Suppose we mastered nuclear fusion to the point where it becomes commercially viable and provided the lion’s share of our energy generation. What would the overall fuel cycle in such an economy be like? Would we produce hydrogen electrolytically from sea water to provide fuel for the reactors? This electrolysis would produce oxygen as a byproduct. Would it be simply released into the atmosphere, or would that cause problems by increasing the oxygen content of the atmosphere? What would happen to the helium that would be generated as byproduct of the actual fusion?
The key thing to remember is how little fuel you need for the energy you get. You would not be able to measure the impact of additional oxygen. Not even next to a plant creating fuel. Same with the benign fusion products. Helium is a valuable gas. No chance a fusion reactor would make enough to make its capture for sale worth the effort.
Worrying about the environmental impact of fuel generation is mostly a non-issue.
Ordinary hydrogen mostly doesn’t take part in any of the useful reactions. For most possible reactions we need either or both deuterium and tritium. Deuterium may be obtained from ordinary water via annoyingly laborious separation. Given it is double the weight of ordinary hydrogen it isn’t especially hard. Tritium is a much harder question to obtain. With a half life of 12 odd years, there isn’t any in the wild.
Which brings us to lithium. Lithium can be irradiated in a breeder reactor to create tritium. It can also participate directly in reactions, including an “aneutronic” reaction. (Scare quotes because it isn’t really.) Going down the route to use lithium to feed fusion reactors (if they became a dominant energy source) could realistically deplete the usefully available lithium in the world enough to really matter. By which one means the economically recoverable surface deposits of the stuff.
OTOH, if someone could work out how to extract the lithium in seawater in an economic manner, it would never be an issue.
Given the lithium is really truly totally gone forever, this is potentially a bigger deal than is often realised.
Note that a fusion reactor would produce copious amounts of thermally energetic neutrons. These would be absorbed by the reactor walls and radioactive atoms would ensue (depending on the materials of the walls, etc.) In addition to making the container vessel radioactive, the material will be quite prone to weakening. So, the walls will have to be designed to be regularly replaced and disposal of the old material will have the usual problems.
Also, gamma rays are produced for extra fun. Then there’s any unburnt tritium that gets scattered arojund.
It would depend upon both the fuel system being used and the type of reactor. Although the primary fusion reaction that occurs in our Sun and most stars on the main sequence is the proton-proton (p-p) chain which uses the most common version of hydrogen, this form of fusion will never be viable for terrestrial power production because of the extraordinarily high density and temperature required and low capture cross section; it only works in the core of a star because of the inexorable pressure of gravitational confinement, and even then it produces such a low rate of specific power that takes most stars many billions of years (in the case of red dwarf stars, hundreds of billions or trillions of years). The only practical forms of compact fusion power production would use deuterium-deuterium (D-D), deuterium-tritium (D-T), and deuterium-helium-3 (D-3He), with D-T being strongly preferred because of its preferable triple product properties at the lowest temperature. (Yes, there are people who believe that they are going to achieve aneutronic fusion with p-11B or 3He-6Li reactions, but the potential of getting sufficient yield achieve a ‘burning plasma’, much less net power yield to sustain fusion conditions, is so much lower than the above reactions that it would basically require some kind of magic to achieve and sustain confinement.)
One of the challenges of the fuel cycle is that while deuterium is stable and relatively common in sea water, tritium has a half-life of 12.32 years and basically doesn’t occur in nature in any quantity outside of a star, so it would have to be bred from 6Li, which is a limited resource. 3He is stable but rare in nature, as it is a decay product of tritium. You may have heard a lot of nonsense about ‘mining’ 3He from the Moon but what this would really involve is sifting through the electrostaticaly charged Lunar regolith for a tiny fraction of 3He which was captured fro the solar wind; it would literally be easier (although still very difficult) to capture it directly in space. Tritium could be bred by using the neutron output from a fusion reactor impinging upon a lithium ‘blanket’, but getting sufficient yield to be sustainable is an enormous engineering challenge that just may not actually be viable since any losses would mean that it would extract less tritium than what is being consumed by the process, and thus would require some other more sustainable form of neutron production.
