Many questions here about power sources, not an exact fit and many are now years old and since we've reached fusion Q=1, perhaps some have better insight on future development. I need a generative non-fossil fuel, powersource for an aircraft.

I know it is impossible to predict the form of future tech. I'm only asking for informed speculation. Educated guess are welcome and approximation is expected, Breadbox? 50 gallon drum? Limousine? Aircraft carrier? An explanation of major components would be helpful.

How small / compact (weight and volume) can we expect to make a fusion power source that:

  • is capable of generating ~200Kw
  • for about 3-5 hours (no idea what the consumption rates would be).
  • include storage for fuel, assuming electrical generation components to be ~1mX1mX3m.
  • Non-cryogenic ceramic superconductors can be assumed if that is helpful.
  • Within 75 year timeframe from present.

I am assuming a tokamak type reactor, but any type theoretical or existing is acceptable. If any more criteria are needed for to help answer please ask I will do my best to provide.

Edit: the device will be used to power an aircraft in atmosphere.

The 1x1x3 dimension above refers to the actual electric dynamo producing the electricity not any size limitation on the fusion device.

Weight estimates would be nice. But volume/ size is more important, fuel capacity is also a important.

  • 6
    $\begingroup$ Room temperature superconductors? On top of fusion? Commercial fusion in the next 75 years is more plausible then room temperature superconductors in the next 75 years in my opinion. That is allowing room temperature super conductors is a second layer of speculation on top of fusion,. $\endgroup$ Commented May 1 at 21:11
  • $\begingroup$ For reference, current battery prices are about 0.2 dollars per watt-hour, so that a the desired one megawatt-hour battery should cost about 200,000 dollars. One megawatt-hour batteries neatly packaged in standard shipping containers are readily available for purchase right now. No fusion required. The point is that this is available right now, and unless the fancy fusion-based thingy is smaller and cheaper, it makes no sense to wait. $\endgroup$
    – AlexP
    Commented May 1 at 21:23
  • $\begingroup$ What's the electrical power being used to run? $\endgroup$
    Commented May 1 at 22:28
  • 2
    $\begingroup$ "Educated guess are welcome and approximation is expected". What does your plot require? Go with that. $\endgroup$
    – RonJohn
    Commented May 2 at 16:19
  • 1
    $\begingroup$ Could people stop making excuses like “go with what seems fitting to you” or “hey roomtemperature superconductors is more advanced” and just answer the question? If your comment isn’t helpful or improves the question it shouldn’t be made. $\endgroup$
    – Demigan
    Commented May 4 at 13:12

5 Answers 5


TL;DR: probably not a tokamak, probably not much smaller or lighter than a loaded 20ft ISO container.

Fusion reactions produce radiation, and you want to be able to turn that radiation into useful energy, and not into broken computers and cancer. Your fusion reactor's size is therefore going to be constrained by the amount of shielding it requires, and how you can extract energy from the reaction.

In an idea world you use direct energy conversion, where the nice hot charged particles winging out of the reaction are braked in a solenoid, which generates an electric current, and more awkward x-rays are captured in some kind of x-ray photovoltaic structure. Neutrons don't lend themselves to direct energy conversion, and require a conventional Brayon-cycle thing with a gas (probably steam) turbine and big cooling radiators and all the weight and bulk and the whole Carnot efficiency thing. So problem 1 is doing direct energy conversion, because if you can't, you aren't making a compact reactor. This probably (but not definitely) rules out tokamaks.

Problem 2 is neutrons. There are such things as aneutronic fusion reactions, but they're harder than you might initially expect. There are definitely highly neutronic reactions, and fast neutrons are a nightmare to shield against.

If you want compact fusion reactors, you probably have to discard D-T (deuterium-tritium) reactions as they release ~80% of their energy in the form of very dangerous 14 MeV fast neutrons. D-D fusion releases more like ~40% of its energy as 2.45 MeV fast neutrons but that's still quite hazardous and you do get some tritium production which will result in 14 MeV neutrons from D-T side-chain reactions. D-3He needs hard-to-come-by helium-3, and still at least 5% of the energy output will be in the form of neutrons from D-D reactions.

3He-3He is aneutronic, but apparently the range of energies of reaction products is quite wide, which hinders direct energy conversion requiring larger and much less efficient thermal power plants.

p-11B is mostly aneutronic and the fuel is readily available... about 0.1% of the energy will be carried away in neutrons, but these include 3 MeV neutrons from a side-chain involving a 14N fusion product, which are sill problematic. p-6Li looks to be aneutronic, but is hotter and produces a higher proportion of its output in the form of x-rays. And speaking of which:

Problem 3 is x-rays. Fusion reactions that produce less neutrons tend to be hotter. Hotter plasmas radiate more energy through bremsstrahlung x-rays. Those x-rays need shielding, but they also sap energy from the reaction making it harder to sustain, and harder to efficiently extract energy from than ions. X-ray voltaics have been proposed but aren't actually practical yet, and making them work is a separate technical issue to making a fusion reactor.

