The Star Builders

by Arthur Turrell

  I. But not Fukushima, because it happened after the study was published.

  CHAPTER 9 FINISHING THE RACE FOR FUSION

  “The day when the scientist, no matter how devoted, may make significant progress alone and without material help is past… Instead of an attic with a few test tubes, bits of wire and odds and ends, the attack on the atomic nucleus has required the development and construction of great instruments on an engineering scale.”

  —Ernest Lawrence, Nobel Prize banquet speech, 19401

  “Fusion has always been thirty years away and that gets thrown at me at every public outreach talk I do,” Culham’s Professor Ian Chapman says when I ask him if star builders are closer to delivering net energy gain than they were thirty years ago. “It ignores the huge progress that’s been made.”

  While Ian talks candidly about mistakes, he believes that we’re closer than ever to building a star. The data back him up: between 1957 and 2018, the joint combination of temperature, density, and confinement achieved by any star machine increased by a factor of 1 million.

  JET deserves credit for having pushed progress the furthest with its brief generation of 16 megawatts and a Q of 67 percent in 1997—tantalizingly close to the finish line. Most machines have only managed fractions of a percent.

  Although JET’s record hasn’t been broken, progress on parts of the problem have continued apace. The very first devices controlled plasmas for microseconds. In the late 1950s, John Cockcroft’s ZETA machine improved that one thousand times over. JET, built in 1984, runs for a few seconds—another thousandfold improvement. But controlling fusion plasmas for minutes is now possible. In 2003, the Tore Supra tokamak in France managed to control plasma for six minutes and thirty seconds. Also on the accomplishment list: many tokamaks are now able to reach temperatures beyond a hundred million degrees. The EAST tokamak, in China, has managed to combine high temperatures, as high as fifty million degrees, with times of one hundred seconds. Both EAST and France’s WEST tokamak are aiming to control their plasmas for longer than fifteen minutes. Records keep being broken: in 2020, the Korea Superconducting Tokamak Advanced Research (KSTAR) managed twenty seconds at more than a hundred million degrees.2

  The progress on which JET relies, and future tokamaks will build, has largely been won by ever bigger machines, ever bigger teams, and some serendipity. “The present status of fusion is primarily the outcome of a large effort and a lot of dedicated work, and only at a few occasions did we obtain results which were so pleasantly surprising that the community even took time to accept them,” Professor Sibylle Günter, scientific director of the Max Planck Institute for Plasma Physics, told me.

  Sibylle’s list of the breakthroughs begins with the Russian T-3 tokamak that appeared on the international scene in 1969 and completely changed the direction of magnetic fusion research. She also talks about a discovery at her own institution in 1982. A researcher named Fritz Wagner working on the ASDEX tokamak in Garching found machine settings that almost magically caused both the confinement time and plasma density to double. Without this single discovery, tokamaks would probably have to be twice as big.3

  Other progress has come from theory and computation. It’s much easier and cheaper to do an experiment on a computer than in a $2.6 billion star machine, if the simulations are realistic. As both computer hardware and software have improved, so simulations have become better guides to reality. Researchers at JET recently used machine learning to predict disruptions tens of milliseconds before they occurred, giving them time to tweak the plasma to dissipate the buildup of energy and forces. Happily, the insights garnered can be applied to other tokamaks. Such breakthroughs are a step toward more reliable reactors.4

  Ian Chapman thinks these improvements could give the decades-old JET a shot at setting a new world record for fusion energy when it begins a new run of deuterium-tritium experiments. “When we last ran JET with deuterium and tritium in 1997 it was a point in time,” he says. “The power raced up and we lost control of the fuel, and it was a millisecond. We’re going to do that again, but this time we’ll race it up and hold it there, and bring it down in a controlled way… We now understand plasmas better than we did twenty years ago.”5

  Magnetic fusion machines are edging closer and closer to the finish line. If magnetic fusion star builders can get the combined temperature, density, and confinement times just another 1.8 times higher, they’ll hit ignition. That’s when their plasmas not only get net energy gain, but generate enough energy from fusion to remove the need for external heating.6

