Is it possible to build a garbage recycler that turns residues into its constituent elements using plasma and accelerating it through a magnetic field?

The technology level is around that of the typical space opera: starfaring civilizations. Physics and chemistry are the same as ours.

Energy input and disposal is not a concern (we could include a fusion plant or ten to the project, but we do not have unlimited power!).
This is a space based recycler, so hard vacuum is easy to access, but there can be atmosphere artificially provided if necessary. Resting frame of reference.

The material to be recycled is frozen (few Kelvins) compacted heterogeneous mixes of waste in 10m sided cubes (in the order of tens of thousands of tons). The cubes are covered in a film that prevents out-gassing and keeps any elements or composites that manage to be liquid at the said temperature. Encompassing everything from paper sheets and fish poo to shredded nuclear reactors, cars and solid blocks of reinforced concrete. The blocks can have any element of the periodic table, and all of them are to be expected in any proportion, including all useful alloys, compounds, organics and textiles ever known.

The film can be of any material, though it's planned to be a carbon-based polymer of some sorts. It can be removed from the cube prior to process if necessary. The cube can be made smaller or even shredded if required by the process, but bigger chunks are never going to be provided. If it needs to be made into dust/fine particles, the "atomization" process should be part of the recycler. If the mix needs to be preheated, it will also be part of the process.

There should be no expectation that the mixture has the same proportions of element rarity present in the universe or Earth's crust.

Noble gasses and the most heavy and radioactive elements can be mixed at the end of the process, if it's too hard to differentiate them. PREFERABLY NOT!

It can be an expensive piece of infrastructure: it's meant as a one of a kind per heavy populated star system (in the order of 10^7 citizens). Its purpose is to be the end step of traditional recycling process leftovers.

It's located in space, it can be near the main star or well past the ice giant portion of a star system, wherever you need it.

Size is also not a concern. Single structure designs are preferred over "distributed" infrastructure, for defense purposes, but will do what's necessary. Should be less massive than a little moon (10^19kg).

It does not need to be practical. It's meant as a "safeguard" against interstellar resource blockades (like subsidized agriculture in some countries).

It need outputs in the magnitudes proposed to be able to keep up with demand during wartime on systems whose planets do not have enough tectonic activity to produce veins of vital heavy metals and elements.

During peacetime it's used to just simply complete a loop of resources in space where necessary, maintained by deep space dwellers to avoid being captives of planet surface inhabitants.

As strategic infrastructure, its existence comes first. More productivity and more efficiency are the secondary really important points, and economic sense is the third concern, but some systems need one to be anything else than colonies.

Previous research and ideas



I've envisioned a toroidal fusion power plant that is fed a stream of fine grained garbage along with the usual fusion fuel. The stream would get turned into plasma and then ejected to a magnetic accelerator that uses atomic weight and magnetic fields on the curves of a loop to separate the stream into its constituent atoms in separate receptors.

Is this or another design feasible?

How does it work?


Synchrotron, the shape that resembles the proposed design

Bonus points:

  • Approximate volumetric size of the plant (capable of processing 1 cube per day)
  • The order of magnitude of energy or fuel consumed or produced (for any fusion or other kind of reaction(s) you choose for the process)
  • How well does it scale (up and/or down)
  • Containment method for the elemental outputs


Based on answers, it seems clear that mass is really relevant to the segregation process. Please, feel free to include a centrifuge or whatever you come up with to use mass to your advantage.

  • $\begingroup$ Sure it's feasible for a space-faring civilization with unlimited power. $\endgroup$
    – RonJohn
    Aug 3, 2018 at 8:29
  • $\begingroup$ It's not unlimited. That it's not a concern means we could dedicate two nuclear power plants to it. I'll note it, thanks. $\endgroup$
    – Oxy
    Aug 3, 2018 at 9:01
  • $\begingroup$ It can be removed from the cube prior to process if necessary Not neccesary given that The blocks can have any element ... ever known $\endgroup$
    – user3106
    Aug 3, 2018 at 12:37
  • $\begingroup$ @Jan well, if the process has a net output of energy, the envelope can be included. It may even be fuel for the reactor. If it ends up being an expensive process, then removal of the envelope will allow meeting the efficiency concern, thus "necessary". $\endgroup$
    – Oxy
    Aug 3, 2018 at 12:51
  • $\begingroup$ Related to one of my earliest questions. I don't believe it's a duplicate, but it will have additional insight. $\endgroup$
    – JBH
    Aug 3, 2018 at 15:38

3 Answers 3


TL:DR feasable? Yes. Practical? No.

