IN MARCH 1951, the president of Argentina, Juan Per贸n, announced the results of a secretive project on Huemul Island in northern Patagonia. His scientists had achieved nuclear fusion, he said, harnessing the reaction that powers the sun to herald a future in which energy would be sold in 鈥渉alf-litre bottles, like milk鈥�. But things soon turned sour when researchers returned from Huemul to report that the whole thing was an expensive, embarrassing fraud.
The Huemul hoax was an extreme case. Arguably, though, it set a pattern for the long quest to harness star power for virtually limitless clean energy here on Earth: audacious claims followed by disappointment, rinse and repeat. It explains the tiresome persistence of the old joke that fusion has always been 30 years away, and always will be.
Yet here we are again. In the past year alone, private fusion firms have received more investment than in the entire history of this enterprise. 鈥淭he feeling among investors is that fusion will happen,鈥� says , a fusion scientist and founder of Fusion Energy Insights, a membership organisation for the energy industry. Some companies are even promising commercial fusion reactors in a decade. 鈥淧rogress is happening very rapidly,鈥� says at Lawrence Livermore National Laboratory (LLNL) in California. 鈥淎s you get closer and closer, things start to take off.鈥�
Advertisement
What is hard to discern, however, is whether recent advances at big, state-funded fusion projects, together with new technologies and reactor designs in development at private firms, really amount to a tipping point. When you carefully unpack the hype, is it plausible that fusion reactors will be supplying electricity to grids by around 2030?
How fusion works
The basic physics of nuclear fusion is straightforward. Fission, which is what existing nuclear power plants do, releases energy from the decay of heavy atoms like uranium. Fusion, by contrast, involves fusing the atomic nuclei of very light elements, usually hydrogen, to form heavier nuclei. The total mass of every new nuclei formed is less than the mass of the original pair that fused to form it, and the missing mass is released as enormous amounts of energy. This is the process that powers the sun, after all.
Making a miniature sun on Earth is far from simple, however, because self-sustaining fusion only happens at the ludicrously high temperatures and pressures found in the cores of stars. Even if you successfully start fusion in the lab, the reaction will quickly fizzle out. 鈥淚t鈥檚 incredibly difficult,鈥� says Lee Margetts, a fusion scientist at the University of Manchester, UK. 鈥淲e鈥檙e trying to do what the sun is doing, without [producing] solar flares, and in a much smaller container.鈥�
Solar flares are intense eruptions of plasma, erratically bursting into space at millions of kilometres per hour. They are a hallmark of the immensely turbulent nature of this material 鈥� the superhot, supercharged state of matter in which fusion takes place in the sun鈥檚 core. 鈥淧lasma is a wily, beautiful state of matter,鈥� says Windridge. As solar flares suggest, its behaviour is also hard to predict and even harder to control.
That is the challenge for those chasing fusion at experimental facilities across the world. In many cases, these use doughnut-shaped reactors called tokamaks, which deploy magnetic fields to levitate and so control the plasma while injecting energy to kick-start a fusion reaction. Only when you get your plasma hot and dense enough for long enough will you reach 鈥渋gnition鈥�, meaning the fusion reaction becomes self-sustaining.
On this, there has been progress. In August, a team led by Yong-Su Na at Seoul National University in South Korea achieved a fusion reaction for 30 seconds at temperatures about six times that of the sun鈥檚 core using the Korea Superconducting Tokamak Advanced Research (KSTAR) device. And in December 2021, at the Joint European Torus (JET) in Oxford, UK, researchers set a new global fusion record by producing 59 megajoules of heat energy in a reaction that lasted 5 seconds. 鈥淚f you could hold it happily for 5 seconds, it suggests you could hold it happily for 5 hours,鈥� says Windridge.
Meanwhile, to avoid the trials of magnetically levitating plasma, Kritcher and her colleagues at LLNL鈥檚 National Ignition Facility (NIF) are pursuing another technique called 鈥渋nertial fusion鈥�. This involves aiming 192 powerful lasers at a peppercorn-sized capsule that implodes and crushes atoms together at its centre, creating a short, sharp burst of fusion. In August last year, NIF reported a world-first self-sustaining nuclear fusion reaction, whereby fusion creates enough heat to make more fusion happen. 鈥淭hat really was a big step forward,鈥� says at Princeton University.