Aneutronic and aneutronic-ish reactions tend to be very hot, and lose so much energy through bremsstrahlung that they aren't really practical to run in a magnetic confinement fusion system like a tokamak. The x-rays they produce are much more energetic, too... 6 to 20 times higher energy than thse from cooler reactions like D-T. Athermal fusors which try to keep electron temperatures down and minimize bremsstrahlung have been proposed, and seem to be the best technology for your needs.

is capable of generating ~200Kw

Oddly, the problem here is that because there's only so far you can shrink down all the auxiliary bits... the cooling, the power conditioning, especially the shielding... the actual fusion bit in the middle can actually provide quite a lot of power if you wanted it to, because even if you did shrink it all the rest of the stuff can't be shrunk. Tri-Alpha (as linked by Monty in another answer) propose a "truck sized" reactor that can generate 100 MW. Focus Fusion propose a 5 MW device based on a dense plasma focus that seems a little smaller, though not by much:

enter image description here

("Fig. 11 Artist’s conception of Focus Fusion 5 MW generator", from Focus Fusion: Overview of Progress Towards p-B11 Fusion with the Dense Plasma Focus. The bit between the inner red rings is the layered x-ray voltaic structure, and the bit outside that and inside the outermost red ring is shielding.)

They don't state weight, or say anything about additional auxiliary support systems which presumably need to exist to make the whole thing work. Optimistically, you might end up with something that could fit in a large van. You probably wouldn't be powering something merely truck sized... trains and ships and maybe even large aircraft seem much more likely

Within 75 year timeframe from present.

75 years is a long time. 75 years ago was before the first demonstration of nuclear power generation was made with the EBR-1, 8 years before Sputnik was launched. There weren't any integrated circuits back then, and even the theory behind lasers was barely in its infancy.

Triple-product improvements being what they are, practical fusion seems likely within that timescale (the "forever 50 years away" people tend to prefer thought-terminating cliches to looking at actual progress) but actually what that practical fusion will look like is anyone's guess. I'm not sure that tokamak-style D-T fusion with tritium breeding and thermal power plants could ever be particularly compact... aircraft carriers yes, smaller things maybe not. If any of the nonthermal fusors work, or someone comes up with a new trick though, you're probably in luck.


TAE Technologies is working on a fusion reactor design "the size of a truck" with the current goal of a commercial prototype by 2030.

The TAE reactor is a long, relatively narrow machine, that would likely fit quite well into the similarly shaped body of a large aircraft. In theory, such a reactor could fit into a large commercial airliner, depending upon its mass.

It uses Hydrogen-1 and Boron-11 as fuel, producing reaction byproducts Carbon-12, Helium-4 and Beryllium-8 which decays very quickly to Helium-4. No neutrons are released by this reaction or its decay products, meaning that it has no need for heavy shielding, and the fuel would be relatively easy to store.

Since the break-even point of this type of reactor would depend upon its size, I would not anticipate that it could be reduced much below its 'truck' size. However, with R&D, a viable net-energy-producing reactor and all of its ancillary equipment and fuel might be reduced to the size and weight of a loaded large shipping container.

That's certainly feasible for a large aircraft's power source.

  • 2
    $\begingroup$ That timeline seems... unlikely. $\endgroup$
    – codeMonkey
    Commented May 2 at 14:39
  • $\begingroup$ @codeMonkey: Startup timelines. If something takes 20 years, it's fine to say 10. The problem isn't the timeline though. The problem with that startup is the size they claim, which I don't think many others do, if any. Could be something unique about the process they want to utilize $\endgroup$
    – Argyll
    Commented May 2 at 18:28

Unless we fundamentally misunderstand how fusion is going to work, this ain't happening at all, let alone in 75 years.

The problem with fusion is that you need to be able to dump an ENORMOUS amount of energy into the plasma/fuel for it to actually reach the temperature where the Coulomb force between particles is, on average, countered by the strong force. Plasmas (and pretty much every other form of matter, for that matter) dissipate heat at a rate proportional to the temperature gradient between it and the surroundings; to raise anything to fusion temperature involves creating a massive temperature gradient, meaning that whatever you're trying to fuse needs to have energy constantly pumped into it to supply what was lost.

You can't even store it in a vacuum to prevent conductive losses - fusion plasmas are so hot that they lose energy due to emitting radiation, not by conducting kinetic energy away. The stuff you need to keep a plasma at temperature is already way outside your size constraints and definitely outside your weight constraints - keep in mind that the current best fusion reactor that might have produced power at some point in its operational history is literally an entire building.

Then again, computers were also entire buildings about 75 years ago, and they're handheld now. So if you want to follow technological progress, say that we spend the next ten years figuring out how to build a fusion core - a device made primarily of unobtainium that can, with supplied input power and fuel, cause fusion to take place at a much lower temperature than would cause radiative cooling issues, so that you don't need to carry ten thousand tons of liquid nitrogen and cooling systems with you on your aircraft. Let's say that, like with computers, a fusion core's size will halve every two years with minimal rise in cost, and that one fusion core, when supplied adequate hydrogen, can generate as much power as you need (we're already pretty much assuming cold fusion, so...).

