A New Nuclear Technology Discussion Thread
- Thread starter UserJoe
- Start date
It's an amusing reversal, until one contemplates that the Navy cares about propulsion sans refueling, and the DOE's NNSA is the primary agency responsible for designing nuclear bombs.
How would it be better from a proliferation perspective? This proposal would be turning out completed bombs, no?
But from a proliferation perspective you're concerned about a large inventory of, say, 5% U235, because it would accelerate enrichment to 90% if the party had a clandestine enrichment facility outside of the monitoring regime. That reasoning goes out the window if you have an actual completed device ready to go. In that scenario the proliferation has already happened and actual physical possession of the device is all that matters, which drastically worse.
The magnet structural forces actually go as B**2. There has always been a problem in tokamaks about which limit you hit first, the structural limit or the first wall loading limit. People have now gone beyond 20 T with REBCO magnets.
Not all fusion concepts need HTS. Helion is one that doesn't need them. Tokamaks need them because they need high steady state magnetic fields. This is because of the low ratio of the plasma pressure to the magnetic pressure which is called the plasma beta. A steady state FRC high field design would need to use them. I would expect that a TAE reactor design would use superconducting magnets. Here's an old steady state D-He3 FRC conceptual design that uses superconducting magnets. The design also gives an idea of what the power balance is up at high temperatures. It uses a combination of direct conversion and a turbine power system to remove the wall heating from the neutrons and photons.
Tokamak designs always had blankets and shielding inside of the magnetic field coils even when they were LTS magnets at least back to the first one I ever looked at which was the STARFIRE design from 1981. The REBCO HTS revolution was about allowing compact fusion devices.
Last edited:
It's really not that hard to follow. The containment structure is considerably smaller and simpler to begin with, and the plant does not require a hot fluid circulation, i.e. you only need a single, non-irradiated, thermal loop. That's a massive part of the BoP of a nuclear plant.
Moreover, on CAPEX, you don't need to reprocess and/or cask your fuel rods, another big expense and a big physical part of the plant.
On OPEX, the whole fuel cycle is gone, you just need to make the bulk fuel material. No zirconium rods, no titanium support structures, it's suddenly a very simple device.
Pretty much every part of the plant is highly simplified. The only thing you add is a bomb making facility. Small detail
(no but for real, the entire plant is genuinely a lot simpler, but it's an unknown how a real, controlled-proliferation bomb making facility would look and what it would ultimately cost. There's a chance that observers would require this kind of facility to work fundamentally differently from the super low cost cold war era bomb factories)
Ultra low yield, unusable as bombs and very easy to control. Literature points at ~100kgTNTe yield bombs as being feasible and preferable in order to keep containment size reasonable. 10kt would require ~300m (1000ft) diameter containment and would cause induced seismic activity, so not ideal for a commercial plant.
But the inventory size can be shown to be (and fundamentally limited by) your enrichment facility. At the super low amounts of U-235 that would be needed for a power plant, you can go with ultra low yield centrifuge styles like microaccelerators (which are basically beefed up mass spectrometers).
All of proliferation is about being able to prove that everything that's visible is everything there is. Anybody can have a secret underground enrichment facility, the point is that what you're showing the observers is all you need and are using for your purpose. And the fact that you can show much less inventory being needed and physically being available at the plant would help proliferation concerns a lot.
That makes way more sense. I was thinking way further back.
I mean that sounds like a manageable explosion size, but the problem is with current technology the amount of fissile material doesn't really get smaller as the yield gets smaller. You still need a critical mass, you achieve the reduced yield by de-tuning the neutron initiator timing or similar, it's the same critical mass with fewer neutron generations. Give me a 100kg TNTe bomb and I have 100% of the uranium-235/plutonium-239 inventory I need to build a kiloton TNTe bomb.
Not sure I follow. The amount of enrichment resources needed to go to a critical mass of weapons grade U235 is not that small, it's a cascade where almost all the resources go to lower enrichment steps, because far more material needs to be processed at lower enrichments. And, the efficiency of a low yield device is going to be very low because it's been deliberately de-tuned, and because of the isotopes you use (either highly enriched U235 or Pu239) there will be very little other fissile isotopes bred. Whereas a LWR using LEU will get a significant fraction of the total energy yield from plutonium bred from U238, which will be mostly absent in a bomb design. So overall the energy yield per enrichment effort will be very poor. You might need less upfront but you reach burnup limits on LEU in about 3-5 years so the savings aren't that much.
