transferred to drive electrical production. The lithium blanket protects the outer portions of the reactor from the neutron flux. Newer designs, the advanced tokamak in particular, use lithium inside the reactor core as a design element. The plasma interacts directly with the lithium, preventing a problem known as "recycling". The advantage of this design was demonstrated in the Lithium Tokamak Experiment.

Deuterium fusion cross section (in square meters) at different ion collision energies

Fusing two deuterium nuclei is the second easiest fusion reaction. The reaction has two branches that occur with nearly equal probability:

D + D	→ T	+ 1H
D + D	→ 3He	+ n

This reaction is also common in research. The optimum energy to initiate this reaction is 15 keV, only slightly higher than that for the D-T reaction. The first branch produces tritium, so that a D-D reactor is not tritium-free, even though it does not require an input of tritium or lithium. Unless the tritons are quickly removed, most of the tritium produced is burned in the reactor, which reduces the handling of tritium, with the disadvantage of producing more, and higher-energy, neutrons. The neutron from the second branch of the D-D reaction has an energy of only 2.45 MeV (0.393 pJ), while the neutron from the D-T reaction has an energy of 14.1 MeV (2.26 pJ), resulting in greater isotope production and material damage. When the tritons are removed quickly while allowing the ³He to react, the fuel cycle is called "tritium suppressed fusion".[86] The removed tritium decays to ³He with a 12.5 year half life. By recycling the ³He decay into the reactor, the fusion reactor does not require materials resistant to fast neutrons.

Assuming complete tritium burn-up, the reduction in the fraction of fusion energy carried by neutrons would be only about 18%, so that the primary advantage of the D-D fuel cycle is that tritium breeding is not required. Other advantages are independence from lithium resources and a somewhat softer neutron spectrum. The disadvantage of D-D compared to D-T is that the energy confinement time (at a given pressure) must be 30 times longer and the power produced (at a given pressure and volume) is 68 times less.

Assuming complete removal of tritium and ³He recycling, only 6% of the fusion energy is carried by neutrons. The tritium-suppressed D-D fusion requires an energy confinement that is 10 times longer compared to D-T and double the plasma temperature.[87]

A second-generation approach to controlled fusion power involves combining helium-3 (³He) and deuterium (²H):

D + 3He

→ 4He

+ 1H

This reaction produces ⁴He and a high-energy proton. As with the p-¹¹B aneutronic fusion fuel cycle, most of the reaction energy is released as charged particles, reducing activation of the reactor housing and potentially allowing more efficient energy harvesting (via any of several pathways).[88] In practice, D-D side reactions produce a significant number of neutrons, leaving p-¹¹B as the preferred cycle for aneutronic fusion.[88]

Both material science problems and non proliferation concerns are greatly diminished by aneutronic fusion. Theoretically, the most reactive aneutronic fuel is ³He. However, obtaining reasonable quantities of ³He implies large scale extraterrestrial mining on the Moon or in the atmosphere of Uranus or Saturn. Therefore, the most promising candidate fuel for such fusion is fusing the readily available protium (i.e. a proton) and boron. Their fusion releases no neutrons, but produces energetic charged alpha (helium) particles whose energy can directly be converted to electrical power:

p + ¹¹B → 3 ⁴He

Side reactions are likely to yield neutrons that carry only about 0.1% of the power,[89]^:177–182 which means that neutron scattering is not used for energy transfer and material activation is reduced several thousand-fold. The optimum temperature for this reaction of 123 keV[90] is nearly ten times higher than that for pure hydrogen reactions, and energy confinement must be 500 times better than that required for the D-T reaction. In addition, the power density is 2500 times lower than for D-T, although per unit mass of fuel, this is still considerably higher than for fission reactors.

Because the confinement properties of the tokamak and laser pellet fusion are marginal, most proposals for aneutronic fusion are based on radically different confinement concepts, such as the Polywell and the Dense Plasma Focus. In 2013, a research team led by Christine Labaune at École Polytechnique, reported a new fusion rate record for proton-boron fusion, with an estimated 80 million fusion reactions during a 1.5 nanosecond laser fire, 100 times greater than reported in previous experiments.[91][92]

SuperOx was able to produce over 186 miles of YBCO wire in nine months for use in fusion reactor magnets, dramatically surpassing the company's previous production targets.