As for the byproducts of the fusion power process, helium-4 is chemically inert and thus effectively not a contaminant, and because it is only produced on Earth very slowly by radioactive decay in the mantle, it is a scarce resource that has many scientific applications. Oxygen naturally binds to just about anything other than the noble gases, and the amount that would conceivably be produced is an insignificant fraction of what is currently in the atmosphere. The biggest ‘pollutant’ from common fusion reactions are fast neutrons, which again can be used to breed tritium but will also damage reactor and would pose a health hazard for unshielded people nearby, as would many spallation products produced by neutron collisions.
Although you didn’t ask, there are substantial hurdles to practical fusion power production that are often glossed over in many of the hype-laden popsci articles and entrepreneurial pitches; aside from the fuel limitations described above, extracting sufficient power from the system in a useable fashion to maintain continuous confinement (or, for inertial confinement methods, achieving confinement pulses repeatedly) doesn’t just require achieving a burning plasma with breakeven net plasma power output (Q>1) but actually requires net energy yield through all losses in the system, which is more likely to be somewhat more than Q>10, and potentially Q>100. So just because a system gets a reported greater yield from the system than required to heat the plasma to Q>1 conditions (as reported by last year by the National Ignition Facility at LLNL) doesn’t mean it has been demonstrated as viable for power production, and many of the startups claiming to be within a few years of commercial viability but have yet to even demonstrate a breakeven condition are an order of magnitude away from whole system net power yield, and in many if not all cases just improving on power throughput efficiencies alone just may not be sufficient to get to a sustainable net fusion power yield state.
If you want to actually learn about the different approaches to nuclear fusion power generation and the history behind them, Matthew Moynihan has a good series on YouTube that is pretty accessible for a layperson but goes down into the weeds on what has been tried and not found not viable, and what approaches still offer some promise. Personally, I’m doubtful that we will achieve nuclear fusion power production in the foreseeable future without some fundamental breakthrough in physics such as making muon-catalyzed fusion viable or otherwise substantially reducing the threshold to achieve fusion conditions, because we are just blocked in by the sheer amount of power input, fundamental limitations on material thermal capability, and scaling problems that limit specific yield. But I’d love to be proven wrong.
Stranger
There’s some very thorough stuff here, thanks for that.
OP should know lithium also helps with for energy transfer and conversion. Fusion produces a lot of energy, but the process for getting electricity out isn’t straightforward — it’s not like we can run steam pipes inside the tokamak donut. Most of the energy from DT fusion goes into the neutron, which conveniently leaves the plasma, and interacts with the surrounding structure. Intercepting it with lithium results in heat (and tritium, see above), which we can use for steam to push a turbine.
Separating and concentrating that tritium and getting it back into the reactor is a whole chunk of research dollars on its own.
Tritium is also made in heavy water fission reactors (e.g., CANDU).
Doesn’t seem like a huge deal.
World electrical production is about 1.04e20 J/yr. Fusion releases 2.8e-12 J per tritium nucleus. Divide to get 3.71e31 tritia/yr, or 6.17e7 moles tritium.
You get one tritium per Li6, and about 5% of all lithium is Li6. So we need 1.23e9 moles lithium/yr, or 8640 t.
Total lithium production in 2023 was 180,000 t. Obviously there are huge inefficiencies in fusion, but 10x that number is still well under current consumption and even 100x is not that unreasonable (and you can use the remaining Li7 in batteries). Known lithium reserves are ~100M tonnes, and that will undoubtedly grow over time. Fusion is potentially such a high-value use that it will justify pulling from low-grade sources.
Obviously all this requires a lot of stuff, including the tritium breeding system and enrichment equipment. That’s still unproven. But I don’t think the raw quantity of Li6 is the long pole here.
Well, eventually it would be. Like coal, it’s ultimately a non-renewable resource, and would run out long before (for instance) we ran out of hydrogen.
There’s quite a lot in the ocean, though a pain to extract. Maybe enough time to get p-B fusion working. Or just harvest the energy from the p-p fusion reactor that’s already operational.
Correction to the above:
Li7 can produce tritium as well. The reaction is more rare (except at very high energies), but it has the advantage that it doesn’t “eat” the neutron. So you can get perhaps two tritia per neutron.
In fact this is crucial if the reactor isn’t to be a net consumer of tritium. Each fusion consumes one tritium and emits one neutron. So you had better produce more than one tritium per neutron. And probably substantially more since there will be a number of inefficiencies in the process.
Doesn’t change the order-of-magnitude view, at any rate.