It's more complicated than that. Because enrichment is a cascade with most of the effort at lower enrichments, it is a very dangerous scenario to have a large plant enriching to low enrichment (eg 5%) for primarily innocent applications with a small part of its output going to a clandestine facility, because a clandestine facility starting from 5% can be much smaller than one starting from natural enrichment.
You need to make highly enriched fissionable material. You need to recover that material, since the fraction fissioned in a small bomb is small. It ends up in that cavern mixed with large amounts of molten salt. That's reprocessing! The fission products also have to be removed from that salt. More reprocessing! And instead of making fuel rods, you have to make bombs. They are more complex and the have serious security costs.
Thinking about it I don't think there's any choice about recovering it because the enrichment is way too high, the salt will become dangerously reactive if the fuel is allowed to concentrate over time, especially since the salt will provide some degree of neutron moderation. And with all the fission products the salt will probably heat itself enough to evaporate if just left alone somewhere, which would concentrate whatever dissolved fuel there might be. And it's going to be very spicy because the enrichment is so high. Lots of criticality accidents waiting to happen. It's going to be the kind of situation where containers with a diameter greater than like a centimeter are forbidden in the entire building where the reprocessing happens (it being a law of nuclear safety that if a bucket exists anywhere on-site, someone will eventually use it as a place to put fuel in solution and cause a criticality accident).
The whole concept is the type of thing Rickover was talking about in his famous testimony to congress.
Z-pinches do not have a clear flow path and natural magnetic nozzle for the plasma. There are plasma configurations (FRCs and some others) that that have a natural magnetic nozzle. There are some private fusion companies that are specializing in space propulsion. Google "fusion space propulsion" and you'll find a lot of stuff. Here's an example with a unique concept that also works without fusion.
On a less serious note: yes, it's a waste nightmare.
On a more serious note: this has actually been thought about quite extensively - as UserJoe perfectly illustrates: on paper - and this is way less of a problem than you make it out to be. Again, the actual mass of fissionable material being injected into the cavity is absolutely tiny at any point in time and after it successfully detonates, there is no easy way to get it to criticality. And you can tune your purging interval to whatever you require to be less of a risk to regulators and politicians than any other fission reactor.
It's a design that is safe purely through the sheer amount of scaling down you can do, much like the promise of truly small modular reactors. And I project that this design will be just as successful as those!
Z pinches in general may not, but Zap is specifically building a cylindrical shear-stabilized confinement that gives it that clear flow path and nozzle.
Zap's confinement is closed on all but the top by thick layers of flowing liquid lithium. And the top is where the electrodes go.
The issue though is that the fission efficiency for a low yield explosion is very low, so very little of the fuel is consumed. It vaporizes, then condenses out and dissolves in the salt.
You suggested 100 kg TNTe bombs, that will still require at least a few kilograms. The best tritium boosted primaries with reflectors etc still need about 3.5 kg of Plutonium, it would be more for Uranium but say 3.5 just for the sake of argument. The efficiency is very low, <1%, so effectively all of the fissile material is vaporized and remains in the reactor. 1 GWth with 100 kg TNTe would be a bomb every 7 minutes. So it's a metric ton of fissile material every day and change. That's a lot. I don't know where the fluid becomes a self-criticality risk but it does provide moderation and it is weapons grade material so unless you're completely distilling out all the fissiles VERY frequently it's going to be scary stuff.
Realistically I don't the interval can be less than every couple days? And you're going to need a large inventory unless you're recovering the fuel from the salt and making new bombs with very short turnaround.
The extremely low efficiency of a low yield device is a really big problem in a bunch of ways. You're not getting high fission efficiency without a fusion second stage and a much higher yield.
Even without criticality in the salt, this is using fissionable material three orders of magnitude less efficiently than a conventional reactor. The economics without recycling would be utterly impossible.
Actually I think I got the math badly wrong, 1 GWth with 100 kg TNTe bombs would be every half second or so. So rather than having to process the salt very frequently I'm comfortable declaring the proposal outright lunacy with any sort of recognizable fission-based design. The need for a critical mass is an absolute deal breaker for these lower-yield designs, there's way too much inventory needed to acquire, fabricate, build up temporarily in the salt coolant, or reprocess. The only way around this would be to achieve ICF-like compressions of the fission fuel, which is not possible with chemical explosives.