  Far out in the lead in achieving the trio of temperature, density, and confinement simultaneously are three state-sponsored tokamaks: Japan’s JT-60, South Korea’s KSTAR, and the Joint European Torus (JET). The stellarator Wendelstein 7-X is not far behind them, achieving impressive feats for a relatively new machine based on a different technology.7

  For now, it looks as though the private sector fusion firms are lagging behind—though some are keeping the details of their progress secret. At the back of the pack are those firms aiming to do fusion reactions that aren’t deuterium-tritium: they have a very, very long way to go because the conditions required are more extreme. On the other hand, many private sector star builders have been playing an impressive game of catch-up. The real question for them is whether they’ll be able to continue the rapid rate of progress and leapfrog ahead.8

  The huge progress made by magnetic confinement fusion machines means that net energy gain is within reach. And even if JET can’t make it all the way, another machine is coming that almost certainly will.

  We’re Going to Need a Bigger Tokamak

  There’s one aspect of the current fleet of magnetic fusion machines that is holding back progress. It’s a lesson that has been learned time and time again in fusion, a lesson that the Big Bang, stars, supernovae, and nuclear weapons have been telling us all along: fusion works best on big scales. For conventional tokamaks, the confinement of plasma gets better the bigger the machine is. Going from JET, at a three-meter (approximately ten-foot) radius, to a tokamak twice that size would improve confinement four times over.9

  In 2006, recognizing the need for a machine that can run with deuterium and tritium and has the necessary scale to go beyond net energy gain, a group of thirty-five countries got together and, in a ceremony at the Elysée Palace that was attended by four hundred guests, signed an agreement to build an almighty tokamak in the South of France, in Cadarache, near Marseille. The members of this thermonuclear club are China, the European Union, India, Japan, Korea, Russia, and the United States, with the brunt of the costs borne by the EU. These nations represent more than half the world’s population. It took them twenty-two years of planning, arguing, design, and preparation to agree to build the new machine: ITER.

  When it is completed, ITER will be the world’s largest tokamak, and one of its key objectives will be to demonstrate net energy gain. It’s a behemoth. Much of it is designed on principles that have been tested on JET, but it will operate with more than ten times the volume of plasma—and bigger plasma volume usually means more stability. Instead of having a three-meter-radius torus, ITER will have one that is more than six meters.

  ITER’s engineering is more complex than that of any other machine yet built. The tokamak’s superconducting magnets will need to be formed from one hundred thousand kilometers (approximately sixty-two thousand miles) of niobium-tin wire. Each finished magnetic coil will stand seventeen meters (about fifty-six feet) tall and nine (about thirty feet) wide and be cooled to four degrees above absolute zero. Eighteen of these will be arrayed around the tokamak to generate a magnetic field of thirteen Tesla (3 million times the Earth’s magnetic field), storing up tens of thousands of megajoules of magnetic energy. ITER will take up 180 hectares (equivalent to 250 soccer fields), and when finished, its structure will have a mass equivalent to three Eiffel Towers.

  While ITER is situated in the South of France, the work being undertaken fo
r it is truly international. There are more than fifteen satellite sites around the world providing expertise and parts. The project’s current director general, Dr. Bernard Bigot, has described it as an “extraordinary human adventure” and he’s not just talking about the three thousand or so staff who will be needed to run ITER; this is a fusion experiment on behalf of the planet.10

  As with many large and complex international collaborations where juicy contracts are available, construction has been beset by delays. The machine was initially due to turn on in 2016, but that target date has slipped by, and the first plasma isn’t scheduled to appear until 2025, with deuterium-tritium experiments expected to begin even later, in 2035. As of the start of 2020, construction was two-thirds complete.11

  ITER has been designed to win the race to build a star. It’s aiming to hit a plasma Q of five—that is, five times as much power out as in. It will also try to reach a Q of ten for a few hundred seconds, generating five hundred megawatts of power for the fifty megawatts of power put in. The plasma will reach temperatures of 150 million degrees Celsius (270 million degrees Fahrenheit). The gain in energy won’t be enough for a commercial reactor, and ITER won’t deliver a single watt of power to the grid, but it’s enough to show that fusion power is possible. If it works, ITER will shift fusion from fantasy to tangible power source.12

