Tokamak or stellarator: what is the difference?

What the two machines share

Fusion fuel has to be heated to temperatures at which atoms fall apart into a plasma of charged particles, and no wall could survive touching it. Both machines solve this the same way: charged particles follow magnetic field lines, so a strong field shaped into a ring can hold the plasma away from the walls. EUROfusion calls this “magnetic confinement”, in which the hot plasma particles are contained in a magnetic “cage”. The basic idea is explained on How a tokamak works.

A simple ring of field lines is not enough, however. For the cage to hold, the field lines have to wind around the ring in a helix, a little like the stripes on a candy cane. The real difference between a tokamak and a stellarator is how they produce that twist.

The tokamak: a current inside the plasma

The word tokamak is a Russian acronym for a torus-shaped chamber with magnetic coils. Coils standing around the ring produce the field that runs the long way round. The twist comes from a second ingredient: an electric current of millions of amperes flowing through the plasma itself. The Max Planck Institute for Plasma Physics (IPP), which builds both kinds of machine, puts it simply: a tokamak creates part of its magnetic confinement with an electric current flowing in the plasma.

That current is induced by a coil in the hole of the ring, the central solenoid, which works like the primary winding of a transformer, with the plasma as the secondary. In ITER, the international tokamak under construction in France, the central solenoid is designed to initiate and sustain a plasma current of 15 megaamperes for 300 to 500 seconds. In TCV, the same job is done by the ohmic coil, the OH coil, while sixteen poloidal coils shape the plasma (see TCV).

A transformer can only push a current in one direction for a limited time, so a tokamak works in pulses. IPP notes that tokamaks without auxiliary facilities run in pulsed mode. Those auxiliary facilities exist: TCV, for example, sustained a plasma current driven only by microwaves for 1.9 seconds as early as 1999.

The stellarator: a twist built into the coils

The stellarator was conceived in 1951 by the American physicist Lyman Spitzer. It has no transformer and no net current flowing around the plasma. According to IPP, the helical twist of the field lines is produced only by external coils, and that is what makes the stellarator suited to continuous operation.

The price is geometry. A tokamak is symmetric: every slice through the ring looks the same. Without the plasma current, a stellarator has to give up that symmetry, so both the coils and the plasma take on complicated, twisted shapes. IPP describes an advantage hidden in this complexity: the extra freedom makes the magnetic field easier to optimise. Finding a good shape is a mathematical problem, and IPP’s stellarator theory group spent more than ten years searching systematically for an optimised field before Wendelstein 7-X was built. Its predecessor, Wendelstein 7-AS, ran from 1988 to 2002.

Wendelstein 7-X in numbers

Wendelstein 7-X in Greifswald, in north-eastern Germany, is described by IPP as the world’s largest fusion device of the stellarator type. Its magnetic field of 3 tesla comes from 50 non-planar and 20 planar superconducting coils, cooled to about minus 270 degrees Celsius. The machine is 16 metres in diameter and 5 metres high, weighs 725 tonnes and holds a plasma volume of 30 cubic metres. Its first plasma, in helium, was produced on 10 December 2015, and scientific experiments with hydrogen began on 3 February 2016.

The machine is designed for plasma discharges of up to 30 minutes. In its 2025 campaign, which ended on 22 May 2025, it held a world-record value of the so-called triple product – density times temperature times energy confinement time, the key measure of progress towards a power plant – for 43 seconds, a record for long discharges. Its energy turnover, the heating power fed in multiplied by the duration, reached 1.8 gigajoules over a plasma lasting 360 seconds.

IPP is careful about what this means. For short plasma durations of a few seconds, tokamaks – the highest values were reached by JT-60U in Japan and JET in the United Kingdom – remain the clear front-runners. For longer durations, IPP first said Wendelstein 7-X was ahead; in an update of June 2025, taking into account JET pulses of up to 60 seconds, it described the two as joint leaders.

Strengths and weaknesses

Put side by side, the trade-off looks like this:

• Tokamak: simpler, symmetric coils and decades of experience. EUROfusion calls it the most advanced and best investigated fusion device design available today. The cost is the plasma current, which has to be driven and controlled, and pulsed operation unless extra current drive is added.
• Stellarator: no plasma current and the possibility of continuous operation. The cost is complicated coils that are hard to design and build, and a smaller body of experience; IPP itself notes that tokamaks are better studied because of their simpler design.

IPP says the experiments on Wendelstein 7-X are meant to test whether a stellarator could be a technically simpler route to a fusion power plant than a tokamak. Europe’s main line, from JET to ITER and the planned demonstration plant DEMO, remains the tokamak – the line in which medium-sized machines such as TCV take part (see Swiss Plasma Center).

A Polish thread in Greifswald

Wendelstein 7-X was partly assembled by Polish hands. Under an agreement with IPP signed in 2007, the Institute of Nuclear Physics of the Polish Academy of Sciences (IFJ PAN) in Kraków assembled the busbar system that carries current to the superconducting coils and made its ultra-low-resistance electrical connections, between 2007 and 2012. It also helped to design equipment and support structures for the machine’s outer vessel and made 30 sets of mechanical parts for polychromators, instruments used to diagnose the plasma.

Polish researchers work on tokamaks too, including TCV; the wider picture is in How does Poland take part in European fusion research? And if you want to know whether any of these machines has yet produced more energy than it used, see Has fusion ever produced more energy than it used?