pbsspacetime
patreon
The Final Barrier to Nearly Infinite Energy (Script)
🕑 Added 2025-02-25 18:00:06 +0000 UTCThey say fusion is 50 years away, and they've been saying that for 50 years. But if so why are billions suddenly being pumped into fusion startups? Well to train LLMs, but there's a reason the technobrats are bullish on fusion in particular. The fact is, the technological challenges have been chipped away and in many cases solved over the past decades, and there's really no one deal-breaker difficulty remaining. One of the final challenges is deciding on the physical vessel to contain our mini artificial stars--and we have some pretty sci-fi options that are neally ready to go.
Solar power is the ultimate in clean energy—and it’s delicious when you can get it, which isn’t everywhere all the time. But solar power is just fusion power that’s happening far away. Slam hydrogen nuclei together into helium and a bit of the original mass is converted into energy. In the core of the Sun, that energy mostly radiates into empty space, with a miniscule fraction hitting the Earth and powering nearly all of life. A minisculer fraction hits our solar panels, and only in, well, sunny places and times.
There’s a limit to how far we can scale solar energy collection until we build our Dyson sphere. Why build a cage around a star if you can instead build a star inside a cage? What if we can run mini-Suns right here on the surface of the Earth? That’s the goal of the artificial fusion reactor. Today I want to talk about a specific aspect of fusion—and that’s how to confine hydrogen fusion in a way that leads to a net energy output. I want to talk about how to bottle a star. I’m going to reference ITER a lot—that’s the European Union’s fusion experiment—and I’ll refer to it because it’s the largest and arguably the most advanced, and is projected to achieve “first plasma” this year. I do want to emphasize that there’s a LOT of exciting stuff in recent years from smaller efforts—both national programs and increasingly private endeavors. We’ll come back to some of the more speculative advances another time.
Let’s start by taking whatever lessons we can from the Sun. Our home star sustains a steady fusion rate for billions of years because it confines its nuclear inferno by crushing it under the weight of 100,000 Earths. Single protons—hydrogen nuclei—are fused in a series of steps to produce helium-4 plus a lot of energy. The Sun’s proton-proton chain runs at something like 100 million Earth atmospheres of pressure and over 15 million Kelvin temperature, generating the equivalent of 20 million times the entire world’s nuclear arsenal in energy every second.
So the Sun both crams protons close together and gets them moving really fast in order to get them to stick. We need to do the same—we need to both compress and heat our nuclei. And the resulting fusion needs to output more energy than was used to reach that state. One big efficiency upgrade is to focus on the most energy-rich fusion reactions. Rather than going all the way to helium from single protons like the Sun does, we can fuse the heavy hydrogen isotopes deuterium and tritium. The Sun can achieve net energy gain starting with protons because of the incredible densities it reaches. Even with heavy isotope fusion, have to compensate further with a lot higher temperatures—
in fact something like 100 times the temperature of the core of the sun.
There is no material in the universe that wouldn’t be instantly vaporized at the temperatures needed for fusion. How are we supposed to even contain our fusion reaction, let alone crush it to the needed pressures?
Well, there are two broad approaches. Inertial confinement and magnetic confinement. Inertial confinement attempts to slam nuclei together with some sort of shock. In 2022 the National Ignition Facility achieved net energy output for the first time by forcing fuel together with lasers. But it’s not the first inertial confinement approach to succeed. The hydrogen bomb is a type of inertial confinement where the thing slamming nuclei together is an fission bomb—an old-school atomic explosion. The latter is rather uncontrolled, and so not suitable for sustained energy production. But even the controlled inertial confinement approaches have the issue of rather bursty energy output, and so are generally not thought to be the most likely option for our first-generation commercial fusion reactors.
Those are more likely to be magnetic confinement reactors. The hydrogen in a reactor is so hot that positively charged nuclei are stripped of their negatively charged electrons making a plasma of electrically charged particles that can be manipulated by magnetic fields. These fields are created by powerful electromagnets made from superconductors, and those superconductors have to be cooled to near absolute zero. So you have this center-of-a-star-hot plasma, mere meters away from matter that’s colder than intergalactic space. So to bottle our star we have to create the largest temperature gradient in the known universe—which I guess is something to be proud of all on its own.
Magnetic confinement reactors work by squeezing the plasma so it flows around a torus, and then compressing the torus with magnetic field and heating it by blasting in energy via particle beams until fusion is achieved. This has to be done cleverly because if you just run plasma in a circle leads, various instabilities that are fatal for your plasma, and perhaps for your reactor…
The leading solutions for stable toroidal plasma confinement are the tokamak and stellarator. The details of their differences aren’t too important here. Tokamaks are more widely used—for example by ITER, so I’ll cover this just briefly. Tokamaks use three systems of magnets. The first system is the Torodial system, which is the main containment system and gives the plasma the donut shape. The next system is the poloidal system which helps shape and position the plasma near the inner wall. This is to help compress the plasma for more reactions to occur. Lastly, the central solenoid drives the direction of the plasma, spinning the plasma in a circle. Tokamaks suppress instabilities by …
In contrast, stellarators combine all three magnetic systems into one using a 3D design, making them significantly more complicated.