  We’re Going to Need a Bigger Laser

  Inertial confinement fusion has also seen huge improvements, with NIF by far the leading machine. Many had hoped NIF would reach ignition when it was completed in 2010. The laser performed well. The experiments on nuclear stockpile stewardship that dominate NIF’s laser shots are said to have been a success. Scientific experiments probing conditions in planets and supernovae have been a hit. But on the inertial fusion energy program, progress was initially much less forthcoming. Through 2011 and 2012, the implosions were riddled with instabilities. While neutrons were generated and energy released, the results were orders of magnitude less than what had been expected from parsimonious mathematical models and simple computer simulations.13

  Following sharp criticism from Congress and the US Department of Energy, NIF’s scientists changed direction and started trying to understand what was going wrong in the implosions, and why they weren’t behaving like the simulations suggested. New staff—including new director Dr. Mark Herrmann and new chief scientist Dr. Omar Hurricane—were brought in and the program was refreshed. With new ideas, NIF found its stride. The experiments are now breaking records for inertial fusion, and they’re not a million miles behind JET’s world-best record for energy yield.

  Between 2011 and 2019, the fusion energy released in the best shot on NIF rose sixty times over. In 2018, NIF first reached an energy yield of 3 percent. In 2020, NIF’s scientists announced that they had exceeded this. This is too much fusion to be occurring just due to the energy coming from the laser. Making the analogy with fire, it’s robust evidence that heat isn’t just coming from the match. After a rocky few years, NIF is back in the race.14

  Just as with the star builders using magnetic confinement, computers have played an important role in NIF’s success. In 2018, Livermore’s supercomputer, Sierra, became the second fastest in the world. Better computing means that it’s possible to simulate more of the complex ways that plasmas can misbehave, or use machine learning to guide experiments.15

  But the most significant progress has come from better control of how the laser and the plasma interact in the hohlraum, the gold box that houses the capsule of fusion fuel. All of those weird plasma waves and instabilities can transfer energy from one beam to another, or turn the plasma in the hohlraum into a mirror. By being careful about how energy is distributed across the 192 beams, NIF’s scientists have managed to tame some of the laser-plasma interactions. Through experimentation, they discovered that reducing the density of the gas between the hohlraum and capsule a little, but not too much, allowed more of the laser beam energy to get into the gold. That meant more energy getting into the capsule, and more strongly driven implosions. Capsules are now regularly imploding at almost 400 km/s, more than thirty times Earth’s escape velocity (even faster than NIF scientist and astronaut Dr. Jeff Wisoff traveled during space shuttle launches).

  If NIF achieves ignition, it will be a sudden jump from brief spark to burning fusion plasma. As Mark Herrmann said, “In the simulations, it really is a cliff edge.” While the gap from 3 percent to 100 percent may seem large, the improvements in energy yield that NIF has seen since 2011 have come in factors of between five or six. NIF’s scientists aren’t very many improvements away from net energy gain.

  Omar Hurricane told me that plasma conditions were a much better measure than yield of energy alone of how close experiments are to ignition: “The metrics to focus on are actually the pressure, confinement time, or temperature, or you can also look at it as density times radius [of the hotspot within the fuel capsule]. The yield itself doesn’t give a good impression of how close you are, as it’s like approaching a cliff.”

  Different inertial fusion experiments are harder to compare with each other than magnetic fusion ones. The MagLIF experiment at Sandia National Laboratory has generated substantial numbers of neutrons using deuterium-deuterium reactions, but it isn’t running with tritium.16 Excitingly, First Light Fusion just started running with deuterium and tritium in 2020—but it hasn’t yet published results on yield or the conditions it’s reaching. All of which means NIF still tops the scoreboard for energy gain from inertial fusion.

