Tag Archives: physics

Fusion power: The tritium problem

In my previous post on fusion power, I discussed how fusion power will not be possible on Earth without using tritium. I then discussed how Tritium is exceptionally *rare* on Earth (the entire planet has only enough Tritium to power Chicago for a month or so). Now I will discuss why it won’t be possible to make enough Tritium to solve this pressing issue with fusion power. Thus I hope to answer the question posed in my first post: fusion power is hopeless (within my lifetime at least).

The 25kg of Tritium produced by Canada’s CANDU fission reactors represents the sum total of all the tritium here on Earth. If fusion advocates want to have enough Tritium to start up loads of fusion power plants, they’ll need a lot more than that.

The first method may be to just produce more tritium the way Canada does. The CANDU reactors use heavy water (water with the hydrogen-1 atoms replaced with hydrogen-2 aka deuterium) to moderate free neutrons and cool the system. When the deuterium in the water interacts with a neutron, it may capture it and transform into tritium (aka hydrogen-3). This tritium is highly radioactive with a half-life of just 12ish years, so Canada separates it out and stores it in shielded containers.

To get a tiny bit more technical: a single atom of uranium-235 undergoes fission and releases either 2 or 3 neutrons (average 2.5). 1 neutron is needed to keep the fission reaction going, 1 neutron can hit deuterium and transform it into tritium, and we can estimate that the remaining 0.5 neutrons will be lost due to hitting some other atom in the area that isn’t deuterium or uranium.

But now we have a process to create tritium, so can’t we just scale it up? Unfortunately no, this is unfeasible. Remember that it takes 1 atom of uranium (which weighs 235 atomic units) to create 1 unit of tritium (which weighs 3 atomic units). 235 ÷ 3 is about 78, so it takes 78 kilograms of ultra-enriched uranium to produce a single kilogram of tritium. The CANDU reactors themselves don’t even use highly enriched uranium, their uranium contains only about 1% U-235, so for them it takes 780 kg of uranium for each 1 kg of tritium they produce.

Even if we could ensure that extra 0.5 neutrons wasn’t lost, that doesn’t get us very far. And to be honest, even this math undersells the problem. That 780 kg of uranium to produce 1 kg of tritium *only works if we extract every single possible neutron from the fissile uranium*. In practice, the nature of half-lives means that the uranium produces most of its neutrons quite early and then neutron production quickly decreases, to the point that it’s producing too few neutrons to reliably continue a chain reaction. The actual CANDU reactors replace their uranium regularly to maintain a high enough level of neutrons, and they use about 100,000 kg of uranium to produce about 0.1 kg of tritium each year. That is just too much uranium (all of which needs significant work to keep it safe and secure) to feasibly scale this process up to produce lots and lots of tritium.

Sure there are other neutron source, and other targets for producing tritium. But the math is still the same. You need an extraordinary amount of some very heavy element (the kind that undergoes fission) to create a tiny amount of a very light element (tritium). Those heavy, fissionable elements are very dangerous and pose potential national security threats, so must be treated with high levels of security. The tritium is *also* dangerous, but has the added problem of requiring extra steps to separate it from the heavy-water mix that it’s a part of. Producing tritium with fission is a dead end if we want enough of it for fusion.

But of course no one wants to produce tritium using fission, they want to produce tritium using *fusion*, and here’s where our problem goes from bad to worse.

Most tritium-fusion advocates have a clear picture in their mind of how it will work:

  • Tritium and deuterium will fuse in a reactor, and this fusion will produce neutrons
  • The reactor will be surrounded by a blanket of (usually) lithium
  • The lithium will absorb a neutron and undergo fission, creating a new unit of tritium
  • That tritium will be fed back into the fusion process, continuing the cycle indefinitely

It all seems to simple, but every single step of this has problems, and no one has yet even demonstrated a coherent *plan*, let alone demonstrated a practical *system* for creating and extracting tritium in this way. It’s akin to saying “fusion is easy, just get the atoms close together” without having any plan for how to do that.