To some extent, the magnetic field is a relatively solved problem. There are no doubt refinements to be made in the form of the field, and definitely in the generation of the field with new superconductor technology. That’s something we can come back to another time. Currently, one of the biggest engineering and research challenges in fusion is in how to build the actual physical containment chamber for that field and the plasma it holds. How do we build a wall that can hold a mini-Sun? The magnetic fields certainly helps by creating what’s called the ‘pedestal’ region, where the plasma is cooler and sparser, relative to the middle where the fusion reactions are happening, this helps create a small buffer between the fusion plasma and the wall. So the material doesn’t literally have to withstand contact with millions-Kelvin plasmas. But it still has to do a lot.
That wall is the first point of contact between the plasma and the outside world—and we call that plasma-facing surface the first-wall. While the magnetic field keeps the highest temperature parts of the plasma out of direct contact with the wall, a lot of energetic particles do reach it, and the wall needs to withstand that. It then needs to take the heat generated by this incoming radiation and transport it to somewhere useful to actually generate power. Finally, as if this wasn’t enough, the chamber wall needs to create new fuel for fusion. Let’s look at these one by one.
Within the coiled ribbon of millions-of-Kelvin plasma, heavy hydrogen nuclei collide and release energetic gamma rays and very fast-moving neutrons and helium nuclei. Those gamma rays and helium nuclei help keep the plasma hot. So hot that it radiates x-rays. This thermal radiation bakes the outer wall of the chamber, and the wall has to very quickly conduct this heat to the cooling system quickly or be at risk of melting. Even more energy is carried by the neutrons. These neutral particles are not affected by the magnetic field and so fly straight out of the plasma. They deposit kinetic energy into the wall, and this is the main source of heat energy from fusion. Some neutrons hit nuclei in the wall, potentially creating unstable isotopes. Some helium and hydrogen nuclei also end up with enough energy to escape the magnetic field and reach the first wall. Some of these are tritium nuclei, which are also highly unstable. As a result of the neutron and tritium bombardment the wall becomes more radioactive over time.
These impacts also erode the wall, with the wall’s own nuclei getting smacked from the surface into the plasma chamber in an effect called sputtering. This degrades the shielding over time, and also contaminates the plasma.
Finally, an imperfect wall structure can lead to runaway instabilities in the plasma called “edge localization modes”. These can result in localized failure of the magnetic confinement, and to large amounts of energy being dumped into the wall, which is not good for the wall. This can also trigger larger magnetohydrodynamic instabilities in the plasma, causing the plasma to shed a lot of energy and potentially kill the fusion reaction.
Assuming our wall can survive the above, the heat generated within the wall will be carried away by some “working fluid”--water is a popular choice, but also molten salts and even lithium.
The working fluid both cools the wall and heats the fluid to the point that it can turn a generator.
We’ve just generated electricity. Nice. But even if we got more power out than we used to create our plasma, this process is not sustainable. The deuterium part of the fuel is enormously abundant—there are 33 grams per cubic meter of seawater, which means basically limitless. But the tritium is not abundant—it has a half-life of 12 years and so is almost non-existent naturally. The reactor has to make its own tritium. That’s not super complicated. Hit a lithium nucleus with a neutron and it’ll split into two tritium nuclei. So the trick is to place a layer of lithium behind the first wall. That wall will slow neutrons from the fusion reaction, extracting some of their energy, and the neutrons will then hit the lithium and breed tritium. Obviously the lithium has to be replaced, but fortunately lithium is moderately abundant—with thousands of years worth in the ground and millions of years worth in the ocean.
One major issue with this is that the fusion reaction doesn’t produce enough neutrons to breed a sustainable amount of tritium. To deal with this, a neutron multiplier is needed—this is a layer of material that amplifies the neutron flux—one neutron goes in and two come out to feed the tritium breeder. That way each fusion reaction replaces the tritium it uses.
So now that we know all the miracles that the reactor wall has to pull off, let’s look at the options for materials.
The most traditional pick for a fusion reactor first wall is the metal tungsten. It’s metal, so it’s structurally strong. And it has the highest melting temperature of all metals, so can stay solid under heavier bombardment. It’s also less inclined to hold onto as many of the radioactive tritium as other materials. And, very importantly, it has a low sputtering rate—fewer tungsten atoms ejected into the plasma.