Why is fusion so hard to perfect?
But ignition is just the start. These state-funded fusion experiments were never designed to be power stations. After fusion has briefly been achieved, the reactor is shut down, the results are pored over and the experiment is revamped. We can do fusion, says Cowley. 鈥淲hat we don鈥檛 know is whether we can do it at a cost the consumer wants to pay for their electricity.鈥�
As if to hammer that point home, over the past year, NIF scientists have been unable to repeat their self-sustaining reaction. Tiny imperfections in the imploding capsules can mean fragments of their diamond casing is shattered into the plasma, disrupting the fusion reaction. Oh, and each capsule currently costs around 拢1 million.
Even with the tried-and-tested tokamak design, experimental reactors can only run in short bursts. JET鈥檚 5-second record was limited by the reactor鈥檚 copper coils, which overheat if they are bombarded for too long by the neutrons released during fusion. The unwieldy size and complexity of experimental tokamaks also means they can take decades to build 鈥� and many more years to master. The tokamak at the $22 billion International Thermonuclear Experimental Reactor (ITER) in France, which should fire up in the late 2020s, has been afflicted by the spiralling costs and delays that often plague big international science projects.
On top of all that, nobody has made the leap from ignition to net-energy generation. Obviously, getting more energy out than you put in is a prerequisite for any power station. NIF holds the current record, generating 70 per cent of the energy that was put into the imploding capsule. The team behind ITER says it will get out 10 times the energy that goes in 鈥� but not until 2035. And these figures don鈥檛 take into account all the energy it takes to power the lasers 鈥� or, in the case of a tokamak, all of the energy needed to power the magnets. 鈥淭he performance has to be a heck of a lot better before we get there,鈥� says Cowley.
Tony Donn茅, the programme manager of EUROfusion, a group of national fusion labs that includes JET, reckons we will be into the 2060s before any appreciable amount of electricity is being produced by fusion. All of which may leave you wondering why a profusion of private companies has landed on 鈥渃ommercial fusion in a decade鈥� as a reasonable ambition.
Fusion start-ups
There has certainly been a flurry of investment. According to the from the Fusion Industry Association (FIA) in Washington DC, the 30-odd companies it represents have declared private funding to the tune of $3 billion in the past year, plus an additional $117 million from governments.
One of the firms enjoying significant backing is , based in Massachusetts, which has recently raised $1.8 billion after demonstrating magnets made from a novel high-temperature superconductor that pack twice as much power as those in today鈥檚 best tokamaks. This should mean that you can put more pressure onto the fusion plasma, preventing heat from escaping, and allowing you to build smaller reactors. CFS claims it will have a reactor putting out more energy than goes in by 2025 鈥� and producing electricity continuously by early 2030. 鈥淲e鈥檙e not so much in discovery mode, we鈥檙e in execution mode,鈥� says Bob Mumgaard, the company鈥檚 CEO.
He is keen to draw parallels with recent advances in human space flight spurred by private industry. 鈥淣ASA鈥檚 Space Shuttle and SpaceX鈥檚 rocket are following the same laws of physics designed with the same tools, but executed for different goals by different types of organisations,鈥� says Mumgaard. By modernising rocket technology, and making rockets reusable, the likes of SpaceX have made space flight much cheaper.
Mumgaard鈥檚 optimism also stems from the fact that the big, state-backed fusion projects have delivered a genuine understanding of the physics of plasma containment, which private companies hope to leverage to rapidly innovate and produce better reactors.
CFS and others in the race to fusion are 鈥渂uilding on the shoulders鈥� of this new grasp of plasmas, says Mumgaard. In fact, a large part of the research that many public labs now do is running computer codes and doing calculations for private companies. 鈥淲e鈥檙e in that baton-passing phase,鈥� says Windridge. 鈥淧eople tend to be now viewing the main challenges in fusion as engineering challenges.鈥�
Which isn鈥檛 to say these engineering challenges will be easy to overcome. Here again, though, there are signs of progress. CFS isn鈥檛 alone in having raised investment off the back of demonstrations of technology 鈥� such as artificially intelligent plasma control, more powerful lasers or novel fuel sources 鈥� and new designs that push beyond the traditional tokamak doughnut.
in the UK and ENN in China, as well as the planned UK government-funded STEP reactor, will use a new type of spherical tokamak. This allows reactors to be even smaller because, for the same volume, spheres have a smaller surface area than doughnuts, so the magnetic fields can push more effectively. Like CFS, Tokamak Energy says it will have pilot reactors running in the mid-2020s, with commercial reactors following in the early 2030s.