  While NIF has made great strides, to get that bit further requires more energy getting into the capsule’s hotspot and kick-starting the fusion reactions. Each doubling of the energy dumped into the hotspot gives, roughly, ten times as much energy out. From the secret Halite-Centurion experiments, inertial star builders know that it is physically possible to get net energy gain if only they can build a laser big enough. Those experiments suggest that slamming a fusion fuel capsule with five to ten megajoules will result in net energy gain. But NIF is only fielding 1.8 megajoules right now.

  So scientists at NIF are thinking about squeezing a few more joules out of their already phenomenally big laser. As Jeff Wisoff told me, it’s expensive to add energy to the laser, so they have to try everything else first—but he said that a “modest upgrade” might be possible. Dr. Bruno Van Wonterghem, who spent years designing NIF and who took me on a tour of its steel arteries, told me about the improvements, reeling off what they’d done already, “but the ultimate is the bigger hammer”—that is, making the laser even bigger. Bruno smiled just a little when he told me that NIF had capacity for a 50 percent increase in laser energy. When I asked about the cost of a big laser upgrade, Livermore said that it would be a tiny fraction of the capital cost of NIF and Bruno gave the impression that it would be substantially less than their operating budget. If their sponsors at the Department of Energy go for it, he thinks they could do the work in a year or so. That might just be enough to get them to ignition.17

  “When you look around the world,” Jeff Wisoff tells me, “and say, ‘Where is it credible we might get ignition in the next ten years?’ I think NIF is the best shot.” Unsurprisingly, NIF’s director, Mark Herrmann, agrees, saying of ignition: “It needs a certain scale, energy, pressure; this is the only facility that has the prospect of doing that in the next decade.”

  The Race for Net Energy Gain Is Still Open

  With machines like JET and NIF ever so close but each limited in some way, and ITER not even attempting net energy gain until 2035, there’s a chance that another machine could swoop in and win the race.18 Mark Herrmann told me that he spends a lot of time making sure NIF doesn’t fall behind other countries’ efforts. The international competition to achieve net energy gain looks set to become fiercer. Several other countries are planning, or have even built, laser fusion machines. They can learn from NIF—in particular, they know that because NIF has reached 3 percent, they probably don’t need a laser that much bigger t
han 1.8 megajoules to hit 100 percent of net energy gain.

  France’s Laser MegaJoule was completed in 2014. I went to visit it in 2011, when it was under construction. As at NIF, a catacomb of thick, radiation-blocking walls had been constructed around the spherical target chamber. Because it hadn’t yet been irradiated by the first experiments, I was able to poke my head into the shiny, echoey chamber and marvel at how it had been engineered to submillimeter precision and yet still was ready to withstand bombardment by millions upon millions of high-energy particles. Right now, Laser MegaJoule is fielding slightly less energy than NIF and it only works in the less-proven direct drive mode (which doesn’t use a hohlraum), but it’s one to watch.

  Russian scientists revolutionized magnetic fusion when they revealed the tokamak. Now Russia has a plan to get out ahead on laser fusion. They’re building a 192-beam laser facility, called UFL-2M. It will have 2.8 megajoules of energy, albeit in a less useful color than NIF’s infrared beams. Russia also has a dark horse in the race: a fission-fusion hybrid reactor. Some star builders don’t like fission-fusion hybrids because they lack many of the benefits of fusion alone. But some have passionately supported them as a more certain stepping stone that would bring net energy gain more easily within reach, and would allow for quicker development of fusion technologies.19

  China is perhaps the most ambitious of all when it comes to fusion. “If you’re prepared to do two or three different concepts and take a bit more risk, you can go faster,” Ian Chapman told me. “That’s what China is saying.” If China follows through on its plans, it will be at the forefront of both tokamaks and laser fusion within the next decade, leaving Europe and the USA behind. Not only does it already have an existing 0.2-megajoule laser, Shenguang III, and EAST, a tokamak that reached temperatures of 100 million degrees Celsius (180 million degrees Fahrenheit) in 2018, it also has plans to build ignition scale devices that will rival or even surpass NIF and ITER respectively. “It’s ambitious but not inconceivable,” Ian said. “I think for China it’s realistic because they’re prepared to put the resources behind it.”