In a plasma that is embedded in a magnetic field (known as a magnetized plasma) the fusion rate scales as the magnetic field strength to the 4th power. For this reason, many fusion companies that rely on magnetic fields to control their plasma are trying to develop high temperature superconducting devices. In 2021, SuperOx, a Russian and Japanese company, developed a new manufacturing process for making superconducting YBCO wire for fusion reactors. This new wire was shown to conduct between 700 and 2000 Amps per square millimeter. The company was able to produce 186 miles of wire in nine months.[101]

Even on smaller production scales, the containment apparatus is blasted with matter and energy. Designs for plasma containment must consider:

• A heating and cooling cycle, up to a 10 MW/m² thermal load.
• Neutron radiation, which over time leads to neutron activation and embrittlement.
• High energy ions leaving at tens to hundreds of electronvolts.
• Alpha particles leaving at millions of electronvolts.
• Electrons leaving at high energy.
• Light radiation (IR, visible, UV, X-ray).

Depending on the approach, these effects may be higher or lower than fission reactors.[102] One estimate put the radiation at 100 times that of a typical pressurized water reactor. Depending on the approach, other considerations such as electrical conductivity, magnetic permeability, and mechanical strength matter. Materials must also not end up as long-lived radioactive waste.[93]

For long term use, each atom in the wall is expected to be hit by a neutron and displaced about 100 times before the material is replaced. High-energy neutrons produce hydrogen and helium via nuclear reactions that tend to form bubbles at grain boundaries and result in swelling, blistering or embrittlement.[102]

Low-Z materials, such as graphite or beryllium are generally preferred to high-Z materials, usually tungsten with molybdenum as a second choice.[99] Liquid metals (lithium, gallium, tin) have been proposed, e.g., by injection of 1–5 mm thick streams flowing at 10 m/s on solid substrates.

Graphite features a gross erosion rate due to physical and chemical sputtering amounting to many meters per year, requiring redeposition of the sputtered material. The redeposition site generally does not exactly match the sputter site, allowing net erosion that may be prohibitive. An even larger problem is that tritium is redeposited with the redeposited graphite. The tritium inventory in the wall and dust could build up to many kilograms, representing a waste of resources and a radiological hazard in case of an accident. Graphite found favor as material for short-lived experiments, but appears unlikely to become the primary plasma-facing material (PFM) in a commercial reactor.[93][103]

Tungsten's sputtering rate is orders of magnitude smaller than carbon's, and tritium is much less incorporated into redeposited tungsten. However, tungsten plasma impurities are much more damaging than carbon impurities, and self-sputtering can be high, requiring the plasma in contact with the tungsten not be too hot (a few tens of eV rather than hundreds of eV). Tungsten also has issues around eddy currents and melting in off-normal events, as well as some radiological issues.[93]

Accident potential and effect on the environment are critical to social acceptance of nuclear fusion, also known as a social license.[104] Fusion reactors are not subject to catastrophic meltdown.[105] It requires precise and controlled temperature, pressure and magnetic field parameters to produce net energy, and any damage or loss of required control would rapidly quench the reaction.[106] Fusion reactors operate with seconds or even microseconds worth of fuel at any moment. Without active refueling, the reactions immediately quench.[105]

The same constraints prevent runaway reactions. Although the plasma is expected to have a volume of 1,000 m³ (35,000 cu ft) or more, the plasma typically contains only a few grams of fuel.[105] By comparison, a fission reactor is typically loaded with enough fuel for months or years, and no additional fuel is necessary to continue the reaction. This large fuel supply is what offers the possibility of a meltdown.[107]

In magnetic containment, strong fields develop in coils that are mechanically held in place by the reactor structure. Failure of this structure could release this tension and allow the magnet to "explode" outward. The severity of this event would be similar to other industrial accidents or an MRI machine quench/explosion, and could be effectively contained within a containment building similar to those used in fission reactors.