First, the high energy neutron created during fusion carries 80% of the reactions heat energy. That means it can’t just be used to create new tritium, if we want this reactor to actually be a *power plant*, we need that neutron to also boil water and drive a steam turbine. That means 1 neutron per fusion reaction isn’t enough.

So we add elements to the lithium blanket that “breed” more neutrons. These elements, when hit by a neutron, create 2 or more neutrons in turn. Great, now we can turn 1 neutron into as many as we want, with some of those neutrons being allowed to produce heat for our steam turbine, and the others creating tritium to go back into the reaction. Mission solved, right?

But this creates our first fundamental problem with the blanket. We have a blanket of solid lithium, plus beryllium and/or lead to act as neutron multipliers, and that blanket is constantly being destroyed by neutrons *as a necessary step to produce more tritium*.

This blanket isn’t *just* there for making neutrons mind you, it will also need to carry away the heat energy from those neutrons to be used to boil water. That means it will need all the steel piping, diagnostic sensors, and other components necessary to safely transfer heat away. That means the blanket needs a lot of non-lithium components in it, and those components *can also be destroyed by the fast neutrons*. Then again, as the lithium in the blanket is already expected to be destroyed by the fast neutrons, that can compromise the precise structural system needed for heat to be transferred and for diagnostic sensors to get an accurate reading. A single microscopic crack in the blanket could through the entire system into chaos, and a crack is inevitable when the whole point of the blanket is to be destroyed.

Not to mention the helium problem. A neutron can pass through a few atoms of lithium before it hits one to create tritium. That tritium can get locked within the structure of the wall and beta-decay into helium-3 before it can be extracted. Helium-3 can build up inside the walls of the reactor, and it aggregates together since it can’t bind to the steel or other elements that make up the wall. These cavities of helium-3 reduce the structural integrity of the whole system, as you suddenly have a section of wall sitting on top of gaseous helium instead of solid steel. This again is catastrophic to structural integrity, and it’s hard to make helium obey you at the best of times, even worse when it’s trapped inside a metal lattice. So it won’t be so easy to just remove the helium from the system.

But even if you magically create a wall that won’t be destroyed through neutron damage, and can breed enough new tritium to continue the reaction, there is no working proposal to extract that tritium and feed it back into the reactor. If the lithium blanket is a solid wall, then new tritium will be trapped inside that blanket, it cannot simply flow out straight through a solid wall. You might try to inject helium purge gas, to purge out all the trapped tritium, but now you have a different problem.

Those high energy neutrons are dumping enormous amounts of heat into this wall as well, there will be violent thermocycling as the walls heat with neutron radiation and are then desperately cooled back down by heat exchange (this heat is needed, remember, to boil the water for the steam turbine). The areas of the wall containing lithium will be crushed or sintered together, blocking any helium from removing the trapped tritium within.

Finally, a solid wall of steel and lithium with have thermal conductivity that is unpredictable and well-nigh impossible to model. The elements will touch as microscopic points, and those points will again move around as the wall expands and contracts from thermal forces. So you won’t have a wall that heats evenly and can be cooled evenly. Some patches will hit thousands of degrees and can warp or melt the wall if they aren’t cooled down fast enough.

So then maybe a solid wall won’t work for us, but how about a liquid wall? Lithium and our neutron multiplier atoms can be heated to hundreds of degrees until they liquify. Then we can surely have a homogenous liquid that heats and cools evenly, where helium cannot get trapped in solid lattices, and we can pass this liquid over chemical filters to extract all of the precious tritium from within. Simple right?

But this is even less probable of a solution. A river of molten lithium is highly conductive, and as it needs to be pumped through and around the high powered magnetic fields of the fusion reactor, it will trigger an unbelievable electromagnetic braking force. Remember the “braking radiation” I discussed in the previous post? We aren’t done with electromagnetics ruining our nuclear party. The conductive liquid metal passing through the magnetic fields will create a massive *secondary* magnetic field that acts to oppose the moving conductive fluid that created it. So our river of molten lithium will suddenly create a powerful force pushing it in the *opposite direction* that we need it to go in, forcing an ungodly amount of pumping power to be spent just to force it to move the correct way.

Then there’s the fact that a river of molten lithium will be highly corrosive, trying to tear down whatever piping infrastructure we’re using to pump it.

And finally, we still aren’t done with tritium problems. Yes it’s easier to remove tritium from a river of molten lithium than a lattice of solid lithium, but we still have a problem. Tritium diffuses through solid metals at high temperatures, meaning it will quickly leak out of our molten lithium, through the plumbing, and into any water we’re using as a coolant for this whole nightmare of a system (remember, we still NEED water coolant to drive our steam turbine if this thing will make any power!). This doesn’t just steal the tritium we need to continue the reaction, it produces tritiated water, highly radioactive and a severe biological safety hazard.

So no, we can’t *just make more tritium* to run our fusion power plants. We can’t just scale up our systems for creating tritium from fission. And any proposal to make tritium from the fusion itself is still an entirely separate unsolved problem, *on top of the unsolved problem of making fusion power in the first place*. The tritium created will try its level best to destroy whatever wall or plumbing infrastructure we use to create it. It will decay into helium-3, which will try even harder to do the same. And the wall itself is an unsolved problem of heat exchange, hydrodynamic forces, gas exchange, and structural integrity, all of which has to be *perfect* so that enough tritium is secured to continue the reaction.

And sure, you can lose a few atoms of tritium here and there, but any lost tritium will have to be repaid by redirecting more of the fusion reactor’s neutrons towards tritium and using fewer neutrons to create actually electrical power. You know, *the actual point of a nuclear reactor*. Even getting a positive production of tritium is an untested problem, *without* trying to create any nuclear power whatsoever.

There are some things that work great in a lab bench but just don’t work when scaled up. This is exactly the problem that sank Amyris, a biotech company I blogged about long ago. They had a good foundation of synthetic biology, and they thought they could easily scale up from a benchtop setting that made micrograms of product to a factory setting making the kilograms needed for industrial use. They just couldn’t do it, and they filed for chapter 11 bankruptcy in 2023. It’s not that what they wanted to do is impossible, it’s that it is difficult, expensive, and there are cheaper options available.

That, I think, is what will ultimately sink nuclear fusion for the remainder of my lifetime, not that it’s impossible, that it just isn’t worth it. Any problem humanity faces might be theoretically possible, we could redirect the entire National Science Foundation funding to making better AIs for Civ VI, for example, and we’d certainly get something better than what Firaxis gave us. But is that value for money, compared to everything else the NSF funds? No.

And so, is fusion power possible? If you feed the entire budget of the NSF into it, maybe. But even if we can heat our homes by burning money, that doesn’t make it a good investment. The NSF and other agencies want to fund areas of research that will create their own positive return on investment, that will enter a virtuous cycle where profits from the technology get reinvested into improving the technology further.

The NSF’s investment into genetic engineering did exactly that. In the 1970s, at the same time fusion was being funded and nuclear fission wasn’t yet politically toxic, the NSF was giving tiny amounts of funding to biologists and chemists to study genetics. To find out how we can read and use genes to our advantage. But unlike fusion power, genetics entered a *virtuous cycle*, where their projects were so successful they could start profitable companies with them, and those profitable companies then reinvested back into genetics technology. The human genome project was completed around the year 2000, and in 1970 knowing the entire genome of an individual would have seemed like *more* of a pipedream then getting power out of nuclear fusion (we already had nuclear fission, remember!).

But fusion never entered a virtuous cycle. There were never fusion successes that could go on and make profits to reinvest. And *that* is why fusion was abandoned, not because of some evil conspiracy, but because of simple failure.

Failure which I believe will continue. Because still today, investing in fusion just isn’t worth it. We can get power in a dozen of other ways, so long as we don’t have NIMBYs blocking every solar array or gas turbine needed to power our modern society. And don’t forget that fusion will have to face those same NIMBYs, you can’t compare a theoretical fusion plant to a currently built gas or solar plant. You have to include the NIMBY factor of convincing the same people who think 5G is giving them cancer that fusion power *won’t* give them cancer. Or the same people who don’t want a new apartment because it’s *ugly* that a fusion plant *won’t* be ugly.

The technological limitations of nuclear fusion power are still unsolved. They are solvable but solving them isn’t worth the amount of money needed considering the extreme amount of known unknown and even unknown unknowns. We don’t know how to do it, and we’re better off spending our money on things we do know.

Sure, let the NSF keep funding the national ignition lab, you can even keep funding ITER if you think it will help. But startups like Commonwealth Energy aren’t going to provide power to the grid within my lifetime, and it isn’t worth spending our entire GDP to solve these problems sooner than that.

Stars with a Story

I wanted to tell a short story about a star I found interesting. It’s nothing like I’ve blogged about before, but I haven’t blogged in a long time and I need to write SOMETHING.

At the heart of the Crab Nebula is a little star that I will call Crabby. Its real name is PSR B0531+21, but no one’s going to remember that name so Crabby it is. Crabby is a star that exploded in 1054 AD, and the Crab Nebula is the remnants of that explosion.

What’s so special about Crabby is that we know *exactly* when it exploded, because Chinese astronomers recorded it. For almost every other supernova remnant in the galaxy we don’t have that kind of information, it’s hard to tell whether something exploded a million years ago or a *billion* years ago because we’re viewing it from so far away. So knowing the exact *year* that Crabby exploded is like knowing someone’s time of birth down to the millisecond. It gives us incredible precision that we normally don’t get in astronomy.

And that precision has let scientists come up with some wacky theories, such as the theory that Crabby is a “strange star.”

Now what’s a strange star exactly? Isn’t it a little rude to call a star “strange” in the first place? Well it requires a little bit of quantum mechanics to explain this, but read on because I promise it’s worth the read.

See, when a star normally collapses, the force of gravity becomes so great that it compactifies the matter within to a ridiculous extent. The core of a collapsing star is usually made of heavy elements like iron, and the collapse forces those atoms as close together as physically possible. There’s a simple rule in quantum mechanics: two things can’t occupy the same exact space, and this rule leads to an absolute limit of density: the density where atoms can’t get any closer together or they’d literally overlap.

Here on earth atoms aren’t usually that close together. Liquids and gases of course let atoms slosh about, but even solids on earth usually have their atoms arranged in intricate patterns with a little bit of space in between each atom. White dwarfs crush this down into the absolute limit of density, around a ton or more per cubit centimeter.

So white dwarfs like this are ultra compact objects with ultra compact structures. But gravity can do even better than a white dwarf.

See the strength of gravity depends on mass: more mass, more gravity. So what happens in the cases of a VERY massive white dwarf? Their high mass means they want to crush things down even denser, but we’ve already reached the limits of atomic density, so where else can we go?

Well remember that atoms themselves are mostly made up of empty space, there’s a lot of emptiness between the nucleus and the electrons, for example. Atoms themselves can be crushed down into just their neutrons, with each proton absorbing an electron to become a neutron. When we crush down atoms in this way, we remove all that empty space they contain, and that lets us crush the whole star down to *even greater levels of density*.

At this level we have a “neutron star,” a star so crushingly dense that it’s the size of a city with the mass of the sun. Our friend Crabby is theorized to be a neutron star. But it might possibly be even more special.

Some have theorized that Crabby is a bit too cold for its age, and how could it have lost all that heat? The answer may bring us to the fringes of theoretical physics.

See just like atoms can be crushed down into neutrons, neutrons can be crushed down further. Neutrons are made up of 3 quarks: 2 down quarks and 1 up quark. But those quarks are much smaller than the neutron itself, and between them is mostly empty space. So once again, if the gravity of the star is strong enough, a neutron star can be “crushed down,” until its neutrons are crushed into quarks.

