Research into fusion has gone down a blind alley, but a means of escape may now be at hand
IN THE winter of 1968 three British physicists went to Moscow to examine a machine called a tokamak. This fusion reactor was a newly devised competitor to America’s approach to fusion, known as the stellarator. The Russians said the tokamak left the stellarator in the dust. The Americans demurred. But the British found that the Russians were right. The tokamak was far better than the stellarators of the day at holding in place the hot soup of atomic nuclei and electrons, called plasma, that is fusion’s fuel. Stellarators thus dwindled, and the tokamak became the preferred way to try to turn fusion into a practical and useful technology.
Fusion’s promise was of copious, safe, clean power generated from deuterium, a heavy isotope of hydrogen that makes up about 0.016% of the “H” in “H2O”, and tritium, an even heavier form of hydrogen that can be made easily from lithium. Fusing deuterium and tritium generates helium (and also a neutron), together with a lot of energy. But that promise has not been fulfilled. An old joke—that commercial fusion is 30 years away, and always will be—is more true than funny. The latest tokamak, the International Thermonuclear Experimental Reactor, or ITER, being built in France, will (according to current plans) open for fusion a decade late, in 2027, at a cost of at least $15 billion. That is more than twice the original price tag. No one seriously expects a commercial successor before the middle of the century.
In recent years, though, something curious has happened. The sidelined stellarator has started to make a comeback, as computing power almost unimaginable in the 1960s has been brought to bear on the difficulties that dogged it. There is no guarantee that it will now succeed where the tokamak failed. But real hope, rather than the fingers-crossed-behind-the-back sort, is coursing through the fusion fraternity. For, in November, a German stellarator called the Wendelstein 7-X will start operating. And the Wendelstein 7-X is the first stellarator which can, according to that computing power, create perfectly the magnetic fields required for fusion.
The ideal and the good
Atomic nuclei are positively charged. Like charges repel. It is therefore hard to force two nuclei into sufficient proximity for the strong nuclear force, which holds nuclei together, to exceed the repulsive power of electromagnetism—thus permitting the nuclei in question to fuse into one. Temperatures of millions of degrees are needed to make nuclei move too fast for the repulsion to matter. High pressure, to concentrate them and increase the chances that they will encounter each other, also helps.
Controlling such hot, pressurised plasma—in particular, bottling it up so that it cannot touch the wall of its chamber and thus lose heat (and also damage the wall)—requires magnetic fields. If these fields are not perfect, the plasma will leak out.
Tokamaks, which have hollow, doughnut-shaped fusion chambers, do their bottling with two magnetic fields. One is generated by superconducting electromagnets that loop around the chamber and through its central hole (see diagram). The other comes from an electrical current induced in the plasma itself. This simple combination creates magnetic lines of force that corkscrew around the plasma, confining it as a smaller doughnut-within-the-doughnut. Cranking up the fields’ strengths creates an ever-denser doughnut, which increases the plasma’s temperature and pressure until it reaches the point where the nuclei within can fuse.
The price paid for a tokamak’s simplicity, though, is that the field weakens towards its outside edge, and its lines of force tend to drift. The plasma drifts with them and, as a result, sometimes touches the chamber wall. By contrast, the fusion chamber of a stellarator and the magnets that surround it look like something Gaudí might have imagined: a mess of twists, turns and asymmetries. In theory, this complexity means that drift in one part of the chamber is offset in another, differently oriented part. On a full circuit of the chamber, the plasma is squeezed evenly all the way around.
In the 1960s designing and building stellarators was an art-form as much as a science. Hence the preference for tokamaks. But supercomputers and precision engineering have changed that. The reasons for preferring tokamaks to stellarators may thus have vanished. The Wendelstein 7-X will be the test of this.
Fingers will still be crossed, of course. Computer models are not reality, as an American project called the National Ignition Facility has discovered to its cost. (NIF is designed to carry out what is called “inertial confinement”, by hitting pellets of frozen deuterium and tritium hard with lasers, to heat and compress them at the same time. It fits its design specifications perfectly, but still refuses to generate more energy than it consumes.) Earlier experiments with a smaller stellarator do however mean that the machine’s masters at the Max Planck Institute for Plasma Physics are pretty confident.
Even if the Wendelstein 7-X does perform as predicted, though, the behemoth that is ITER will not go away. The fallacy of sunk costs and the national pride of the host and the other participants in the project will see to that. But ITER may find itself relegated from being the flagship of fusion to acting as a proving ground for technology, such as neutron-resistant materials, that ends up being used in stellarators.
None of this, meanwhile, answers the question of why fusion power is needed at all. Even if stellarators work well, the 30-year rule, or something pretty close to it, is likely to apply. And, by the middle of the century, the world’s energy landscape will probably look completely different from now. Perhaps there will, indeed, be a gaping hole in supply that only fusion can plug. More likely, cheap photovoltaic and energy-storage technology will mean that much of humanity’s energy comes from a different fusion reactor—one 150m kilometres away, called the sun.