First Tokamak

Russian tokamak T-1, 1958.

In the early years of magnetic confinement research, various devices such as pinches, stellarators and magnetic traps were developed in parallel. In the Soviet Union (now Russia), scientists first explored the tokamak principle in small tabletop experiments using toroidal ceramic vessels. Because the ceramic material caused plasma contamination, the T-1 tokamak with a steel chamber was constructed.

This first tokamak started operation in 1958 and demonstrated the potential of the concept for plasma confinement. The name stands for “TOroidalnaja KAmera i MAgnitnyje Katushki” — toroidal chamber with magnetic coils — and was probably suggested by Igor Nikolaevich Golovin, deputy director of the Laboratory of Measuring Apparatus, which later became part of today’s Kurchatov Institute. Many larger and smaller tokamaks were subsequently built.

One of the small devices, TM-1 (“tokamak malyj”, or small tokamak), was built in the early 1960s and is now the oldest tokamak still in operation. After numerous upgrades, reconstructions and relocations, it operates as the GOLEM educational tokamak at the Czech Technical University in Prague.

Temperature Record at T-3

A major breakthrough came with the larger T-3 tokamak, which achieved plasma temperatures of about 10 million kelvin. The results, presented in 1968, were independently confirmed in 1969 by a British team from Culham Laboratory using laser Thomson scattering. The success of T-3 convinced fusion laboratories around the world to pursue the tokamak concept; at Princeton, the Model C stellarator was converted into the ST tokamak, helping to trigger the worldwide tokamak boom.

Tokamak Boom

National Spherical Torus Experiment (NSTX) at the Princeton Plasma Physics Laboratory, 2009.

CLEO, built at Culham in the United Kingdom, began experiments with active plasma stabilisation in 1971. It was soon followed by other tokamaks, including the Soviet TO-2 (1972), Japan’s JFT-2 (1973) and the French TFR (1973). Experiments also revealed the danger of plasma disruptions and relativistic runaway electrons. During a disruption, plasma energy and current can be lost very rapidly, producing severe loads on tokamak components and, in some cases, accelerating electrons to extremely high energies.

External Heating

Cutaway view of the SPARC tokamak design.

Neutral beam heating was proposed at the Princeton Plasma Physics Laboratory in 1964 and later applied to tokamak experiments such as ORMAK at Oak Ridge. In this method, ions are accelerated to high energies, neutralised and then injected into the plasma. Neutral beam injection subsequently became one of the principal auxiliary heating methods used in large tokamaks.

Radio-frequency heating methods were also developed. Electron cyclotron resonance heating (ECRH) uses microwave radiation to heat electrons, while ion cyclotron resonance heating (ICRH) uses lower-frequency radio waves to transfer energy primarily to ions.

H-mode

Despite progress in external heating, experiments on PLT and other tokamaks showed that increasing the heating power could degrade energy confinement in the low-confinement regime. A major breakthrough came in 1982, when the German ASDEX tokamak discovered a regime with substantially improved confinement. It became known as high-confinement mode (H-mode), while the previous regime was termed low-confinement mode (L-mode).

In H-mode, an edge transport barrier forms, producing steep density and temperature gradients and reducing turbulent transport. However, H-mode can also give rise to edge-localised modes (ELMs), quasi-periodic instabilities that eject bursts of particles and energy towards the divertor and other plasma-facing components. Large ELMs could cause unacceptable transient heat loads in a fusion reactor.

Quiescent H-mode, an ELM-free high-confinement regime, was later discovered on the DIII-D tokamak. ELMs can contribute to particle and impurity exhaust, but large ELMs are undesirable because of their intense heat loads. Research therefore focuses on mitigating or suppressing them while maintaining adequate impurity control. One important technique is the application of resonant magnetic perturbations (RMPs), which can modify or suppress ELM activity.

Divertor

Early tokamaks commonly used limiters to define the plasma boundary, but direct plasma interaction with limiter surfaces could introduce impurities. Divertor configurations offered a better solution by shaping the edge magnetic field so that open field lines guide particles and heat away from the confined plasma towards dedicated target plates. Divertors had already been explored in stellarators before being adopted by tokamaks.

