Tokamak

The tokamak emerged in 1958 at the Kurchatov Institute in Moscow not because Soviet physicists were uniquely brilliant, but because magnetic confinement fusion demands a specific geometry. Plasma hot enough for fusion—100 million degrees Celsius, seven times the temperature at the Sun's core—vaporizes any physical container. Steel, tungsten, ceramic, diamond—all sublimate instantly at these temperatures. The only vessel that can hold fusion plasma is a magnetic bottle, and physics constrains that bottle to one shape: a doughnut.

Igor Tam and Andrei Sakharov predicted the geometry in 1951. Igor Golovin coined the term in 1957: toroidalnaya kamera magnitnaya—toroidal chamber magnetic—shortened to tokamak. The name describes the solution. A torus avoids end losses. In a linear chamber, particles racing along magnetic field lines escape through the open ends. In a torus, the field lines close on themselves. Particles spiral endlessly, trapped.

But a simple torus isn't enough. External magnets generate the confining field, but this field alone is unstable. The plasma bulges, kinks, collapses. The breakthrough was realizing you could drive a current through the plasma itself, creating a secondary magnetic field that twists around the primary field like stripes on a barber pole. This helical configuration stabilizes the plasma. The topology—torus plus helical field—emerges not from ingenuity but from physics. This is path-dependence at the design level: once you require toroidal confinement, the helical field follows.

What had to exist first?

The theoretical understanding that fusion requires temperatures where hydrogen nuclei overcome electrostatic repulsion. Two protons repel each other. To force them close enough to fuse, they must collide at speeds corresponding to 100 million degrees. Below this threshold, fusion doesn't happen. This knowledge emerged from thermonuclear weapons research in the late 1940s.

The 1951 realization that magnetic fields could confine plasma without physical contact. The Soviet physicist Lev Artsimovich proved that plasma ions—electrically charged—spiral along magnetic field lines. Strong enough fields can trap them indefinitely. But sustaining those fields requires engineering that didn't exist in 1951.

The capacity to generate and sustain megagauss magnetic fields over cubic meters. The first tokamaks used copper coils cooled by liquid nitrogen. Modern devices use superconducting magnets operating at -269°C, four degrees above absolute zero. Building magnets that generate 5-tesla fields continuously without overheating or quenching required decades of materials science.

Geopolitical declassification. In 1958, the United Nations convened the Second International Conference on the Peaceful Uses of Atomic Energy in Geneva. The Soviet Union, the United States, and the United Kingdom declassified their fusion research simultaneously. Tokamak designs became public knowledge. International collaboration became possible.

The T-1 device at Kurchatov was simple by modern standards: a glass vessel 67 centimeters in diameter with copper magnetic coils, achieving plasma confinement measured in milliseconds. But it demonstrated the principle. Plasma didn't escape. Temperature rose. The torus worked.

Lev Artsimovich led the program from the first experiments in 1951 through iterations that progressively improved confinement time and temperature. T-3, completed in 1962, sustained plasma for 20 milliseconds at 10 million degrees. T-4 in 1968 reached 50 milliseconds at 15 million degrees. By the time tokamak research went fully public in 1968, Soviet devices outperformed all Western alternatives—stellarators, magnetic mirrors, pinch devices. Western fusion programs pivoted to tokamaks within two years.

The tokamak exhibited niche-construction at the international scale. Success in Moscow created selection pressure for tokamaks everywhere. ITER, the 35-nation fusion megaproject under construction in Cadarache, France, is a tokamak. Every commercial fusion startup—Commonwealth Fusion Systems in Massachusetts, Tokamak Energy in the UK, TAE Technologies in California—bases designs on the Soviet geometry from 1958. The ecosystem locked in.

But the tokamak also reveals path-dependence as constraint. The configuration is complex. Plasma current must be driven continuously using external power. Field coils require cryogenic cooling that consumes megawatts. Instabilities—edge-localized modes, sawteeth, disruptions—can terminate confinement instantly, releasing the entire plasma energy into the chamber wall. Managing these instabilities demands real-time control systems processing terabytes of sensor data per second.

Alternative designs exist. Stellarators use purely external magnetic fields, eliminating the need for driven plasma current. Inertial confinement compresses fusion fuel with lasers, avoiding magnetic fields entirely. Magnetized target fusion combines both approaches. Any of these might bypass tokamak limitations. But seven decades of optimization have made tokamaks the default. The research ecosystem has accumulated knowledge in tokamak physics, tokamak engineering, tokamak materials science. Switching to alternatives would require rebuilding this entire knowledge base.

Today, tokamaks approach breakeven: the point where fusion energy output exceeds energy input. JET in the UK achieved 67% of breakeven in December 2022, producing 69 megajoules from 59 megajoules of input. ITER aims for ten times energy gain—producing 500 megawatts from 50 megawatts—by 2035. If fusion becomes economical, the tokamak—a Soviet design from the early space age—will power the grid.

The conditions created the design; the design structured the research ecosystem; the ecosystem became path-dependent. Magnetic confinement fusion is a tokamak problem not because tokamaks are optimal, but because seven decades of accumulated knowledge make switching to alternatives prohibitively expensive. Founder-effects locked in the geometry: the first successful design captured the research ecosystem, and momentum has sustained it ever since. The solution space contains other topologies, but the infrastructure investment makes exploration prohibitively expensive.

The first successful design often becomes the permanent solution—not because it's best, but because it's first.