Heh, yeah, that was what I was going at with the bomb frequency. Thanks for actually doing the math, I figured it had to be in the order of seconds.
It's complete lunacy, but it's damn cool lunacy.
I think I know the picture on Zap's site that is making you think there is a clear flow path and nozzle. It is a misleading image and there is stuff up past where it is cut off. Even if there was a clear path it would not be good for a plasma propulsion nozzle. The wall interactions would quench the plasma. You need a magnetic nozzle to guide the plasma flow and keep it off the wall.
It was a conversation with a fusion engineer working on a tokamak experiment that gave me this impression. It had nothing to do with their website.
TechCrunch is reporting that Helion just raised another $425MM, largely to bring production of supply-chain-limited components in-house. They cited capacitors, semiconductor power electronics, and magnetic coils as things they were buying in quantities large enough that outside suppliers were taking too long to deliver.
Bringing things in-house was also something SpaceX did, to good effect, most famously the turbopumps on their Merlin engines and a notorious support element in a cryogenic tank that a supplier screwed up, leading to an explosion of the second stage.
There's a story of SpaceX asking for bids on a component. A supplier gave a rather expensive bid, so SpaceX went with an internal program. Some months later the supplier got back in touch to see what had happened, and was informed that SpaceX had designed and qualified a part and was now manufacturing it at a small fraction of the supplier's bid. Vertical integration is one of SpaceX's not-so-secret advantages.
This seems to be about manufacturing capability for the Microsoft and other future power plants. The parts are probably custom and specific to them so it might make sense to take on the expense of building in-house manufacturing capability and hiring employees to do it. It should also help them protect the recipe to their secret sauce. Some of the stories I've read about this say that Helion has started to run its new Polaris experiment. That is the one that's supposed to show net electricity generation.
I just noticed this and want to disagree.
I find there's a "searching for dropped keys under the lamppost" vibe to tokamaks (and to some extent stellarators). Sure, their physics is perhaps the best nailed down. But the engineering and in particular the economics? If that's not hopeless, it's not far from it.
Alternatives, in particular Helion, are trading more aggressive physics for potentially much easier engineering.
I think the phrase "also works without fusion" needs to wait until fusion is a viable energy source.
Since I was polluting the other thread, I'm curious why @UserJoe (and apparently @paulfdietz) have such confidence in Helion building a commercial plant in 2028. There's been a lengthy discussion and I thought we were all pretty aligned that the TRL here is pretty low and the claims of viability so fast is pretty clearly at best puffery and at worst an investor lure, but apparently this isn't consensus?
I'd at least want to hear UserJoe out since he's a long term high quality contributor but my impression aligns with yours. You skip D-T not because it's so easy to get other reactions working, but because if other reactions really were easy then D-T would be like falling off a log. Even if you don't consider D-T practical for power generation, it still makes sense to retire risk.
Why is it claimed by some that it would be more practical to get helium-3 off the moon than source it on Earth?
Because it's convenient to think that if you're obsessed with space and you want a gold rush to make space happen.
Probably.
Less cynically, the terrestrial production of helium-3 is predominantly beta decay of tritium as a side effect of nuclear weapons programs. That ...doesn't scale well, or at least we should hope we're never in a position where it scales by an order of magnitude or more. There are trace amounts of 3He in natural gas deposits and the like, but the natural radioactive processes that yield it are a lot less common than the alpha decay which yields 4He, so "trace" usually means "below 1 ppm" and hence it's pretty expensive to extract. So, if there is a potential for a mass market of helium 3 (beyond the current uses of neutron detectors and cryogenics and similar), and if cheap space travel/resource extraction is at least vaguely plausible in the near future, it's only a mostly ridiculous (rather than completely ridiculous) idea.
Near term, the cheapest approach would be dedicated tritium manufacturing, followed by sitting around in tanks for a few half-lives and occasionally harvesting off the 3He. That'd be a security and proliferation nightmare though, and doing it outside DOE or the equivalents in other countries would probably be impossible. I guess maybe lithium spallation could scale up, but I'm not sure that would be power-neutral, since you'd basically be cracking a lithium atom and then fusing it back together again.
I think so. Grabbing at anything to justify a 1950's vision of space settlement a reality.
Also, He-3 from radioactive mixtures, specifically containing lots of isotopes of hydrogen, is super hard. You can't centrifuge it out because H-D and T radicals have the same mass, nearly the same density, nearly the same boiling point, etc. And they're almost certainly going to be present in higher concentrations. Burning out the hydrogen is a nightmare as well.
Making He-3 is surprisingly difficult, whatever pathway you choose. Although I'd still put it well below doing commercial fusion with it.
Edit: I should probably not be ultra-obtuse with my references here, as I doubt many people are familiar with the Pons and Fleischmann experiment as it's so long ago. Back in the... 80s? two Utah chemistry scientists claimed to have done cold fusion. One of the ways they thought to be able to prove this is to show the production of particular isotopes of helium and/or hydrogen in the samples. The rates of production were so low and the mixture was so contaminated with similar concentrations of hydrogen and helium isotopes that even specialized external labs were not able to conclusively discriminate between possible molecules/atoms to say if helium was being produced by fusion, all due to their ultra-similar atomic weights.
Hence when people start talking about fusion and harvesting or purifying this part of the periodic table, we all collectively think back to the good old days of cold fusion.
Making He-3 is surprisingly difficult, whatever pathway you choose. Although I'd still put it well below doing commercial fusion with it.
Edit: I should probably not be ultra-obtuse with my references here, as I doubt many people are familiar with the Pons and Fleischmann experiment as it's so long ago. Back in the... 80s? two Utah chemistry scientists claimed to have done cold fusion. One of the ways they thought to be able to prove this is to show the production of particular isotopes of helium and/or hydrogen in the samples. The rates of production were so low and the mixture was so contaminated with similar concentrations of hydrogen and helium isotopes that even specialized external labs were not able to conclusively discriminate between possible molecules/atoms to say if helium was being produced by fusion, all due to their ultra-similar atomic weights.
Hence when people start talking about fusion and harvesting or purifying this part of the periodic table, we all collectively think back to the good old days of cold fusion.
Last edited:
If you do a google search on "helium 3 from the moon" you'll find a lot of information about it including this front-page article from last year about a startup that wants to do it. The moon has more helium-3 than the earth because helium-3 gets deposited there by the solar wind.
From a discussion about the recent EAST tokamak results about a long duration run in the other thread:
There was an an even longer run a few weeks ago at the WEST tokamak. That run was 1337 seconds. I would be surprised if the end time was not decided ahead of time. There is no doubt that there are some fusion reactions occurring in the experiment, but these experiments don't use DT and are not meant to produce a lot of fusion. Producing a lot of fusion would just cause problems for them because they would be more difficult to work with because of the shielding and activation products. I don't know any details about them, but they appear to be studying long term current drive and control of the plasma. There is probably lots of useful information coming out of them just as there is from the many other tokamaks doing experiments without many fusion reactions occurring in them.
Helium is notoriously slippery. Hard to believe it just sits around there waiting to be pulled off the regolith.
From the other thread:
I'm not sure why you think it's an investor scam. Helion has competent scientists and are constructing and running experiments at a scale that was previously only possible by large government research programs. Their work is an extension of a smaller scale government run experimental program. I haven't seen their data but presumably their investors are allowed to see it and have it vetted by experts that they hire to do that. They have shown the scaling that they are using, and it has presumably been shown to be consistent with their experiments. One easy check the people vetting it could do would be to see if the Helion empirical confinement equations are consistent with what TAE has measured and published in their FRC experiments. The adiabatic compression scaling that is the key to them reaching fusion conditions is well understood and past government run experiments have been in agreement the theoretical scaling. They need to start the compression with a target plasma at good enough conditions and confinement and have a capacitor bank with enough energy to reach the required target compression ratio for fusion. I have no idea what your TRL number is based on or even what number is the maximum (5?,10?). So far, they've shown they can form FRCs, merge FRCs and compress FRCs while maintaining what they think is good enough stability and confinement. Their new experiment is supposed to demonstrate high repetition rate, net energy gain, capturing the energy to generate electricity. They will also need to show that they can make their own fuel and a gas processing and purification system. The ability to do those will determine if they can build a functioning power plant. Helion and their new investors have enough confidence that they raised a lot of additional funding for a factory to build the parts needed for the power plant. I wouldn't characterize my confidence level that they can do that by 2028 as high. There's a lot of moving parts and the engineering problems to get their systems to run reliably (if the physics works) at a high duty factor will be a challenge but it's not as complex as a tokamak power plant would be. That being said, my confidence or lack thereof had no bearing on me doing some good-natured ribbing about the possibility of Megalodon getting egg on his face.