However, it also has a major weakness with regards sputtering—in that those tungsten atoms that do join the plasma have a pretty unfortunate effect. Tungsten has 74 protons in the nucleus compared to hydrogen’s one proton. That means tungsten’s electrons are much more strongly bound. So while the plasma’s hydrogen is fully ionized—liberated of electrons— the tungsten atom will retain 20 or 30 electrons, even inside the plasma. So now when a tungsten atom bumps into another particle in the plasma, some of the energy of the collision goes into exciting one of those electron. That electron then drops back to a de-excited state, emitting a photon in the process. And that photon then escapes the plasma, taking that energy with it.
The larger the atomic number of the atom, the more electrons it can retain in the plasma, and also the smaller the energy gap between electron levels. That means more opportunities ot absorb energy from the plasma and radiate it away.
This effect is called line emission cooling, and it’s a great way to lose a lot of energy from your plasma very quickly. Even a little bit of a heavy element like tungsten in the plasma can make it very difficult to keep it hot enough to sustain fusion.
Tungsten has major advantages, but the issue with pollution of the plasma was enough to convince the ITER team to try something different. At least to start with. In fact, they planned to go in the opposite direction to tungsten in terms of location on the periodic table by using beryllium for the plasma-facing layer. Beryllium is the 4th element on the periodic table, so even though it will end up polluting the plasma, those contaminants tend to be completely ionized. There are very few electrons in atomic shells to radiate away the plasma’s heat. In fact, beryllium actually removes pollutants. It can capture oxygen impurities, which themselves are released from oxidized metals in the vessel wall when the machine is running. Oxygen in the plasma can cause the plasma to cool just like tungsten.
Beryllium has other major benefits. It has an extremely high thermal conductivity, allowing it to quickly transfer heat from the front plasma-facing surface to the back surface with the cooling systems. And one of the most compelling benefits of Beryllium is that it acts as a neutron multiplier—doubling the neutrons available for tritium breeding.
So what’s the downside? Beryllium does have a much higher sputtering rate than tungsten—and even though the ejected impurities don’t harm the plasma nearly as much, this does mean the beryllium surface erodes much more quickly than tungsten. That means the wall needs to be replaced more often. Also, fluctuations in the plasma will induce electrical currents in a beryllium wall, and those currents experience massive forces from the strong magnetic field. This can damage the wall and even cause it to fail while the reactor is running.
A huge disadvantage of beryllium is that it’s extremely toxic. Erosion of the wall fills the reactor chamber with beryllium dust, which cannot under any circumstances be allowed to get out of the reactor if you want to keep your nuclear engineers "operational." Finally, beryllium is super rare. ITER would need 12 tonnes of the stuff for its first wall.
ITER had originally planned a beryllium first wall, but just in 2023 decided that the hassles outweighed the benefits and shifted their plan to good ol’ tungsten. This is a pretty safe choice for ITER as an experimental facility, but for commercial fusion some kind of neutron multiplier feeding the tritium breeder is needed. ITER is currently testing various options.
So we still have the problem of tungsten polluting the plasma. There are other options for the plasma-facing surface though. Boron is currently getting the most attention from the ITER crew, since it’s the next lightest element after beryllium. The idea is to build a tungsten wall, and then sprinkle a little bit of boron powder into the reactor while its running. This should then evaporate and coat the wall, reducing the amount of tungsten that cooks off into the plasma. Boron does the icky side effect of retaining tritium, which means the inner wall could become dangerously radioactive over time.
Meanwhile, going the other direction on the periodic table from beryllium we get lithium, which if you remember we were using to breed tritium behind the first wall. At first glance lithium seems a non-starter for the plasma facing wall because lithium metal has a melting point of 182 degrees Celsius. But who says the first wall really has to be solid. There’s this idea to coat the reactor chamber in a layer of liquid lithium, which has many advantages. You can’t really damage a liquid structurally—Seriously, go hit a lake with a hammer and see what happens. Lithium nuclei would end up in the plasma instead of the tungsten behind it. But this is actually good for the plasma. Getting plasmas up to temperature requires physicists to add large amounts of energy, both with EM fields and with “neutral beam injectors”. Experiments like the Lithium Tokamak Experiment-Beta (LTX-β) at the Princeton Plasma Physics Lab suggest that a little bit of lithium in the plasma may help the plasma get up to temperature more easily. Finally, the lithium can even double as the coolant and the tritium breeder if it is moved through and behind the tungsten wall.
As I said, ITER is supposed to produce its first plasma this year, but is projected to achieve its first fusion reaction in 2035, smashing two deuteriums together to
produce helium-3. Now its first commercial grade deuterium-tritium fusion is slated for 2039. If all that works then I guess we build a lot of them. But it's also possible that one of the many smaller enterprises will beat ITER--they certainly talk a big game, with some projecting first-fusion in the next few years. But that's a story for another episode. Personally I don't care who wins--I just want my infinite free energy so I can train a godlike LLM to take over writing spacetime.
