WINTER is uncompromising on the plains of north-east Germany. Today a keen wind is blowing in unchecked from the Baltic, and ears are hidden beneath woolly hats. But there is also a brightness in the air, and a cheek-tingling warmth. The sun is out. An orb of burning gas 150 million kilometres away is doing its business.
How we would love to bring that searing power a little closer to home. Harnessing nuclear fusion, the process that fuels the sun, would mean practically limitless, carbon-free energy 鈥 no small deal in a warming world. Trouble is, we have been pursuing this dream for five decades or more. Always, it has seemed another five decades away. Our technology simply does not allow us to reliably command the stars.
Advertisement
This sunny December day on Germany鈥檚 Baltic coast could mark a decisive step towards changing that. The switch is being flipped on the 鈥 a machine you could definitely describe as the world鈥檚 most powerful microwave oven, and perhaps the future of its energy supply, too.
Housed at the Max Planck Institute for Plasma Physics on the outskirts of Greifswald, Wendelstein 7-X is not the largest, furthest advanced or best funded attempt to build a nuclear fusion reactor. Those accolades all belong to the under construction in the south of France. But ITER is beset by delays, cost overruns and even doubts about whether it鈥檚 the right design for the job. A new bloom of alternative fusion projects, the biggest of which is Wendelstein 7-X, is now providing some competition. Can they finally ignite success?
It鈥檚 not difficult to grasp the 鈥渨hy?鈥 of nuclear fusion. The amount of energy the sun beats down on Earth each year on Earth. But the 鈥渉ow?鈥 has always been a little more problematic 鈥 with good reason. 鈥淯ntil about 70 years ago people didn鈥檛 even really know how the sun shined continuously all day long,鈥 says physicist of the Griefswald Institute.
We do now know the basics. Stars like the sun consist largely of hydrogen gas that fuses to make helium, an atomic rearrangement that releases a vast amount of energy. The first difficulty is that hydrogen atoms don鈥檛 like being fused. Within the sun, it takes temperatures of 15 million 掳C and pressures over a 100 billion times those at Earth鈥檚 surface.
That鈥檚 a combination we can鈥檛 hope to mimic on Earth 鈥 so we don鈥檛 try. Hydrogen has two heavier, more readily interacting isotopes, deuterium and tritium. These will fuse into helium at just a couple of times atmospheric pressure, with one gram of deuterium-tritium fuel yielding the combustion heat of over 10 tonnes of coal (see diagram). The catch is this reaction requires even hotter conditions: around 100 million 掳C.
At such searing temperatures, deuterium and tritium exist not as atoms, but as an ionised, charged 鈥減lasma鈥 of atomic nuclei stripped of their electrons.
No conceivable reactor material can withstand the heat of this plasma, so it must be confined by other means. Most commonly magnetic fields are used, although other methods have been tried (see 鈥Driven by inertia鈥). Charged particles move along magnetic field lines, so weave the right shape of magnetic field around a plasma and you can stop its particles escaping and melting the reactor sides 鈥 in theory, at least. In practice, currents within the plasma tend to create their own magnetic fields, making it highly unstable. Creating, containing and sustaining a plasma to ignite fusion for any length of time has been the central frustration of fusion for the past 50 years.
ITER 鈥 the acronym is Latin for 鈥渢he way鈥 鈥 should show the way forward. It represents the combined response of 35 nations hoping to make fusion commercially viable within the next 20 years. Preparatory work began on the project in 1998, with construction of the 15-storey-high, hangar-sized reactor building at Saint-Paul-l猫s-Durance in Provence starting in 2010.
ITER had a predecessor, the Joint European Torus (JET), that still operates at the Culham Centre for Fusion Energy near Oxford, UK. JET, inaugurated in 1983, can keep a plasma stable for seconds at a time, enough for fusion ignition but not 鈥渂reakeven鈥 鈥 the crucial point of creating more energy than is used to fire up the reactor. The plan for ITER is to maintain fusion conditions for minutes at a time, and produce 10 times as much energy as was put in. If all goes smoothly, an even larger follow-up reactor, called DEMO, will actually produce electricity sometime in the 2030s.
But not all is going smoothly. In 2008, the date of ITER鈥檚 first plasma was set for 2017. By 2010, it was 2020. Now it is unlikely to happen before 2025, with breakeven a few years after that, says Steven Cowley, director of JET and a member of the ITER collaboration. The latest meeting of ITER鈥檚 ruling council in November last year until June this year. Meanwhile, the project鈥檚 costs have spiralled from $5 billion to over $21 billion.
Rumbles of discontent are becoming loud. Committees in both the and the have begun to question whether fusion research offers value for money. China, meanwhile, a collaborator on ITER, seems increasingly intent on going it alone, with plans to build its own experimental fusion reactor by 2030.
Hulkin鈥 donuts
Wendelstein 7-X is intended to offer an alternative way. ITER and JET are both tokamaks, devices that contain the fusion plasma within a doughnut shaped vessel surrounded by huge supercooled, superconducting magnets. Wendelstein 7-X is, by contrast, a stellarator 鈥 basically a doughnut-shaped device surrounded by huge supercooled, superconducting magnets.
鈥淚t took the equivalent of 2000 microwave ovens to heat plasma to 1 million 掳C鈥
There is a crucial difference. Whereas magnetic fields in a tokamak have to be generated both inside and outside the plasma for stable operation, in a stellarator both the reactor doughnut and the surrounding magnetic fields have a complex, asymmetric design to ensure that every particle, wherever it is in the plasma, experiences the same force (see diagram). At least in supercomputer simulations of the reactor, this allows for continuous, stable operation indefinitely 鈥 a great boon for a commercial device. 鈥淭he advantage of the stellarator is you switch it on and it operates,鈥 says Remmelt Haange of ITER, who has worked on both that reactor and the Wendelstein 7-X device.
A flash of light lasting about a tenth of a second on 10 December was the first hint that it might actually work, as a helium plasma was injected into the Wendelstein device and heated to 1 million 掳C, using an unprecedented microwave equivalent of 2000 kitchen ovens. Pending the results of further tests over the past few weeks, the plan is to at a ceremony this coming week, on 3 February.
The aim over the next few years is to build up to maintain stable fusion conditions for about half an hour. If that works, another, bigger machine will produce 3 gigawatts of thermal power, and about 1 gigawatt of electricity 鈥 about the same as a medium-size coal-fired power station.
As is the way in fusion research, it鈥檚 a long, slow road 鈥 but at Wendelstein 7-X is convinced the stellarator is the best way forward for fusion. 鈥淚 personally think the chances are high, that鈥檚 why I work here,鈥 he says. 鈥淏ut we鈥檙e nowhere near showing that鈥檚 more than a belief.鈥
Hopefully the evidence should be there by the mid-2020s, around the time ITER is due to start up. 鈥淭he timescales fit rather well,鈥 says Wolf. 鈥淚f ITER then produces net energy gain, we can answer the question of how to proceed.鈥 David Campbell at ITER thinks that if things get that far, it won鈥檛 be an easy decision. 鈥淚t鈥檚 not a simple shoot-out. You鈥檒l be balancing the physics advantages of the stellarator against the engineering advantages of the tokamak,鈥 he says.
But with both ITER and Wendelstein 7-X mere stepping stones towards a commercial fusion reactor, viable fusion is still decades in the future. Could there be a quicker way?
of the Plasma Science and Fusion Center at the Massachusetts Institute of Technology thinks so. 鈥淭he fusion scientific community really wants to see ITER succeed, but on a reasonable timeline,鈥 he says. With an eye on that 鈥渞easonable鈥, in 2014 Whyte and his colleagues one-tenth the size of ITER which, they claim, could reach breakeven before it. The secret lies in advances made in superconductors since both ITER and Wendelstein 7-X were conceived. ITER needs the world鈥檚 largest liquid helium plant to cool its superconducting magnets, which are made out of a niobium-tin alloy. Whyte鈥檚 ARC device 鈥 standing for 鈥渁ffordable, robust, compact鈥 鈥 will have new, commercially available high-temperature superconductors for its magnetic coils, made of barium copper oxide plus a rare earth metal. These require only much cheaper nitrogen cooling.
Otherwise, it鈥檚 largely the same tried-and-tested technology. 鈥淲e are not reinventing the wheel when it comes to the basic science of plasma confinement and stability,鈥 says Whyte. Even so, Whyte estimates a first device capable of producing 200 megawatts of electricity will cost something like $5 billion and take five years to build.
Whyte is not the only one to think this way. is a commercial company exploiting technology pioneered at the Culham Centre, and plans to use the new superconductors to build a spherical tokamak. Similar in conception to ARC, this design squishes the tokamak鈥檚 conventional doughnut geometry into a sphere with a narrow hole running down the middle, a shape rather like a cored apple. The goal is to make a 180-megawatt reactor small enough to fit in a living room, says David Kingham from the company. 鈥淲e are aiming for fusion energy gain in five years, first electricity in 10 years and commercial deployment in 15 years,鈥 he says 鈥 for a total cost of around 拢2 billion.
Others are less confident these 鈥渟mall fusion鈥 projects will succeed. Campbell points out that it took almost 20 years for the superconductors selected for ITER to reach a point where they could be made into large, high-quality magnets. , head of theory for Wendelstein 7-X, worries that any significantly smaller geometry will have much higher heat flows than either ITER or Wendelstein 7-X, where they are already one-fifth the flux at the sun鈥檚 surface. 鈥淚 think it can be done, but it makes it more complicated,鈥 he says. 鈥淚鈥檓 rather sceptical, but I think it鈥檚 good that these projects exist.鈥
The general feeling is that, for a problem as complex as nuclear fusion 鈥 and one with such a potentially huge payoff 鈥 more competition can only be a good thing. Certainly Haange is wishing Wendelstein 7-X well. 鈥淭he machine looks fantastic, it鈥檚 a pleasure to see it up and running,鈥 he says. 鈥淚t鈥檚 like diesel and petrol engines 鈥 they are different principles, but it鈥檚 useful to have both in the end.鈥 Perhaps the race for nuclear fusion is finally about to ignite.
Don鈥檛 mention cold fusion
Far, far away from the galaxy of mainstream nuclear fusion, a small but dedicated band of rebels is still devoted to the heresy of cold fusion. The idea that nuclear fusion reactions actually don鈥檛 need huge temperatures or big kit to happen has no agreed theory to back it up, and has had a bad rep since ultimately unreproducible claims were made back in the 1980s. But rebranded as 鈥渓ow energy nuclear reactions鈥 (LENR), it lives on.
Perhaps the loudest of the mavericks is Italian engineer Andrea Rossi, who says he has been operating his 鈥溾 cold fusion reactors, with a fuel of nickel powder and hydrogen, since 2011. Now in partnership with US company Industrial Heat, Rossi claims to be operating a 1-megawatt reactor producing heat for a secret customer in the US for a one-year trial. Russian researcher Alexander Parkhomov and others have, they say, reverse-engineered the nickel-hydrogen reaction in the E-Cat and generated heat from an unknown reaction, more than could be produced by any chemical process.
Rossi has just been granted a , and the Japanese government has restarted funding for LENR research, apparently on the basis of work by Toyota and Mitsubishi. Few researchers give the results much credence, but they have attracted the attention of serious industrial players outside Japan, too. Airbus is one of the few willing to make a public show of interest, hosting a conference on low energy nuclear reactions in Toulouse, France, last October.
Driven by inertia
Magnetic fields are good at containing the searingly hot plasmas needed to make nuclear fusion work (see main story) 鈥 but not that good. Uncharged neutrons made in fusion reactions still leak out and hit the 鈥渇irst wall鈥 closest to the nuclear reaction, causing radiation damage that means replacing this wall regularly. ITER, the biggest attempt at a fusion reactor yet, currently under construction in the south of France, will have a reactor vessel bounded by 440 removable panels weighing 4 tonnes each, and an elaborate robotic system for handling them.
Inertial confinement is a different approach. It relies on zapping fuel pellets with high-power lasers, compressing the resulting plasma and heating it so quickly that fusion occurs before it can fly apart. This is the method used by the US in Livermore, California, which opened for business in 2009. Its main focus is working out what happens within nuclear weapons rather than power generation, but others have taken the principle and aim to use it for commercial purposes.
Canadian company aims to combine the magnetic and inertial confinement approaches. The plan is to spin out a hollow sphere of liquid metal 鈥 a molten mixture of lead and lithium 鈥 to create a self-replacing cavity. This is then rapidly compressed, setting two spinning balls of plasma on course for collision. The firm has recently raised $23 million from investors to fund the next stage of development. A similar hybrid approach based at the , also in Livermore, has recently been granted $4 million by the US Department of Energy.
This article appeared in print under the headline 鈥淔ired up鈥