This design is not really tenable in it's current form for a few reasons.

Firstly plasma is not at all dense (or at least isn't in most man-made fusion reactors) very little hydrogen is used in a fusion reactor and so as soon as you add your finely ground trash it will likely cool down to below the threshold needed for fusion meaning essentially you'd be just as well off (likely much better off for reasons I won't get into) vaporising and ionising your trash with any other form of heat like a laser, strong electric current, concentrated sunlight or intensive radiation (which could be from a fusion reactor) and then putting it through the magnetic seperator.

The other problem with plasma's low density is the fact that the 10m solid cubes you are burning up will produce a massive volume of plasma (like really big.) and this plasma can't be high-density because otherwise particle-particle interaction will dominate and it will be much harder to separate. This means your plant will either have to work incredibly slowly only doing a few grams per year or will have to send plasma through at tremendous speeds which means you'll need much more powerful magnets to send them on those curving paths and will be using much more energy per kg of mass sent through than would be reasonably tenable by any civilisation without something like blackhole power generation.

Your design will also suffer from problems arising from double ionisation. because it separates based on specific charge. Lets say we have two ions going through the detector, Carbon-12 with one electron missing and Magnesium-24 with two electrons missing. Theese two ions will be treated almost identically by the separator the only difference coming from their binding energy which will be so minute that the track would need to be almost a hundred times longer. This is likely to always be a problem too whatever method of ionisation you use, if you ionise it more you'll get double, single and triple ionisation less and elements will be left behind.

all this being said you device could still be used as PART of a larger macro-recycling system, oncee you've taken out everything you can normally (I.E. chemically and physically) some waste is bound to still be there. This is what you then put through a industrial scale mass spectrometer and split into constituent elements.

The massive energy concerns might not be too bad either. In theory all the energy goes into the many steams of fast moving ions and a little bit of cyclotron radiation both sources of energy in their own right that can be easily recycled and fed back into the system. If you can fare space then you can probably also generate a lot of energy very quickly.

So while i can't see this becoming the main way we deal with trash it might very well be part of a larger system for dealing with that really pesky rubbish that won't go away.

EDIT: regarding your bonus points Size could vary really quite a lot in general bigger plants will be more successful at separating elements out more cleanly and will be easier to maintain while smaller plants will process waste faster but might prove much harder to construct.

Best case scenario for energy consumption is the enthalpy of formation of the waste which for any space faring civilisation is basically 0 more likely it's going to be determined by the rate at which waste needs to consumed, the efficiency of their tech, the elemental composition of their waste, and a thousand other variables we don't know.

it scales really well actually in fact i'd advice building these plants in deep space where vacuum is cheap and you can spread out as far as you want.

The elemental outputs are fist going to be sent through a solenoid so that energy can be recycled from the fast moving matter streams, the now slow elemental beams can simply be sent to electrodes that will extract energy from the charge of the ions. In the case of gaseous elements its going to be a lot harder, for the reactive elements i'd advice using a calcium electrode which you then use electrolysis to extract the gas back off of. Noble gases are going to be harder but not impossible you might need to fire them into somekind of solvent they dissolve in (a pure bitumen perhaps) then extract them later.

  • 1
    $\begingroup$ Scale is not really a problem, it can be 20 km3 if it need be. It's meant as a one per star system on populated systems, wich can easly be 5e7 peoples. Efficiency is also not a concern, it's as you mention, the last step on the recycling pipeline meant to recover the most valuable and hard to recover elements. I'll add clarification. $\endgroup$
    – Oxy
    Aug 3, 2018 at 10:22
  • $\begingroup$ I appreciate a lot the points about size relating to quality of output, speed and expenses of the machine. That kind of gradient greatly helps. And of course the reminding that plasma poses density problems. Do you have any idea about how to solve the double ionization problem? $\endgroup$
    – Oxy
    Aug 3, 2018 at 11:23
  • 1
    $\begingroup$ in most cases it should be possible to chemically separate out the atoms that were doublely ionised after you've collected the atoms. i.e. in the case of carbon and magnesium you can simply burn it. the carbon will come out as CO2 gas and the magnesium oxide will stay as a solid. if you what a more direct approch you could always try firing tuned Free electron lasers at those streams to further ionise certain specific ions but not others this would hopefully split the stream of mixed ions in two. $\endgroup$
    – Ummdustry
    Aug 3, 2018 at 12:09

Very likely

It would probably work. Centrifuges are used to enrich uranium (meaning separate U-235 from U-238) by turning it into a gas and then exploiting the fact that the two isotopes have different weights, so when you spin them, they will move at different speeds. This appears to be using a similar idea.

It would be a beast of a thing to design and make it capable of doing any element. But if today we have uranium gas centrifuges, a plasma disintegration centrifuge is plausible, even one sensitive enough to be a universal disassembler. It would likely take decades or centuries of technological development past what we have today, but if your story is far enough in the future, that's not a problem.

Efficiency: Turning things into plasma requires the expenditure of a lot of energy. This plant would likely require a nuclear reactor, probably of a heavily optimized design that directly converts the garbage to plasma in the volumes required and only incidentally produces enough conventional power to keep itself running.

Containment method: The elemental outputs will likely not be energetic enough to undergo nuclear reactions, so you can just shoot them into a chamber full of water (one per individual element). That will decelerate the ions enough to stop them in the receptacle and turn them back to ordinary atoms. This will release Cherenkov radiation (which you can likely use to generate power, improving the efficiency somewhat). The atoms will react with the water; only noble gases like to hang around in naked elemental states (never mind that they'll violently 'liberate' electrons to refill their own depleted shells), so you will need to use ordinary chemical engineering processes to refine them out.

[Disclaimer: I am not a physicist. Someone who is can probably nitpick holes in this.]

  • $\begingroup$ You've reminded me of isotopes! Completely forgot that... So instead of 110~ output "bins", it may end up needing 500 or a thousand... depends on the radioactivity of the input. But using the same process as uranium enrichment would base it off weight, not charge. I wonder if the magnetic process can be effective enough to dissociate based only upon protons and manage to ignore neutron differences in isotopes... (Guess we'll have to build done to find out!) $\endgroup$
    – Oxy
    Aug 3, 2018 at 9:35
  • $\begingroup$ The water containment method is a problem. We went that far to dissociate everything to end up with elements. I know it's better to need two hundred chemical plants to dry and separate every element from hydrogen and oxygen than to need one process to recycle every molecule present in the input (probably thousands of variations, if not magnitudes more). But it's a problem. $\endgroup$
    – Oxy
    Aug 3, 2018 at 9:52
  • $\begingroup$ It's not so much "nitpicking holes" as noting the vast chasms. (See @Ummdustry's good answer.) The energy requirements for this are huge and if we had that kind of cheap energy to play with, there are much easier ways of recycling trash. (Not that space is a good place for a garbage dump, but it takes more energy to turn matter into a plasma than it does to lift it entirely out of Earth's gravitation field.) $\endgroup$
    – Mark Olson
    Aug 3, 2018 at 13:06

Good thing you've got fusion plants

Because you're going to need a lot of energy. Of critical importance is the ionization energy of all the elements. As plasma contains normal elements stripped of their electrons, enough energy will need to be pumped into each trash cube to liberate enough electrons to form a plasma.

The energy required to energize/plasma-ize 10m^3 of random garbage will be immense. At the low end, let's continue consider hydrogen. It takes 1312 kJ/Mol for complete ionization. But hydrogen is easy. It's already a gas.

Iron is trickier. It's a solid and does not yield energy through fusion. Iron's ionization energy is 762.5 kJ/Mol for the first electron. Ionization energies rapidly increase for each electron there after.

I don't know if iron has to be heated to boiling before it can turn into a plasma. If it does then the energy requirements for this reactor are staggering. Assuming linear specific heat values at all temperatures of 1kg of iron of 449 J/kg K; it takes 1407166 J to boil 1kg of iron. Add in the heat of vaporization and it gets even more expensive.

Reactor Design Considerations

There's a couple things you'll need to account for in this reactor.

  • Don't let the plasma touch anything. I've seen temperatures of 30,000 Kelvin in the literature. With the energies under discussion, the plasma temperatures may be considerably higher. Normal matter does not perform well when hit with high energy atoms. This is a long standing problem in fusion power plants under development now.

  • This is not fuel. Anything iron and heavier does not yield energy when fused.

  • Injecting cold matter into this reactor will require ridiculously high energy flows to sustain. Each 10m^3 will need to be heated to plasma temps, which will vary wildly from cube to cube. As mass leaves the plasma chamber, it will take thermal energy that must be restored in order to maintain the plasma.

  • The reactor must contain the rapid expansion of gas phase products. Water expands 1700x when turned to steam. Other solids or liquids may expand even more. Not only must the reactor contain the expanding gasses but also the high speed chunks of arbitrary mass accelerated by those products. In other words, without preprocessing, each cube is a 10m^3 fragmentation bomb.

  • Watch out for the really caustic elements such as fluorine and chlorine. These elements aren't bad when bound to other elements at room temperature, but this reactor is not normal. Special precautions will have to be taken or the reactor may eat itself.

Design trade-offs

Okay, so somehow the reactor works. It separates each element and doesn't destroy itself. You've expended an enormous amount of energy to get a very hot gas. Well done!

However, you now have a big hot cloud of random elements that need to be sorted, cooled and turned into products that can be sold. How will you sort these? How will you prevent unhelpful chemical reactions when the atoms cool down?

Having done a little research into hydrocarbon plasmas, I can say without a doubt that this area is crazy complex and very hard to manage, even with single element plasmas. This reactor must accept all elements in any ratio or quantity. This is the pdf manual for a real life plasma calculator that only does one element plasmas. All of those input variables will change between batches and perhaps within each batch.

  • 1
    $\begingroup$ Well, those numbers are workable. Ten thousand tons of iron are 1.791×10^8 moles, multiplied by the energy required to ionize a mol you provided is 1.365×10^14 joules =~ 37910 MW h. For a power plant to produce this amount of energy in a day, it must ouput 1,58 GW. Thats three nuclear reactors. Not the worst I heard. Boiling the iron is 200 MW more. I suspect that the magnetic coils and supercomputers required will not be comparable, but worst case scenario, it can need about two to three times that power. "So you are saying there's a chance?" $\endgroup$
    – Oxy
    Aug 3, 2018 at 14:01
  • $\begingroup$ Heating and ionization can be done slowly in order to not uncontrollably deconstruct the installation. Remember that one cube per day is a rate of work. There can be many cubes slowly heating, and the more volatile and less demanding elements that already vaporized off of one cube can be separated together with the slow elements of a much earlier cube. The main issue is to sort the plasma, of course... Do not mind the complex calculations, the technology is far from ours. There have been patient scientists with studying the properties of each plasma for tens of generations. $\endgroup$
    – Oxy
    Aug 3, 2018 at 14:07
  • $\begingroup$ @Oxy Sure, there's a chance and one can handwave all manner of things to make a plasma recycler work. However, I can't help but think that there's a cheaper, less energy intensive method to recycle random stuff than to turn it into a plasma. Remember by reducing everything to plasma, you lose a lot of useful compounds like any hydrocarbon. To resynthesize those hydrocarbons will require still more energy. If energy is that cheap then fine but it seems like a waste to me. $\endgroup$
    – Green
    Aug 4, 2018 at 17:07

You must log in to answer this question.

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