The twist-coil alternative to tokamaks — a machine that confines plasma with external magnets alone, trading engineering complexity for physics simplicity.
The stellarator is the oldest magnetic-confinement fusion concept, first proposed by the American astrophysicist Lyman Spitzer in 1951 — predating the tokamak by several years. Like a tokamak, it confines hot plasma inside a toroidal (doughnut-shaped) magnetic cage. Unlike a tokamak, it creates the entire confining field using external coils alone, with no need for a large current flowing through the plasma itself.1
In any toroidal magnetic device, the magnetic field is stronger on the inside of the doughnut than on the outside. This imbalance causes charged particles to drift sideways, and they would quickly escape if nothing corrected the drift. A tokamak fixes this by driving a current through the plasma, which twists the magnetic field lines into helical (corkscrew) paths so that particles spend equal time on the inside and outside, canceling the drift on average.
A stellarator achieves the same helical twist — but does it entirely with the shape of the external magnets. The result is a set of coils so elaborately sculpted that they resemble modern art more than conventional engineering.2
Early stellarators used simple helical windings wrapped around a circular vacuum vessel. Modern stellarators take a more sophisticated approach. The flagship device, Wendelstein 7-X (W7-X) at the Max Planck Institute for Plasma Physics in Greifswald, Germany, uses 50 non-planar superconducting coils, each with a unique three-dimensional shape, arranged in a five-fold symmetric pattern. These coils were designed using advanced computational optimization to produce a magnetic field with specific mathematical properties that minimize particle losses and energy leakage.3
For decades, stellarators lagged behind tokamaks in plasma performance because their complex three-dimensional fields allowed too many particles to escape. The breakthrough came with the development of powerful computers and optimization algorithms in the 1980s and 1990s. Physicists learned to shape the magnetic field so that the drift orbits of trapped particles close on themselves — a property called quasi-symmetry or quasi-isodynamicity, depending on the specific approach.4
W7-X was designed with quasi-isodynamic optimization and began plasma operations in 2015. By 2023, it had demonstrated plasma confinement times and energy densities competitive with similarly sized tokamaks, vindicating decades of theoretical work.
The price of physics elegance is engineering complexity. Each coil in a modern stellarator has a unique, intricate shape that must be manufactured to millimeter precision. The vacuum vessel, shaped to follow the twisting plasma, is far harder to build than a tokamak's relatively simple torus. Assembly and maintenance — particularly replacing components inside the vessel — pose formidable challenges because there is no axis of simple symmetry to exploit.
These manufacturing difficulties have historically made stellarators more expensive per unit of plasma volume than tokamaks, although advances in additive manufacturing (3D printing) and precision winding technologies are beginning to change the calculus.5
Perhaps the stellarator's greatest advantage for a future power plant is its natural ability to run in steady state. A tokamak must find ways to sustain its plasma current continuously — an unsolved engineering problem at reactor scale. A stellarator simply keeps its magnet coils energized; the confining field exists as long as current flows in the external coils, which superconducting magnets can maintain indefinitely with negligible power consumption.
Several stellarator concepts are under active development worldwide, including quasi-helically symmetric designs in the United States (HSX at the University of Wisconsin) and compact commercial concepts from private companies. If the manufacturing challenges can be tamed, the stellarator's disruption-free, steady-state operation could make it a compelling alternative — or complement — to the tokamak on the path to fusion power.