We'll start with the International Thermonuclear Experimental Reactor (ITER), which is one of the most cutting-edge fusion reactors out there. Ten years from now, someone successfully tests a fusion core that is the size of the ITER which can produce as much power as you need. That leaves 65 years until your fusion-powered aircraft needs it. Plugging in the values for the exponential, and taking into account that the ITER's internal volume is about 840 cubic meters, we get a resulting size of $\approx1.38\times10^{-7}$ cubic meters, or about the volume of a cube 5 millimeters across.

Great! It fits in your ship and it generates all the power you need. Of course, a fusion core 5 millimeters across doesn't exactly sound reasonable to viewers/readers, so you can just expand it to fill the 1x1x3m engineering bay and say that the actual "fusion" takes place in a very small part of the core and the rest is just for show/cooling/control/safety mechanisms/whatever.

As for fuel, if you can get it started with a little bit of liquid hydrogen, you could actually use the excess power it generates to electrolyze water in the air and produce both oxygen for your crew (very valuable at high altitude if your aircraft is really high up and needs oxygen recyclers) and hydrogen for the fusion core, which I would imagine powers some sort of ion engine or ramjet.

  • $\begingroup$ +1 I like how you disregard the current assumption that deuterium/tritium are needed. Hurrah for protium. $\endgroup$ Commented May 1 at 21:19
  • 1
    $\begingroup$ Just going to say ITER is not cutting edge. It was designed based upon low temperature superconductors and has not been redesigned to reflect that. The huge boom of new designs in the last 20 years reflect the fact that high temperature ceramic superconductors are now a commercial product. $\endgroup$ Commented May 1 at 21:25
  • 1
    $\begingroup$ Proton-proton fusion is very difficult and inefficient (en.wikipedia.org/wiki/…). It usually results in a diproton which immediately decays back into two protons. It very rarely produces a deuterium nucleus, but the cross section for this is so low we haven't been able to measure it. You're way better off getting natural deuterium than trying to produce it by fusing protium. $\endgroup$ Commented May 2 at 2:01
  • $\begingroup$ I appreciate the answer but I feel you may have missed the question. The 1x1x3 measurement I referenced was.the assumed sise of the actual generator (dynamo?) The fusion core will ring. My question is how small hypothetically can we possibly hope to make a fusion power source. $\endgroup$
    – Gillgamesh
    Commented May 2 at 3:13
  • $\begingroup$ Not with H-H fusion... but there are other fusion reactions. $\endgroup$
    – Monty Wild
    Commented May 2 at 5:04

Might I suggest muon-catalyzed fusion?

The muon is an elementary particle very similar to an electron, but about 200 times more massive. They can bind to atomic nuclei in a similar way to how electrons do, but due to their higher mass, their orbitals are much smaller. A muon can screen the charge of a hydrogen (or deuterium, or tritium) nucleus, allowing it to get much closer to other nuclei than it normally would. So close, in fact, that the strong nuclear force can take hold, allowing nuclear fusion to occur with reasonable probability at room temperature.

Unfortunately, muons have a very short lifetime. Also, while the muon usually gets kicked out after the fusion event, allowing it to find another hydrogen nucleus to fuse, there's a chance it will get stuck to the helium nucleus that results from D-D or D-T fusion. As the helium nucleus has two protons, the muon can only screen half its charge, which isn't enough to catalyze fusing the helium into heavier elements.

In the end, a single muon can only catalyze a couple hundred fusion events before sticking to a helium nucleus, and then decaying into an electron and a couple fo neutrinos.

Also, producing muons requires a significant amount of energy. The most efficient method known today requires more energy than each muon is likely to release before getting stuck to a helium nucleus, even if all the energy it produces could be captured.

However, if your scientists can come up with a more efficient way of producing muons (or solving the helium-sticking problem, but that's likely much more difficult), then muon-catalyzed fusion might be viable.

If so, it could be much more compact than traditional "hot" fusion. No need for big magnets or superconductors or plasma containment. You'd need a bottle of hydrogen isotopes and whatever equipment is needed to produce the muons. Currently, that's done with a particle accelerator; I'm not sure how compact that could be made, and at any rate whatever more efficient muon-producing method your scientists come up with could easily be more compact as well.

You'd still need heavy neutron shielding, though. D-T fusion releases a lot of neutrons, muon-catalyzed or not. And the various aneutronic fusion processes being researched don't work well with muons, since they require nuclei with more protons than hydrogen, which would require more than one muon to screen their charge. And given muons' short lifetimes, you're unlikely to get more than one on any given nucleus at a time.


If you want to have some unexpected development actually giving commercial fusion, you could look at the Polywell reactor.

This invention has an impeccable pedigree, with contributions from Farnsworth and Bussard. Consensus of informed opinion such as this is that this cannot work well enough to be productive, but there is enough wriggle-room that some think it might work if only... Sadly, the development seems to follow the pattern set by cold fusion, room-temperature superconductivity, cloud seeding, and many other things that would be disruptive technologies if only we could get really strong magnets or enough unobtanium. Or maybe the wiffle-ball fusor. But never quite enough evidence to convince us.

If you want a fictional power source, this might be a good place to start. Most of the attempts come with specifications for the proposed power source.


You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .