The Setting

Let's say, for the sake of argument, we have a dyson swarm. This very small torus of swarm mirrors collects over 20 exawatts of power in photons. Most of this power is lost, leaving only around 12 exawatts of power. Most of this is beamed around the solar system, so we will give ourselves about one exawatt to work with.

The Question

With this budget of power, and a large orbit reserved for particle accelerators, assuming we can get at least a 75% efficiency in the process (converting electricity into the moving particles or photons, not the total efficiency of the machine), how would we make the antimatter itself? (what particles would we smash together and how? Linacs? Synchrotrons? Superfluidic photon-smashers?)

What kind of machines would they be? (And as a side note, how efficient would they be?)

The required product is anti-hydrogen snowballs and I expect no 'new' physics to be discovered, but, techniques that have been predicted but not proven are fine. I expect to get less than 0.001% efficiency or so, and most of that exawatt of power will be radiated away, but I have no experience in this field and want to get a rough outline.

TO SUMMARISE: If you wanted to produce antimatter industrially with current physics, how would you do pull it off?

  • $\begingroup$ And also, any good ideas on how to recycle the waste energy? I'd be shocked if there's no way to reclaim some of that energy somehow. How much more efficient can we be if we do a matryoshka antimatter-heat engine system, where the waste heat of one powers the next. $\endgroup$ Commented Jan 6, 2023 at 5:31
  • $\begingroup$ And, depending on how we do it, would we want a big swarm of small particle accelerators, or do we want to build one massive machine? $\endgroup$ Commented Jan 6, 2023 at 5:32
  • $\begingroup$ I also see the issue of needing at least two different accelerator types, as you need to create antiprotons and positrons, and then guide them together with lasers and freeze them with more lasers. $\endgroup$ Commented Jan 6, 2023 at 5:38
  • $\begingroup$ There's two questions here: a worldbuilding question about the most efficient way to produce antimatter; and a physics question about why waste heat is waste heat. $\endgroup$
    – BMF
    Commented Jan 6, 2023 at 16:02
  • 1
    $\begingroup$ This is perhaps a quibble, but 8 is not "most" of 20. If you had 20 eggs and I stole 8 of them, you would not say that I stole most of the eggs, you'd say I stole some but that you kept most. $\endgroup$
    – Tom
    Commented Jan 6, 2023 at 17:49

4 Answers 4


There's a (theoretical) thing called the Schwinger effect, whereby creating an unreasonably intense electrical field can pull electron-positron pairs out of a vacuum (because of vacuum polarization).

Now, positrons aren't super interesting here, because whilst they're a great source of gamma rays they're hard to pack at high density (making them an awkward fuel) and there are a lot of useful things you can do with antibaryons (like pion rocketry and triggering fission and so on) that you can't do with antilepton.

Good news though. It is apparently possible to generate baryons via the same mechanism... there's a relevant physics.SE question which contains some paywalled links I haven't dug up, but you can take a look anyway: Pair production of quarks.

Whilst looking for a related paper, I ran across "Estimates for the efficient production of antihydrogen by lasers of very high intensities" (by Heinrich Hora, 1973, I can't find any free legitimate sources online) which isn't an entirely consumer-friendly piece of work, but the key take-home message is that pair-production of protons and antiprotons can theoretically be done by lighting up a blob of suitably dense plasma with an unreasonably powerful laser. From the conclusion of the paper,

Considering a hydrogen plasma with densities exceeding $n_{cco}$ (1021 cm-3 or a little less due to relativistic effects) where neodymium glass laser pulses of intensities of 1019 W cm-2 corresponding to field strengths of ~ 1011 V cm-1 are incident, the pulse length being assumed sufficiently long (> 10-10 s)... [snip] ...This would again increase $E^{act}$ within the next step of iteration with this last energy, we shall exceed the field strength at 1014 V cm-1 which will cause pair production of protons

Now, making a laser array that can deliver 10 exawatts over a tenth of a nanosecond or more, and then do it many many times over an extended period is left as an exercise for the reader, but if you could pull it off, you could electromagetically separate the spray of particles that came out and pocket the useful ones.

And as a side note, how efficient would they be

Astonishingly bad. Not only will the lasers be fearsomely inefficient (though by the time you get to this point in the future, you'd expect a better grasp of making decent lasers) but an awful lot of the stuff that comes out of your matter-synthesizer will not be very useful to you... lots of unstable heavy particles, mesons and so on, probably an excess of electron-positron pairs. Even if your mass synthesis is relatively efficient (which it won't be, most likely) your antihydrogen synthesis will be somewhat less efficient.