Donn茅 says these are 鈥渧ery aggressive timelines鈥� because private companies don鈥檛 factor in the learning curves you have to negotiate every time you upgrade or redesign a fusion reactor. When JET refurbished its tokamak reactor in the late 1990s, swapping its carbon walls for more advanced metal ones, it took the researchers five years to relearn how to use the machine. Donn茅 reckons the promises of fusion in a decade are more attuned to the investors than the science.
Innovative fusion reactor designs
Building smaller tokamaks undoubtedly presents fresh challenges, says Cowley, but the payoff is that you can make cheaper machines more quickly. He points out that, decades ago, cars used to break down all the time 鈥� but those days are gone because most faults have been eliminated over many iterations. Perhaps fusion engineers can do likewise.
Other companies are pushing radical new designs that they hope will circumvent the problems that have held back fusion reactors. 鈥淲e deliberately picked our solutions to work around those engineering challenges,鈥� says Nick Hawker, CEO of Oxford-based company , whose novel approach to inertial fusion mimics the snapping claw of a pistol shrimp.
Yes, a shrimp. Pistol shrimps stun their prey by quickly closing their giant claw to create a focused shock wave in water, where temperatures rise briefly to extreme levels. For Hawker鈥檚 PhD, he simulated that process and later built it into the technology that FLF uses to create fusion reactions. Instead of focusing many lasers onto a capsule, as occurs at NIF, FLF fires a projectile at high speed and then concentrates the energy onto the fusion fuel capsule using its shrimp-inspired approach. The plan is that it will make inertial fusion cheaper and easy to replicate. 鈥淚s every fusion start-up going to succeed? No, probably not,鈥� says Hawker. 鈥淚s everything worth trying? Yes.鈥�
Many new companies have picked up blueprints discarded by public labs, adapting them and integrating new technology. Canadian company General Fusion, for instance, is betting on a cross between a tokamak and inertial confinement, using magnetic fields to levitate the plasma, followed by liquid metal pistons to compress it. This concept came out of the US Naval Research Laboratory in the 1970s, but without the electronics to synchronise their pistons precisely enough 鈥� and so symmetrically compress the plasma 鈥� it failed. 鈥淲ith the better technology that we have today, and the advances in the plasma physics, we can pull it off,鈥� says the firm鈥檚 founder Michel Laberge.
General Fusion鈥檚 design has the potential to leap one key hurdle in particular: making fusion fuel. A common fusion fuel is tritium, a form of hydrogen, which has to be produced in a separate process to fusion itself. This is a vital part of any operation, but most companies seem to have ignored it, says Donn茅: 鈥淚t鈥檚 like selling a car without having petrol stations around.鈥� But General Fusion is different. The fusion reaction is surrounded by a jacket of liquid metal that includes some lithium. This jacket absorbs the heat effectively 鈥� and at the same time some of the lithium is turned into tritium, producing new fission fuel.
It remains to be seen if solutions like this can work in practice. Even if they do, no-one has yet built the gubbins that would surround a reactor to make electricity non-stop. Exhaust systems, heat exchangers, turbines 鈥� all of it must be custom-built from materials that can withstand the high-energy neutrons released during fusion. 鈥淯nless these companies have a secret research line in materials, after a few months of operation [the reactors] will simply fall apart,鈥� says Donn茅.
So where does all this leave fusion? Clearly, there is a lot of engineering to do before the dream of bottling the power of stars becomes a reality. The private fusion companies argue that all it requires is proper investment, at a scale proportional to the size of the task and the potential payoff 鈥� and it is certainly true that fusion has never really received proper backing from governments. 鈥淚t doesn鈥檛 make it easy to keep the pace,鈥� says Donn茅.
But Cowley warns that we should be careful about comparisons with the space industry. 鈥淲e鈥檇 been to the moon before SpaceX was conceived, we鈥檇 launched hundreds of satellites,鈥� he says. 鈥淭here鈥檚 a lot of difference between that and where we are in fusion because we鈥檝e never made any electricity from fusion. At all.鈥�
