When Will Fusion Power Be Available Commercially?


There are two types of nuclear power: fission and fusion. Traditional nuclear power plants generate electricity with uranium via fission. However, this is not popular due to meltdown risk, nuclear waste, proliferation risk (i.e., bomb), and cost. Fusion, on the other hand, does not have these issues; however, it is still in development.

Figure 1: Fusion reactor SPARC is expected to achieve self-sustaining burning plasma in 2025. (Source: CFS/MIT-PSFC)

“Commercial fusion” refers to generating electricity at a cost similar to carbon-based sources. This requires the fusion reactor to produce more electricity than it consumes and do so at reasonable cost. Currently, this is not operational; however, experiments conducted over the last 10 years suggest that this is feasible. The big question is: When?

Fusion chemistry

Atoms consist of a central nucleus surrounded by electrons, as illustrated in Figure 2. The nucleus is made of protons and neutrons, and the number of protons determines the material type. For example, hydrogen has one proton, helium has two, and lithium has three. Electrons orbit the nucleus and have a negative charge, whereas protons inside the nucleus, stuck to neutrons, have a positive charge.

Figure 2: The atom consists of protons and neutrons in the center, surrounded by electrons.

Matter exists in four different states: solid, liquid, gas, and plasma. This was recognized by the Greeks, who 2,500 years ago said that the world is made of the earth, water, wind, and fire. Fire and lightning are examples of plasma.

One can typically convert a gas to a plasma by heating it to approximately 10,000˚C. This causes electrons to disconnect from their nuclei and float freely in space. One can then heat further, to approximately 150 million degrees Celsius, to fuse nuclei together and release additional heat. This latter step is fusion.

The plasma in a star is similar to the plasma in a tokamak fusion reactor. The sun’s plasma heats Earth’s surface to ~14˚C, and the tokamak’s plasma heats its internal surface to ~600˚C.

The easiest atoms to fuse together are deuterium (1 proton + 1 neutron) and tritium (1 proton + 2 neutrons). These form helium atoms (2 protons + 2 neutrons), isolated neutrons, and more heat, as illustrated in Figure 3. The produced heat can be converted to electricity via a steam turbine or used to make hydrogen gas.

Figure 3: The easiest way to implement fusion is to heat deuterium and tritium gas to ~150 million degrees Celsius. This produces helium, neutrons, and heat.

The fusion machine

The easiest way to fuse deuterium and tritium is with a tokamak reactor, as illustrated in Figure 4. The main components are the toroid cavity, the plasma, the magnets, and the ~1-meter–thick structure that surrounds the plasma, called the “blanket.”

Figure 4: A tokamak fusion reactor uses magnets to confine a plasma, which heats a surrounding blanket.

To achieve commercial fusion, one must evacuate air from the chamber, inject deuterium and tritium gas, add heat to create plasma, move fusion heat from blanket to steam turbine, and produce electricity. For details, please see the U.K. AEA fusion videos.

The quest for commercial fusion

Qplasma is the ratio of the heat output to the heat input. If this is greater than approximately 5:8, then the reactor should be able to maintain fusion without injecting heat. This is referred to as “ignition” or “burning plasma” and is similar to lighting a campfire with a match. After the logs are lit, the match is removed and the fuel burns by itself. This is not occurring now; however, ignition is an important goal and will probably be realized within several years.

The next step is that one needs to produce more electricity than is consumed by the entire system. This is called “engineering breakeven” or “electricity breakeven.”

After that, one needs the cost of electricity produced to be less than that from natural gas or coal. This is called “economic breakeven” or “commercial fusion,” and this is the ultimate goal of the fusion community.

Fusion math

One can define the linear size of a circular toroid as Ro (meters), wherein Ro to the third power is the volume (Ro3). One can also define the magnetic field in the center as B, in units of Tesla (T).

According to fusion physics, the Qplasma heat gain (heatoutput/heatinput) is proportional to Ro3 × B5. Notice we say “is proportional to.” The actual Qplasma value is determined by additional factors, such as plasma temperature, ability to confine the plasma, and fuel type.

The above equation implies that doubling magnetic strength B causes Qplasma to increase 32-fold (25); and doubling linear size Ro causes Qplasma to increase eightfold (23). This also implies that, for a given magnet strength, one can calculate the toroid size needed for ignition.

Fusion machines

Below is a summary of several fusion reactors. The first three in the table are still in development.

Figure 5: Select fusion reactors that are operational or in development

JET is an example of an older initiative (B = 4 T, Ro = 4 meters) that observed confined plasma for several seconds in the 1980s.

ITER is a $25 billion international project (B = 5 T, Ro = 10 meters) that hopes to finish construction in 2025 and hopes to achieve ignition in the 2030s. It is designed to operate for several minutes at a time and remove heat; however, it will not create electricity. A model of ITER is shown in Figure 6, with a person at lower right.

Figure 6: Model of fusion reactor ITER in France that hopes to be operational in 2025. Notice person at lower right (Source: Wikipedia)

ITER will not attempt ignition in the 2020s, as this requires tritium fuel. And tritium makes internal components slightly radioactive and therefore inaccessible for service. Subsequently, ITER will do low-power testing in the 2020s, and then attempt ignition with tritium in the 2030s.

ITER engineers in the 1990s looked at their best magnets and calculated that they needed to build big to get ignition. However, big entails long development times and high costs. According to today’s fusion math, it is less costly and easier to achieve ignition with more powerful magnets and smaller size.

Commonwealth Fusion Systems (CFS) is an MIT spinoff that is currently sitting on $1.8 billion in venture funding and powerful 12-T magnets. This is 2.4× more than ITER’s 5-T magnets (12 T ÷ 5 T), which translates to an 80-fold (2.45) increase in Qplasma, if everything else was equal.

CFS is using its money and magnets to build SPARC (B = 12 T, Ro = 3 meters), a fusion reactor that hopes to achieve ignition in ~2025. To minimize development time, it does not remove heat and runs for only several seconds at a time.

CFS is also working on ARC (B = 9 T, Ro = 5 meters), a reactor that hopes to be operational in the 2030s. ARC is being designed to achieve ignition, remove heat, run continuously, generate electricity, and achieve engineering breakeven. The money raised by CFS is large and indicates that the investment community sees the light at the end of the fusion tunnel.

The fusion landscape

Figure 7 illustrates how Qplasma varies as a function of toroid size Ro and magnetic field strength B.

Figure 7: Qplasma heat gain increases with the toroid size to the third power and the magnetic field to the fifth power.

If Ro is too large, costs increase due to the fact that toroidal volume is proportional to cost. And if Ro is too small, costs increase due to low energy output. Also, if magnet strength B is too small, costs increase due to low Qplasma. In summary, to achieve commercial fusion, one probably needs to be in the vicinity of ARC in the graph.

Cost and time to develop the reactor are somewhat proportional to the toroidal volume. Therefore, if one wants quick data, they need to be small. ITER is the opposite of small and is taking 30 years to develop. This long duration puts it at risk of being obsolete before completed. ITER’s value is that it’s supported much research over many years and made programs like SPARC and ARC possible. However, if $25 billion had been directed by the world’s top fusion scientists’ month to month instead of funding a large development project, we would probably be further along.

CFS’s quest for commercial fusion via advanced magnets is putting pressure on others to keep up or become irrelevant. The British (MAST-U), Chinese (EAST), Koreans (KSTAR), French (WEST), and others are feeling the heat. However, there are still challenges, several of which are summarized below. And it is not clear when these will be resolved.

Challenge No. 1: Confine plasma

To achieve ignition and make electricity at reasonable cost, the plasma needs to be well-confined. This has been a problem due to internal turbulent behavior within the plasma, as well as other issues, and it is not clear when these will be fully resolved.

The amount of time that a particle is confined in plasma is referred to as the “confinement time,” and the Qplasma energy gain is proportional to this time. In other words, if one doubles the confinement time, Qplasma also doubles.

Confinement time is influenced by the shape of the plasma chamber, the placement and number of control magnets around this chamber, the placement and type of sensors that detect plasma, the feedback-control software that controls magnets as a function of sensor readings, the ability to remove helium exhaust gas, and other factors.

Scientists conduct experiments to learn more about confinement. However, for the most part, they have learned all they can from existing reactors. For more knowledge, they need new and improved reactors.

An example of a new device is MAST-U in the U.K. It has 18 control magnets and many sensors and is more capable at plasma research than many of its predecessors. ITER, by comparison, has fewer magnets and an internal shape that was designed 25 years ago. SPARC’s plasma confinement is also modern and is similar to MAST-U.

Figure 8: Side view of MAST-U, a modern fusion reactor in the U.K. (Source: CCFE)

Challenge No. 2: Breed tritium

As noted in the above fusion chemistry equation, tritium is consumed as a fuel. Also, tritium is costly. Therefore, fusion reactors need to make their own tritium. This is referred to as “breeding” and is feasible, as tritium is created when neutrons bump into other atoms such as lithium, as illustrated in Figure 9. In other words, the blanket around the plasma needs to contain lithium and needs to include a mechanism for extracting newly created tritium gas. This is all possible; however, it has never been fully demonstrated.

Figure 9: Neutrons from plasma bombard lithium in blanket to produce helium and more tritium.

Challenge No. 3: Move heat out

The fusion machine generates two types of heat, both of which need to be removed. One is from photons that radiate against the internal surface of the toroid. And the other is from neutrons that penetrate beyond that surface and heat the blanket. Penetration occurs for about a meter; however, most heating occurs within the first 20 cm. The main purpose of the blanket is to capture this heat and move it outward, as illustrated in Figure 10.

Figure 10: Blanket absorbs neutrons, moves heat outward, protects external structure from neutron damage, and makes more tritium via lithium. (Source: MITSPARC)

Typical fusion internal surface heating is 1 MW/m2. This is challenging, especially in the presence of strong magnetic fields and neutron radiation. For reference, a sunbather incurs 0.001 MW/m2 of surfacing heating, and a spaceship re-entering the atmosphere incurs 500× more, as illustrated in Figure 11.