Divertor configurations were introduced into tokamaks during the 1970s, including devices such as TO-2 (Russia, 1972) and DITE (Divertor Injection Tokamak Experiment, Culham, 1979). By directing edge particles and impurities towards dedicated exhaust regions, divertors improved plasma purity and became an essential feature of modern high-performance tokamaks.

Joint European Torus

ITER tokamak construction site.

By the early 1970s, it had become clear that reactor-scale tokamaks would need to be large enough to accommodate substantial shielding, cooling and other systems between the plasma and the magnets. Experiments also showed that larger plasma volumes generally provide more favourable energy-confinement scaling. These considerations led to the construction of the Joint European Torus (JET), which began operation at Culham in the United Kingdom in 1983.

JET was one of the largest tokamaks of its time and was designed with a D-shaped plasma cross-section that allowed advanced plasma shaping. In 1991, it became the first tokamak to demonstrate controlled fusion power using a deuterium-tritium fuel mixture. Because tritium and fusion neutrons activated in-vessel components, extensive remote-handling technology was required for maintenance.

JET pioneered remote maintenance using remotely operated manipulators to replace activated in-vessel components. In 1997, it produced 16.1 MW of fusion power and reached a fusion gain factor of Q ≈ 0.67, the closest any tokamak has come to scientific breakeven. In its final deuterium-tritium campaign in 2023, JET also set a record of 69.26 MJ of fusion energy released in a single pulse before ending experimental operation.

Tokamak Fusion Test Reactor

Model of the ITER tokamak.

Another tokamak capable of deuterium-tritium experiments was the American Tokamak Fusion Test Reactor (TFTR), which began operation in 1982. Although designed with scientific breakeven as a major goal, it did not achieve it. TFTR nevertheless set important performance records, including an ion temperature of about 510 million kelvin in 1995. JT-60 later achieved similarly exceptional ion temperatures and record values of the fusion triple product.

Tokamaks have also made major progress in long-pulse operation. In January 2025, EAST sustained high-confinement plasma for 1,066 seconds, and in February 2025 WEST set a plasma-duration record of 1,337 seconds. Another key measure of fusion performance is the fusion triple product — the product of plasma density, temperature and energy confinement time. JT-60 achieved a world-record triple product in 1997, demonstrating how closely tokamak plasmas were approaching reactor-relevant conditions.

Superconductors

ASDEX tokamak.

Early tokamaks used copper magnetic coils, but longer pulses increased resistive heating and electrical power consumption, encouraging the development of superconducting magnets. The Soviet T-7, commissioned in 1979, was the first tokamak equipped with superconducting toroidal field coils. The French Tore Supra, which achieved first plasma in 1988, later demonstrated the advantages of superconducting magnets for long-pulse operation.

China’s EAST and Korea’s KSTAR were among the first tokamaks to use superconducting magnets for both toroidal and poloidal field systems; EAST achieved first plasma in 2006 and KSTAR in 2008. Niobium-titanium (NbTi) is widely used where moderate magnetic fields are sufficient, while the more brittle niobium-tin (Nb₃Sn) is suitable for higher-field applications. ITER uses Nb₃Sn in its toroidal field coils and central solenoid, while its poloidal field and correction coils use NbTi.

High-temperature superconductors such as REBCO (Rare-Earth Barium Copper Oxide) can carry high currents in very strong magnetic fields and operate at substantially higher temperatures than NbTi or Nb₃Sn. For high-field fusion magnets, REBCO is typically operated well below its critical temperature, often around 20 K, and is being developed for compact high-field fusion devices.

International Thermonuclear Experimental Reactor

The accumulated knowledge of tokamak physics and fusion technology led to ITER, the largest tokamak ever built. Construction of the ITER site began in 2007, and under the current project baseline research operation is planned to start in 2034, followed by deuterium-tritium operation from 2039. ITER is designed to produce 500 MW of fusion power from 50 MW of external plasma heating power, corresponding to a fusion gain factor of Q ≥ 10.