The unstable particles will decay promptly and generate all sorts of inconvenient radiation. The more stable particle pairs might be pulled away and then allowed to annihilate in such a way that you could reclaim a portion of the energy that went into them, but that'll be an inefficient process too.

Ultimately, the efficiency can be driven by the needs of your plot. Clearly, you need enough to drive your ISV Totally Not The Venture Star, and so clearly your efficiency will be good enough to provide that much antihydrogen for you. Wave your hands. It'll be fine.

  • $\begingroup$ Ah, you got me with the last part! Still laughing! Nice. So the consensus is that I can make antimatter all three ways (photons, light and heavy particles). I think that (like BMF suggested) slamming heavy ions together (she said uranium) would (by the sound of it) be simpler and easier. And, the roadmap on how it works is funny! Step 1: Mine asteroid. Step 2: Collect its garbage. Step 3: Slam the waste together with a lot of violence. Step 4: ??? Step 5: Antimatter. $\endgroup$ Commented Jan 7, 2023 at 0:41
  • $\begingroup$ It would be interesting to see how an economy handles needing to produce all of those heavy radioactive nuclei for these machines. Suddenly, nuclear waste is about as valuable as copper. $\endgroup$ Commented Jan 7, 2023 at 0:45
  • $\begingroup$ And it's also noteworthy that they (big evil empire that built the not so ISV Venture star) would have decades of these machine running before they pump the fuel into their ships. $\endgroup$ Commented Jan 7, 2023 at 0:52
  • $\begingroup$ And no! I did did not rip off James Cameron. There's loads of differences! It's bigger and an-and... (Crap, thats it. It's just bigger. A lot bigger.) frustrated Neko noises $\endgroup$ Commented Jan 7, 2023 at 0:55

Laser Pincers

Particle accelerators take so much control that you can only make antimatter from matter 1 particle at a time. Laser Pincers work by firing lasers at each other through a meta-material in such a way that the photon streams are focused into each other at the micrometer scale making about 100,000 times as many photon collisions happen for the energy of your lasers. The result is a cloud of electrons and positrons.

You can then use magnetic fields to separate the resulting matter and antimatter plasma before it has a chance to annihilate. So, not only does this theoretically give your 6 orders of magnitude more efficiency than previous methods, but it's also able to be done as a continuous process assuming you can find a way to manage the heat.

I suspect that the best option is actually to go small instead of big. You just need photons to hit each other; so, if the metamaterial were a small fiber and the lasers shooting into it also relatively small, then heat could be dissipated more quickly. So a single antimatter factory could be a cellular configuration of millions of little laser pincers running around the clock feeding into large antimatter containment chambers.

This also assumes only modern levels of meta-material design. A more advanced civilization could perhaps squeeze many more orders of magnitude of efficiency with better meta-materials. With good enough of a metamaterial you could in theory trap all of your photons in continuous loops until a considerable fraction are converted into matter/antimatter pairs. This could result in actual economically viable antimatter mass production within just 1 or 2 orders of magnitude of ideal conversion which is actually much better than your hoped for 0.001%.

Why this is better than single laser Schwinger effect generation?

The Schwinger effect can not produce anti-hydrogen snowballs by any known mechanism. It can produce positron/electron pairs but does not actually have any established theory for making proton/anti-proton pairs in a single beam system. However, using colliding neutron and proton beams fueled by decaying radioactive material funneled through a similar pincer system, you could in theory generate the needed antiprotons to create your snowballs. So, one pincer system would make the positrons, another would make the antiprotons, and then a series of magnetic fields could be used to combine the antiparticle plasmas into atomic antimatter which can then be cooled and condensed down into an anti-matter snowball.

But the most important feature of pincer beams is efficiency. Industrialization is all about efficiency, and relying on the Schwinger effect is so many orders of magnitude less efficient and than laser pincers, that no advanced civilization would chose it as a means to produce antimatter at a large scale, even if you don't assume particle recycling as an option. But since we are talking about a civilization that can make a Dyson Sphere, I think a little particle recycling is not an unreasonable technological stretch when we are so close to being able to do that sort of thing with modern technology already.

Why it might be better to make positron plasma instead of anti-hydrogen snowballs?

For starters, it is much safer to contain positron plasma than an anti-hydrogen snowball. Just because an anti-hydrogen snowball is solid does not mean it will not violently react with normal matter meaning it still needs to be contained with a magnetic field. Positrons are extremely reactive to magnetic fields because they are pure charged particles. Once you start binding positrons to antiprotons, you start making an electromagnetically neutral material. Technically you can leave it a bit ionized, and still contain it but it will be more sensitive to being knocked out of its magnetic containment than a plasma.

Secondly, going back to the economics aspect of industrialization, anti-protons are harder to make than positrons. Laser pincers can simply be powered by your nearly inexhaustible supply of sunlight, but proton/neutron pincers means you need a very large supply of decaying heavy elements. This is certainly still doable since your civilization is clearly able gather massive amounts of materials into one place to be able to make a Dyson Sphere anyway, but it's not nearly as clean and controllable of a process.

Controlling the flow of electrons compared to solid matter is also much easier; so, when it comes to getting energy out of a positron cloud, all you have to do is zap it with a well regulated flow of electricity, and you will get exactly as much heat out of the system as you need. Getting a controllable flow of power out of interactions between solids is much more difficult to do with any precision. If you accidently spray the snowball with just a bit too much matter, you could kick it through its containment which would be very bad for anyone within a few miles.

  • 1
    $\begingroup$ Good answer, and useful too, but it has to be snowballs, as it is being used in diamagnetic anti-hydrogen powered tractor ships that can reach 0.7C. Think the ISV Venture Star from Avatar. $\endgroup$ Commented Jan 6, 2023 at 16:35
  • $\begingroup$ One way to make it more efficient is to figure out if it were possible to re-culminate the beam and continue to reuses the un-collided photons. I assume you'd want to have massive, individual space station complexes that either make both and export the snowballs, or specialize in one or the other... then have dedicated recombination facilities to turn the clouds of half-fuel into full snowballs. $\endgroup$ Commented Jan 6, 2023 at 16:39
  • $\begingroup$ Thank... y'all? $\endgroup$ Commented Jan 6, 2023 at 18:15
  • $\begingroup$ Thank y'all for the answer... $\endgroup$ Commented Jan 6, 2023 at 18:15
  • $\begingroup$ That just sounds weird. $\endgroup$ Commented Jan 6, 2023 at 18:15

The theoretical limit

The theoretical maximum efficiency of turning electricity into antimatter is 50% due to the Law of Baryon Number Conservation, which mandates that energy be turned into equal amounts of matter and antimatter. So, in the limit of technological advancement, for every 100 MW of production energy, 50 MW of matter-antimatter reaction energy may be produced. (Use $E=mc^2$ to determine the mass instead.)

Robert L. Forward's numbers

Current particle accelerators aren't designed to make antimatter. Their input energy to antimatter efficiency is 0.000002%, at 100 trillion USD a gram. Robert L. Forward wrote a book with Joel Davis titled Mirror Matter: Pioneering Antimatter Physics (Wiley, 1988). In it he discussed heavy ion beam colliders, with anticipated improvements in superconducting magnets, to produce large quantities on the cheap. Beams of uranium atoms could collide to produce $10^{18}$ antiprotons per second (and a lot of nuclear fragments), at an efficiency of 0.01%. For every 100 MW of beam power, 100 W of antimatter-stored energy is produced (and the remaining 999,900 W are waste heat).

The antimatter factory could draw energy from a space-based solar array. Here's Forward's words on that from Indistinguishable from Magic (Baen, 1995):

Where will we get the energy to run these magic matter factories? Some of the prototype factories will be built on Earth, but for large scale production we certainly don’t want to power these machines by burning fossil fuels on Earth. There is plenty of energy in space. At the distance of the Earth from the Sun, the Sun delivers over a kilowatt of energy for each square meter of collector, or a gigawatt (1,000,000,000 watts) per square kilometer. A collector array of one hundred kilometers on a side would provide a power input of ten terawatts (10,000,000,000,000), enough to run a number of antimatter factories at full power, producing a gram of antimatter a day.

You say you want to use the waste heat to make more antimatter, but there's always more sunlight to take advantage of. The machinery to use waste heat will grow exponentially large due to the compounding losses (everything has waste heat, nothing's 100% efficient). You're much better off making more solar arrays.

  • $\begingroup$ Well... for one, good answer, and according to what I've gathered here, I'd ideally build a large ring of antimatter production stations in an even higher orbit than the ring of collectors, which themselves sit over the torus of mirrors. (detailed in that 'destroying a dyson swarm' post) You'd then have half the station dedicated (actually probably less) to laser colliders, producing positronium. $\endgroup$ Commented Jan 6, 2023 at 18:05
  • $\begingroup$ And the other half (probably more) to heavy ion colliders, producing... anti-protonium? Don't think thats a proper term but it is now! Then, just throw them together... slowly... into anti-hydrogen, and freeze it with lasers. And Viola! Hundreds of tonnes of fuel for going on a four-decade-long star-conquering escapade. $\endgroup$ Commented Jan 6, 2023 at 18:06
  • 1
    $\begingroup$ @SamKitsune Yeah, antiprotons are ionized antihydrogen, and you can freeze them into solid blocks if you wish. Note that with heavy ion beams some of your operation will involve mining heavy elements like uranium, either from Earth or across the solar system. If you can get the cost of antimatter to below $10M USD per milligram, antimatter propulsion becomes cheaper than fission propulsion (depending on thrust efficiency), so using uranium isn't a loss. With antimatter thermal using heated hydrogen, 10 milligrams yields about the same thrust as 120 tonnes of conventional rocket fuel. $\endgroup$
    – BMF
    Commented Jan 6, 2023 at 18:22
  • 1
    $\begingroup$ @SamKitsune you talk about building a Dyson swarm in your earlier posts. A 100km^2 solar array really isn't a big deal to build, and one gram a day is nothing to sneeze at. 1000 of those and you're making a kilogram a day, for only a fraction of your total swarm's output. Times ten years? A few tonnes. $\endgroup$
    – BMF
    Commented Jan 6, 2023 at 18:27
  • 1
    $\begingroup$ @SamKitsune I'm not sure, I don't have the book at hand. I imagine you could get away with a variety of heavy elements. Uranium and plutonium are probably the best naturally occurring options though. $\endgroup$
    – BMF
    Commented Jan 6, 2023 at 18:39

Major Frame Challenge

Assuming a merely 1000 tons for your space ship, at 0.7c its kinetic energy is around 36000 ExaJoules. We can double that if you want to brake as well, but a factor of two will not be of concern here.

Assuming a 100% conversion rate1 (haha, massive overestimation), you'd need around 200 tons of anti matter.

As mentioned in another answer, and quoting from projectrho.com making antimatter (specifically anti protons) is totally inefficient.

[...] it will take 1.50327759×10-6 Joules to make 1 antiproton, or 8.988×1017 Joules (899 exaJoules) to make one kilogram of antiprotons.

Note, we need a factor of at least 200,000, so the input energy would be around 18000 yotaJoules. Thanks to wolframalpha we know that this is around 47 times the energy output of the entire sun in 1 second (so with your current 12 exaWatts setup, it'd take around 47.5 years).

And now we're not even talking about anything else. Anti matter on this scale is the most volatile object you can think of. If the anti matter storage tank fails, it will fail apocalyptically and the entire ship is gone within a fraction of a second. Would you build a ship like this? No.

Honestly, I wouldn't go into the details here, just state that it's possible, not how it's possible. That's way better than winging it with all these contradictions and impossibilities. Suspension of disbelief only works, if there's nothing contradicting about it. The more details you specify, there more contradictions can be found. At this point, it's more plausible that the ship flies with magic.

1 just remember, if you halve the conversion rate, you need to double the other numbers (and so on).

  • $\begingroup$ Comments are not for extended discussion; this conversation has been moved to chat. $\endgroup$
    – L.Dutch
    Commented Jan 8, 2023 at 9:00
  • $\begingroup$ I would need to figure out if it is even possible to collimate a laser over 20 ought lightyears. At that point, it's better to combo the superconducting ring-thingy and small antimatter engines when diminishing returns on the ring make it pointless to keep using. And at that point things are easy, compared to how I originally thought they'd work. And besides, the ship only has to be used once, and each one has a sister ship, so catastrophic failure is fine, as long as at least one survives transit (interstellar meteors, GRBs and other space cataclysms) $\endgroup$ Commented Jan 8, 2023 at 13:10
  • $\begingroup$ The OP is only budgeting 1 exawatt for antimatter production, so actually 570 years. $\endgroup$
    – Nosajimiki
    Commented Jan 9, 2023 at 15:57

You must log in to answer this question.

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