What Is Nuclear Fusion? The Reaction That Powers Stars and Not Much Else

A nucleus is held together by the strong nuclear force, and the energy binding each particle in place varies with the size of the nucleus, rising steeply from hydrogen to a maximum around iron and declining slowly after that. Anything that moves a nucleus towards iron releases energy: fusing light nuclei does it from below, and splitting heavy ones does it from above, which is why both fusion and fission produce power. The energy comes out because the products weigh very slightly less than the ingredients, and the missing mass appears as energy according to the relation Einstein published in 1905. The obstacle is that nuclei are positively charged and repel each other fiercely, so getting them close enough for the strong force to take over requires enormous speed, which means enormous temperature, in the range of a hundred million degrees for the easiest reaction, several times hotter than the centre of the sun.

The sun's core is around fifteen million degrees, far below what a reactor needs, and it works anyway for two reasons. The first is quantum tunnelling, which allows nuclei to occasionally pass through the repulsive barrier rather than over it, a process with a very low probability per collision. The second is scale: the core contains an enormous quantity of hydrogen under immense gravitational pressure, and the reaction rate per unit volume is astonishingly low. The power density in the sun's core is only a few hundred watts per cubic metre, less than a compost heap, and the sun is bright because it is unimaginably large. A power station cannot be built that way, which is why terrestrial designs must run much hotter to achieve a useful reaction rate in a small volume, and must use the easiest reaction available rather than the proton chain the sun uses.

The reaction of choice is between deuterium and tritium, two heavy isotopes of hydrogen, which ignites at the lowest temperature of any candidate and produces a helium nucleus and a fast neutron. The practical difficulties follow from that choice:

• •Deuterium is abundant, present in seawater at about one atom in every 6,400 hydrogen atoms, and effectively inexhaustible
• •Tritium is radioactive with a twelve-year half-life and essentially does not exist naturally, so it must be bred inside the reactor by letting the escaping neutrons strike a lithium blanket, which has never been demonstrated at scale
• •At a hundred million degrees nothing can touch the fuel, so it must be held away from any wall, which is what confinement means
• •Magnetic confinement uses powerful fields to hold a plasma in a ring, the dominant design being the tokamak, invented in the Soviet Union in the 1950s
• •Inertial confinement compresses a tiny fuel pellet with lasers so fast that it fuses before it can fly apart
• •The neutrons carry most of the energy, cannot be steered by magnets, and progressively damage and activate the reactor structure, which is a materials problem with no complete solution yet

Two results have changed the conversation. In December 2022 the National Ignition Facility in California produced more energy from a fuel pellet than the laser light delivered to it, achieving scientific breakeven for the first time and repeating it several times since with improving margins. The qualification is substantial: the lasers themselves consume vastly more electricity than they deliver to the target, so the facility as a whole is far from net energy, and it was built for weapons research rather than power. On the magnetic side, the joint European torus in Britain set a record of 69 megajoules over five seconds in its final experiments in 2023 before being retired, and ITER, an international project in southern France involving thirty-five countries, is under construction with the aim of producing ten times the power injected into the plasma, though its schedule has slipped repeatedly and first plasma is not now expected before the mid-2030s. A substantial private sector has emerged alongside, with several billion dollars invested in companies pursuing compact tokamaks with high-temperature superconducting magnets, field-reversed configurations and other approaches.

The attraction is the combination of properties. The fuel is water and lithium, there is no possibility of a runaway reaction since any disturbance stops the process rather than accelerating it, and there is no long-lived high-level waste, since the radioactivity produced is in the reactor structure and decays over roughly a century rather than tens of thousands of years. The power density is high and the output is continuous rather than weather-dependent. Against that sit the difficulties: no reactor has yet produced net electricity, tritium breeding is unproven, the materials that must survive decades of neutron bombardment do not exist yet, and the capital costs are enormous. The honest assessment is that fusion is unlikely to contribute meaningfully to decarbonisation within the timescale that matters for climate, and that it is a plausible energy source for the second half of the century, which is a reason to fund it and not a reason to wait for it.

Fusing light nuclei releases energy because the products weigh slightly less than the ingredients, and it requires around a hundred million degrees to overcome the electrical repulsion between nuclei, far hotter than the sun's core, which compensates with sheer size. Reactors use deuterium and tritium, holding the plasma away from any wall with magnetic fields or compressing pellets with lasers. A laser facility achieved scientific breakeven in 2022, tritium breeding and neutron-resistant materials remain unproven, and ITER expects first plasma in the mid-2030s.