The whole star itself can’t be crushed like this, or else it would get so small that it would become a black hole. But the core of a massive neutron star may be so dense that its neutrons are indeed crushed into quarks, releasing all that empty space within the neutrons, and leaving a quark-rich soup that resembles the instant after the big bang. We have no firm evidence of this happening, but if it happens, then such a star would be called a quark star.

And then some massive quark stars might have a final trick up their sleeves. Remember that rule from earlier, about how 2 things can’t occupy the same space? That rule is what holds each of these type of stars from crushing down too far. White dwarfs can’t let their atoms occupy the same space, it’s only when the force of gravity is *large enough to crush atoms* that a white dwarf gets crushed down into a neutron star.

Similarly, neutron stars can’t let their neutrons occupy the same space, and that’s what keeps them from crushing down further. But when gravity is strong enough to *crush neutrons*, then parts of the neutron star might devolve into a quark star.

But in a quark star, the neutrons are crushed into 2 down quarks and 1 up quark each. Now here’s the funny thing about that quantum rule I told you about: two things can’t occupy the same space right? Well that’s only true *if they’re the same type of thing*. Two neutrons can’t occupy the same space because they’re both neutrons. Same with 2 down quarks, they can’t occupy the same space either. But a down quark and an up quark? They can occupy the same space no problem.

So the neutrons are crushed into 2 down quarks and 1 up quark right? 1 down and 1 up quark can occupy roughly the same space, they can take up 1 “unit” of space together. But that second down quark? It needs to find a *different* space for itself to occupy.

But let’s bring gravity back into the picture: we’ve already seen how it can crush atoms down to neutrons, and neutrons down to quarks, can it crush these quarks even further? Well not directly, quarks are “fundamental particles,” there’s nothing they could crush down into. But there is an *indirect* form of crushing that could happen…

Remember that 1 down and 1 up quark can share the same space but that second down quark has to find a new space. If we could change that second down quark into something else, we could free up that space so gravity could crush further, but how?

Well while all the “regular” matter of our universe, stars, planets, bloggers, is all made of just up and down quarks, there’s more quarks then just them hiding in the fringes of particle physics.

The strange quark is another type of quark, different from up and down. It’s decidedly strange indeed, since it doesn’t exist anywhere in the universe except for fractions of a second at very high energies. But that’s not important, what *is* important is that *it’s another type of quark,” and just like how up and down quarks can share the same space because they’re different from each other, a strange quark *can also share that space* because it’s different from both up and down quarks.

So at the very highest limits of density, when neutrons themselves are crushed and when their quarks are crushed further, the sheer force of gravity might force those quarks to take drastic action. Half of the down quarks from the neutrons might convert into strange quarks, so that they can take up less space with the other quarks and crush the star yet further.

If this is the case, and these strange quarks do exist at the hearts of massive neutron stars, then this may the *only* place they exist for any significant length of time. Usually strange quarks decay because they’re so unstable, but at the crushing densities here, strange quarks are forced to *exist* because it’s the only way to crush down far enough under gravity. Despite being unstable, strange quarks here *can’t* decay, because gravity is forcing them to stay strange.

Now this is all very theoretical, but it begs the question: do strange stars even exist?

Maybe

And it begs another question: is Crabby a strange star?

Maybe maybe

Like I said if the neutrons are crushed into quarks and eventually strange quarks, it will only happen at the very center of the neutron star, a place we can’t observe and can only speculate about. But there is tantalizing evidence: we know exactly how old Crabby is thanks to those Chinese astronomers who recorded its explosion, and our evidence shows it to be a fair bit colder than we’d expect for a neutron star of its age, why could that be?

It’s been suggested that when neutrons are crushed into quarks and then strange quarks, that bursts of neutrinos are created which carry away much of the star’s heat into the cosmos. It *may be* that such a thing happened to Crabby, which is why it’s so cold.

This theory wouldn’t even be possible if we didn’t know how old Crabby is, if it were just another neutron star we would always have questions about if it’s actually just an older star that has had time to cool down, rather than being a young star that is unusually cold. And we owe it all to the astronomers from the past who first saw Crabby in their night sky, nearly one thousand years ago.