In laser-driven inertial containment the larger size of the reaction chamber reduces the stress on materials. Although failure of the reaction chamber is possible, stopping fuel delivery prevents catastrophic failure.[108]

Most reactor designs rely on liquid hydrogen as a coolant and to convert stray neutrons into tritium, which is fed back into the reactor as fuel. Hydrogen is flammable, and it is possible that hydrogen stored on-site could ignite. In this case, the tritium fraction of the hydrogen would enter the atmosphere, posing a radiation risk. Calculations suggest that about 1 kilogram (2.2 lb) of tritium and other radioactive gases in a typical power station would be present. The amount is small enough that it would dilute to legally acceptable limits by the time they reached the station's perimeter fence.[109]

The likelihood of small industrial accidents, including the local release of radioactivity and injury to staff, are estimated to be minor compared to fission. They would include accidental releases of lithium or tritium or mishandling of radioactive reactor components.[108]

A magnet quench is an abnormal termination of magnet operation that occurs when part of the superconducting coil exits the superconducting state (becomes normal). This can occur because the field inside the magnet is too large, the rate of change of field is too large (causing eddy currents and resultant heating in the copper support matrix), or a combination of the two.

More rarely a magnet defect can cause a quench. When this happens, that particular spot is subject to rapid Joule heating from the current, which raises the temperature of the surrounding regions. This pushes those regions into the normal state as well, which leads to more heating in a chain reaction. The entire magnet rapidly becomes normal over several seconds, depending on the size of the superconducting coil. This is accompanied by a loud bang as the energy in the magnetic field is converted to heat, and the cryogenic fluid boils away. The abrupt decrease of current can result in kilovolt inductive voltage spikes and arcing. Permanent damage to the magnet is rare, but components can be damaged by localized heating, high voltages, or large mechanical forces.

In practice, magnets usually have safety devices to stop or limit the current when a quench is detected. If a large magnet undergoes a quench, the inert vapor formed by the evaporating cryogenic fluid can present a significant asphyxiation hazard to operators by displacing breathable air.

A large section of the superconducting magnets in CERN's Large Hadron Collider unexpectedly quenched during start-up operations in 2008, destroying multiple magnets.[110] In order to prevent a recurrence, the LHC's superconducting magnets are equipped with fast-ramping heaters that are activated when a quench event is detected. The dipole bending magnets are connected in series. Each power circuit includes 154 individual magnets, and should a quench event occur, the entire combined stored energy of these magnets must be dumped at once. This energy is transferred into massive blocks of metal that heat up to several hundred degrees Celsius—because of resistive heating—in seconds. A magnet quench is a "fairly routine event" during the operation of a particle accelerator.[111]

The natural product of the fusion reaction is a small amount of helium, which is harmless to life. Hazardous tritium is difficult to retain completely. During normal operation, tritium is continually released.[108]

Although tritium is volatile and biologically active, the health risk posed by a release is much lower than that of most radioactive contaminants, because of tritium's short half-life (12.32 years) and very low decay energy (~14.95 keV), and because it does not bioaccumulate (it cycles out of the body as water, with a biological half-life of 7 to 14 days).[112] ITER incorporates total containment facilities for tritium.[113]

Fusion reactors create far less radioactive material than fission reactors. Further, the material it creates is less damaging biologically, and the radioactivity dissipates within a time period that is well within existing engineering capabilities for safe long-term waste storage.[114] In specific terms, except in the case of aneutronic fusion,[115][116] the neutron flux turns the structural materials radioactive. The amount of radioactive material at shut-down may be comparable to that of a fission reactor, with important differences. The half-lives of fusion and neutron activation radioisotopes tend to be less than those from fission, so that the hazard decreases more rapidly. Whereas fission reactors produce waste that remains radioactive for thousands of years, the radioactive material in a fusion reactor (other than tritium) would be the reactor core itself and most of this would be radioactive for about 50 years, with other low-level waste being radioactive for another 100 years or so thereafter.[117] The fusion waste's short half-life eliminates the challenge of long-term storage. By 500 years, the material would have the same radiotoxicity as coal ash.[109] Nonetheless, classification as intermediate level waste rather than low-level waste may complicate safety discussions.[118][114]

The choice of materials is less constrained than in conventional fission, where many materials are required for their specific neutron cross-sections. Fusion reactors can be designed using "low activation", materials that do not easily become radioactive. Vanadium, for example, becomes much less radioactive than stainless steel.[119] Carbon fiber materials are also low-activation, are strong and light, and are promising for laser-inertial reactors where a magnetic field is not required.[120]

In some scenarios, fusion power technology could be adapted to produce materials for military purposes. A huge amount of tritium could be produced by a fusion power station; tritium is used in the trigger of hydrogen bombs and in modern boosted fission weapons, but it can be produced in other ways. The energetic neutrons from a fusion reactor could be used to breed weapons-grade plutonium or uranium for an atomic bomb (for example by transmutation of ²³⁸
U to ²³⁹
Pu, or ²³²
Th to ²³³
U).

A study conducted in 2011 assessed three scenarios:[121]

• Small-scale fusion station: As a result of much higher power consumption, heat dissipation and a more recognizable design compared to enrichment gas centrifuges, this choice would be much easier to detect and therefore implausible.[121]
• Commercial facility: The production potential is significant. But no fertile or fissile substances necessary for the production of weapon-usable materials needs to be present at a civil fusion system at all. If not shielded, detection of these materials can be done by their characteristic gamma radiation. The underlying redesign could be detected by regular design information verification. In the (technically more feasible) case of solid breeder blanket modules, it would be necessary for incoming components to be inspected for the presence of fertile material,[121] otherwise plutonium for several weapons could be produced each year.[122]
• Prioritizing weapon-grade material regardless of secrecy: The fastest way to produce weapon-usable material was seen in modifying a civil fusion power station. No weapons-compatible material is required during civil use. Even without the need for covert action, such a modification would take about two months to start production and at least an additional week to generate a significant amount. This was considered to be enough time to detect a military use and to react with diplomatic or military means. To stop the production, a military destruction of parts of the facility while leaving out the reactor would be sufficient.[121]

Another study concluded "...large fusion reactors—even if not designed for fissile material breeding—could easily produce several hundred kg Pu per year with high weapon quality and very low source material requirements." It was emphasized that the implementation of features for intrinsic proliferation resistance might only be possible at an early phase of research and development.[122] The theoretical and computational tools needed for hydrogen bomb design are closely related to those needed for inertial confinement fusion, but have very little in common with magnetic confinement fusion.

Fusion power commonly proposes the use of deuterium as fuel and many current designs also use lithium. Assuming a fusion energy output equal to the 1995 global power output of about 100 EJ/yr (= 1 × 10²⁰ J/yr) and that this does not increase in the future, which is unlikely, then known current lithium reserves would last 3000 years. Lithium from sea water would last 60 million years, however, and a more complicated fusion process using only deuterium would have fuel for 150 billion years.[123] To put this in context, 150 billion years is close to 30 times the remaining lifespan of the Sun,[124] and more than 10 times the estimated age of the universe.

The first machine to achieve controlled thermonuclear fusion was a pinch machine at Los Alamos National Laboratory called Scylla I at the start of 1958. The team that achieved it was led by a British scientist named James Tuck and included a young Marshall Rosenbluth. Tuck had been involved in the Manhattan project, but had switched to working on fusion in the early 1950s. He applied for funding for the project as part of a White House sponsored contest to develop a fusion reactor along with Lyman Spitzer. Spitzer went on to develop the Stellarator concept at Princeton. The previous year, 1957, the British had (in spectacular fashion) claimed that they had achieved thermonuclear fusion reactions on the Zeta pinch machine. But this turned out that the neutrons they had detected were from beam-target interactions, not fusion, and the UK had to embarrassingly withdraw the claim.

Scyalla I was a classified machine at the time, so the fact that Mankind had demonstrated controlled nuclear fusion in bulk was hidden from the public. A traditional Z-pinches passes a current down the center of a plasma, which makes a magnetic force around the outside which squeezes the plasma to fusion conditions. Scylla was a twist on this basic idea; it was a θ-pinch, which passes a current around the outside creating a magnetic force in the center [34] Scylla I was a θ-pinch machine, with a cylinder full of deuterium.[33][34] After the success of Scylla-1, Los Alamos went on to build multiple pinch machines over the next few years. But the problem with pinches were that they were unstable because of a whole host of instabilities.

Lots of fusion approaches were pursued in parallel during this period and Princeton emerged as a hub for Stellarator research. It was pioneered by Lyman Spitzer. While fusion did not immediately transpire, the effort led to the creation of the Princeton Plasma Physics Laboratory.[207][208]

The concept of the tokamak originated in 1950–1951 from I.E. Tamm and A.D. Sakharov in the Soviet Union. The tokamak essentially combined a low-power pinch device with a low-power stellarator.[182]

A.D. Sakharov's group constructed the first tokamaks, achieving the first quasistationary fusion reaction.[209]^:90

Laser fusion was suggested in 1962 by scientists at Lawrence Livermore National Laboratory (LLNL), shortly after the invention of the laser in 1960. Inertial confinement fusion (using lasers) research began as early as 1965.

Shiva laser, 1977, the largest ICF laser system built in the seventies

The Tandem Mirror Experiment (TMX) in 1979

Several laser systems were built at LLNL. These included the Argus, the Cyclops, the Janus, the long path, the Shiva laser, and the Nova.[210]

Laser advances included frequency-tripling crystals that transformed infrared laser beams into ultraviolet beams and "chirping", which changed a single wavelength into a full spectrum that could be amplified and then reconstituted into one frequency.[211] Laser research ate money as well, consuming over one billion dollars in the 1980s.[212]

Over time the "advanced tokamak" concept emerged, which included non-circular plasma, internal diverters and limiters, superconducting magnets, operation in the so-called "H-mode" island of increased stability,[213] and the compact tokamak, with the magnets on the inside of the vacuum chamber.[214][215]

Magnetic mirrors suffered from end losses, requiring high power, complex magnetic designs, such as the baseball coil pictured here.

The Novette target chamber (metal sphere with diagnostic devices protruding radially), which was reused from the Shiva project and two newly built laser chains visible in background

Inertial confinement fusion implosion on the Nova laser during the 1980s was a key driver of fusion development.

The Tore Supra, JET, T-15, and JT-60 tokamaks were built in the 1980s.[216][217] In 1984, Martin Peng of ORNL proposed the spherical tokamak with a much smaller radius.[218] It used a single large conductor in the center, with magnets as half-rings off of this conductor. The aspect ratio fell to as low as 1.2.[219]^:B247[220]^:225 Peng's advocacy caught the interest of Derek Robinson, who built the Small Tight Aspect Ratio Tokamak, (START).[219]

In 1991, the Preliminary Tritium Experiment at the Joint European Torus achieved the world's first controlled release of fusion power.[221]

In 1996, Tore Supra created a plasma for two minutes with a current of almost 1 million amperes, totaling 280 MJ of injected and extracted energy.[222]

In 1997, JET produced a peak of 16.1 MW of fusion power (65% of heat to plasma,[223]) with fusion power of over 10 MW sustained for over 0.5 sec.[224]

The Mega Ampere Spherical Tokamak became operational in the UK in 1999.

"Fast ignition"[225][226] saved power and moved ICF into the race for energy production.

In 2006, China's Experimental Advanced Superconducting Tokamak (EAST) test reactor was completed.[227] It was the first tokamak to use superconducting magnets to generate both toroidal and poloidal fields.

In March 2009, the laser-driven ICF NIF became operational.[228]

In the 2000s, privately backed fusion companies entered the race, including TAE Technologies,[229] General Fusion,[230][231] and Tokamak Energy.[232]

The preamplifiers of the National Ignition Facility. In 2012, the NIF achieved a 500-terawatt shot.

The Wendelstein7X under construction

Example of a stellarator design: A coil system (blue) surrounds plasma (yellow). A magnetic field line is highlighted in green on the yellow plasma surface.

Private and public research accelerated in the 2010s. General Fusion developed plasma injector technology and Tri Alpha Energy tested its C-2U device.[233] The French Laser Mégajoule began operation. NIF achieved net energy gain[234] in 2013, as defined in the very limited sense as the hot spot at the core of the collapsed target, rather than the whole target.[235]

In 2014, Phoenix Nuclear Labs sold a high-yield neutron generator that could sustain 5×10¹¹ deuterium fusion reactions per second over a 24-hour period.[236]

In 2015, MIT announced a tokamak it named the ARC fusion reactor, using rare-earth barium-copper oxide (REBCO) superconducting tapes to produce high-magnetic field coils that it claimed could produce comparable magnetic field strength in a smaller configuration than other designs.[237]

In October, researchers at the Max Planck Institute of Plasma Physics in Greifswald, Germany, completed building the largest stellarator to date, the Wendelstein 7-X (W7-X). The W7-X stellarator began Operational phase 1 (OP1.1) on 10 December 2015, successfully producing helium plasma.[238] The objective was to test vital systems and understand the machine's physics. By February 2016, hydrogen plasma was achieved, with temperatures reaching up to 100 million Kelvin. The initial tests used five graphite limiters. After over 2,000 pulses and achieving significant milestones, OP1.1 concluded on 10 March 2016. An upgrade followed, and OP1.2 in 2017 aimed to test an uncooled divertor. By June 2018, record temperatures were reached. W7-X concluded its first campaigns with limiter and island divertor tests, achieving notable advancements by the end of 2018.[239][240][241] It soon produced helium and hydrogen plasmas lasting up to 30 minutes.[242]

In 2017, Helion Energy's fifth-generation plasma machine went into operation.[243] The UK's Tokamak Energy's ST40 generated "first plasma".[244] The next year, Eni announced a $50 million investment in Commonwealth Fusion Systems, to attempt to commercialize MIT's ARC technology.[245][246][247][248]

In January 2021, SuperOx announced the commercialization of a new superconducting wire with more than 700 A/mm² current capability.[249]

TAE Technologies announced results for its Norman device, holding a temperature of about 60 MK for 30 milliseconds, 8 and 10 times higher, respectively, than the company's previous devices.[250]

In October, Oxford-based First Light Fusion revealed its projectile fusion project, which fires an aluminum disc at a fusion target, accelerated by a 9 mega-amp electrical pulse, reaching speeds of 20 kilometres per second (12 mi/s). The resulting fusion generates neutrons whose energy is captured as heat.[251]

On November 8, in an invited talk to the 63rd Annual Meeting of the APS Division of Plasma Physics,[252] the National Ignition Facility claimed[253] to have triggered fusion ignition in the laboratory on August 8, 2021, for the first time in the 60+ year history of the ICF program.[254][255] The shot yielded 1.3 MJ of fusion energy, an over 8X improvement on tests done in spring of 2021.[253] NIF estimates that 230 kJ of energy reached the fuel capsule, which resulted in an almost 6-fold energy output from the capsule.[253] A researcher from Imperial College London stated that the majority of the field agreed that ignition had been demonstrated.[253]

In November 2021, Helion Energy reported receiving $500 million in Series E funding for its seventh-generation Polaris device, designed to demonstrate net electricity production, with an additional $1.7 billion of commitments tied to specific milestones,[256] while Commonwealth Fusion Systems raised an additional $1.8 billion in Series B funding to construct and operate its SPARC reactor, the single largest investment in any private fusion company.[257]

In April 2022, First Light announced that their hypersonic projectile fusion prototype had produced neutrons compatible with fusion. Their technique electromagnetically fires projectiles at Mach 19 at a caged fuel pellet. The deuterium fuel is compressed at Mach 204, reaching pressure levels of 100 TPa.[258]

On December 13, 2022, the US Department of Energy reported that researchers at the National Ignition Facility had achieved a net energy gain from a fusion reaction. The reaction of hydrogen fuel at the facility produced about 3.15 MJ of energy while consuming 2.05 MJ of input. However, while the fusion reactions may have produced more than 3 megajoules of energy—more than was delivered to the target—NIF's 192 lasers consumed 322 MJ of grid energy in the conversion process.[259][2][260][261]

In May 2023, the United States Department of Energy (DOE) provided a grant of $46 million to eight companies across seven states to support fusion power plant design and research efforts. This funding, under the Milestone-Based Fusion Development Program, aligns with objectives to demonstrate pilot-scale fusion within a decade and to develop fusion as a carbon-neutral energy source by 2050. The granted companies are tasked with addressing the scientific and technical challenges to create viable fusion pilot plant designs in the next 5-10 years. The recipient firms include Commonwealth Fusion Systems, Focused Energy Inc., Princeton Stellarators Inc., Realta Fusion Inc., Tokamak Energy Inc., Type One Energy Group, Xcimer Energy Inc., and Zap Energy Inc.[262]