Figure 11: Heating of fusion reactor wall is similar to that of spaceship re-entering Earth’s atmosphere.

Figure 12 shows an example concept of how one might move heat via a blanket. Plasma on the left radiates photons (red) against a silicon carbide metal wall (blue). Behind this wall is a layer of flowing molten lithium lead (purple) and a layer of flowing molten lithium salt (brown). These layers are followed by a ~1-meter–thick reservoir (orange) of molten lithium salt (e.g., FLIBE) that absorbs neutron-based volumetric heat and breeds tritium.

Figure 12: Neutron radiation from plasma is absorbed by blanket and converted to heat.

As noted previously, cost is somewhat proportional to toroidal volume, and most of the volumetric heat is delivered to the first 20 cm of volume within the blanket. Therefore, to be economically viable, the machine must move approximately 20 MW/m3 of energy into this volume via neutrons and move it out via flowing liquid. The average American home consumes ~0.001 MW of electricity; therefore, this is like moving 20,000 homes worth of electricity into a cubic meter via radiation and then moving it out via flowing liquid.

Challenge No. 4: Minimize cost due to neutron damage

Neutrons bombard the atomic lattice structure within solid materials and make them brittle over time. Liquids do not have this problem, as they do not have a lattice structure. Subsequently, the first meter of solid components in the blanket needs replacing approximately once a year. And doing so at a low cost is a challenge. One possible approach is to lift the reactor’s upper half, expose the brittle piece and replace and restore the upper half, as illustrated in Figure 13.

Challenge No. 5: Intellectual property agreement

Many parties hold many fusion patents. Subsequently, each party with an important patent could potentially block manufacturers. A grand bargain in which the holders of the top patents agree to share would be helpful. However, some patent owners might demand more than what others are willing to tolerate. One solution is for commercial fusion to occur after patents have expired. 

One might consider agreement on patents to become more important as we get closer to commercial fusion. However, this is not entirely true. If agreement occurs long before this time, engineers are more likely to make use of technology controlled by others, which could increase their productivity. For example, if the world’s fusion engineers had access to CFS magnets today, they could more easily focus on commercial fusion.

Fusion research 

To discover new knowledge, scientists conceive theories that are tested in experiments. In some cases, they use computers to simulate a theory and predict the outcome of an experiment. If a simulation consistently predicts outcome, it is considered to be based on valid theory. The time it takes to conduct a new experiment influences how quickly scientists can validate theories. In the case of fusion, conducting a new experiment often entails building a new fusion reactor, which takes five to 30 years.

One can break fusion down into separate science problems such as magnet development, plasma confinement research, and blanket research. And within each, one can look at the easiest way to conduct experiments. For example, to develop magnets, one needs a test fixture that tests a magnet, not a big fusion reactor. And to test plasma confinement and exhaust, one needs something like MAST-U, a small reactor with relatively simple magnets.

Engineers can do paper-only design of an entire fusion reactor without leaving their desk. They can start with design goals, such as “design a reactor that is commercially viable and uses existing materials.” And they can calculate costs, calculate physical parameters, and simulate things like force due to magnets. Also, they can assemble a list of not-yet-resolved issues.

Fusion engineers have done all of the above. They have broken down the component parts, they have demonstrated them working with experiments, they have done paper-only designs of commercial fusion reactors, and they have assembled lists of issues that need attention. Before CFS developed powerful magnets, paper-only designs could not achieve ignition with existing technology. However, new magnets changed this, and paper-only designs of commercial fusion are starting to take shape. The primary remaining issue is low-cost heat removal in the presence of high magnetic fields and high neutron radiation. And engineers are working on this.

What would it take to get commercial fusion running this decade?

To accelerate fusion research, a foundation or government could potentially set up an additional research fund directed by the world’s top fusion scientist and supported with hundreds of millions of dollars per year in funding. Decarbonization is likely to cost the world trillions of dollars; therefore, more money for fusion is reasonable. Governments and foundations should be asking, “What would a top team do with more money and what would be the likely outcome?” and “How much would it cost to get commercial fusion running this decade?”


Recent advances in magnets and plasma confinement have led to an influx in capital, much of which has gone to CFS, holder of powerful magnets. Money and magnets are a dangerous combination; however, pesky problems might delay success. To accelerate development, foundations and governments should consider more support, possibly directed by the world’s top fusion scientists.

Further reading: AspenCore climate change solutions series

  1. 10 Things You Need to Know to Solve the Climate Change Problem
  2. How the U.S. Government Can Solve the Climate Change Problem for $100b
  3. Turning Deserts into Factories
  4. Mechanizing PV Solar on Land
  5. How to Reduce the Cost of Electrical Power Transmission
  6. How to Cover Buildings with Solar Skins

When Will Fusion Power Be Available Commercially? Source link When Will Fusion Power Be Available Commercially?