Last edited:
They are not skipping it because it is easy to get the other reactions. They are skipping it because it really doesn't do much of anything for them. They can tell if they would get DT breakeven from just the plasma conditions without the problems that come with DT. They started running their new experiment and will either reach their targets and move on to building a power plant or they won't. They and their investors must have high confidence in the current experiment succeeding because they recently raised $425M to build a manufacturing facility to make the components for the power plant.
EDIT: All of the commercial fusion projects fall into the category of high risk, high reward but most of their big swing experiments still cost less than what it cost to do TFTR and JET back in the 80s. SPARC costs more but still a lot less than ITER. Some of the companies have already fallen by the wayside as I expect most of them will, but the rewards are great if they work. In the case of Helion you would have on-demand power plants that can be built in a factory and deployed in chunks of 50 MWe.
Last edited:
Could you stop the lying and misrepresentation, please?
Where did I say I have confidence in this 2028 goal?
What I objected to you was your obnoxious and deplorable claim that what Helion is doing has no theoretical basis.
OK, so, if you really want to know the things that have no basis in known experimental results or accepted theory:
- The field strength in the FRC that Helion is proposing are way higher than anything observed or inferred, to the point that there is serious doubt that it's even possible outside of super-extreme environments like neutron stars. Moreover, their proposed actual field strengths are definitely not enough to satisfy the... fuck, I forgot the term... the density*temperature product necessary for D-D fusion. let alone D-He3
- The proposed reactor geometry is kind of a fantasy one way or another. Either they use a super compact reactor with close-in magnets of extreme strength and have no appreciable neutron shielding and use almost all D-He3 or p-B11 fusion or something else that's aneutronic OR they have to put in a bunch more shielding - at least about a meter - for D-T or other neutron-generating fusion. In the one case, they're trying to do fusion that has never been done before and is unclear how they would get to an EVEN HIGHER density-temperature product, OR they will have to completely redesign everything they've done so far, kind of voiding the successes they've had so far.
These things combined make it so weird to have any confidence in this company, to the point that you're going to seriously consider it a contender for commercial fusion on their own timeline or being so invested into their bullshit claims that you're calling out others on the internet with strong legal language. By that I don't mean that all their claims are bullshit, just that they have a lot of obvious bullshit among their claims, bullshit that almost everybody in the field seems to be required to publish to have any chance of getting investor money.
I'm fine viewing Helion as legitimate in some sense. They can be a research company that sells - basically - pure unfiltered bullshit to technically illiterate investors and hype to the unsuspecting public with no real journalistic oversight, just so they can continue their science project that may in 20 years yield what they're going for. That's basically what you have to do to raise more money than just a meager university research budget. I get it, there is no money in fusion and there's a lot of hype around much less deserving projects. It's tempting to make a deal with the devil and ride that hype just so you can chase your pet project. Do that 100 times and maybe one of the fusion startups yields something more useful than all academic fusion research!
But it's weird to be in the energy space, discuss these things so much at length, be so knowledgeable about a diversity of projects and still have so much confidence in anything fusion, really. This stuff has so many red flags. Take a step back and consider: even completely solved, 50-year old technologies with not just experimental but commercial success in the past fails all the time - hydrogen and nuclear fission are good examples. Then, completely superior technologies that have been proven in the lab fail all the time and take decades to actually get to market - look at like 90% of silicon anode or solid state battery startups. Then, scientifically theorized or even tentatively demonstrated technologies never make it into actual production, see muon-catalyzed fusion or nuclear waste reprocessing for plain disposal. What Helion is doing is another few steps removed from all that! It's trying to do a type of fusion that has never been done, in a reactor type that has never worked using a direction of scaling that is unclear would even be physically possible, and all that outside of academia in a commercial setting within a few years, where decades of fusion research hasn't even yielded a percent of their proposed progress.