So, especially in the outer solar system, where refining Helium3 from the atmospheres of gas and ice giants is rather trivial
If you first assume that extracting and exporting stuff from the atmospheres of gas giants is trivial, then production of tritium from helium 3 seems pretty credible.
Thing is though, they all have crushingly deep gravity wells. Neptune has an escape velocity from the datum level of 23.5 km/s, and an atmospheric scale height twice that of Earth, meaning that gravity and aerodynamic losses are going to be even higher than on Earth. Even without dealing with atmospheric density or the hazards of operating within a gas-giant atmosphere, the energy cost of raising something from Neptune's datum level to the edge of the gas giant's sphere of influence is 275 MJ/kg. Compare with Earth's 62 MJ/kg. Earth also has abundant solar power available, meaning you don't have to burn up a load of the fuel you're lifting out of the gravity well in order to get any fuel out. There's also presumably a whole load of infrastructure there already.
How might such a production facility look like, especially with the thermal neutron source?
It'll probably look like any other big industrial facility in space, eg. we've no idea having never built one so you're basically welcome to handwave it yourself. Probably boring and boxy, with an outer Whipple shield to protect against micrometeorites. There will be big heat radiator arrays to dump all the heat generated by the neutron source and the decay heat generated by the $\require{mhchem} \ce{n + ^3He -> ^3H + p}$ reaction (which is hot... 24TJ/kg!). They won't be glowing in the visible spectrum... just big slabs of stuff parallel to the Sun's rays so they don't gain additional heat (not that there'd be much of that at Neptune's orbit, but still).
The neutron source will probably be a fusion reactor, given the greater availability of fusion fuels in the outer solar system, with heavy water as one reasonable choice for moderation. D-D fusion seems like the most likely source of neutrons, with the deuterium refined from ice in the outer solar system. That saves burning up yet more of the helium-3.
Now, D+D fusion has two reaction paths:
- $\ce{^2H + ^2H -> ^3H + p}$
- $\ce{^2H + ^2H -> ^3He + n}$
You can't recover the tritium it produces, because D+T fusion is just too easy. It'll almost immediately fuse as $\ce{^2H + ^3H -> ^4He + n}$. This means that you're guaranteed one neutron per pair of deuterons, and interestingly you might also be able to recover that helium three nucleus "for free".
Have a look at this chart showing Lawson criterion curves:
You'll see that it is possible to run a D-D reactor at a temperature too low for D-3He fusion... this isn't the most efficient way to run the reactor, but you're not interested in efficient power generation so much as efficient neutron generation. This sub-optimal plasma means that most of the generated 3He could be recovered from the core, giving you a steady additional trickle of helium-3. You'll need an efficient way to separate it from the helium-4 generated by the D+T side chain, which means you'll also have a load of big gas centrifuges attached to your facility. It'll probably be cheaper to operate those than to fly up twice as much gas from the planet below.
This means that for every 5 moles of deuterium you burn in your reactor, you can extract 1 mole of helium-3 and one mole of helium-4, and must inject one additional mole of helium-3 from some external source, eg. your gas giant mining operations. Your refinery will necessarily need big gas tanks attached to it... these will probably be on the outside, safely away from the emission pattern of the big heat radiators and they'll probably be spherical.
Weirdly for an industrial process, this will be quite net-power-generating... for every 5 moles of deuterium and one mole of helium-3 you put in, maybe as much as 11 MeV per triton. I can't tell how much energy you'd need to feed in, especially given the sub-optimal conditions in your D+D reaction, but the chances are you'll be able to run all the rest of the facility and have energy to spare to do other things like beamed power, or laser launch. Maybe there are big microwave antennae or laser turrets on the outside of the facility, too.
Whatever the eventual shape of the reactor, and the technologies you use, there'll be a jacket around the core to provide shielding and cooling. This would be where you'd run your fertile feedstocks for breeding tritium from, eg. lithium-6 or helium-3. Helium-3 has a pretty high thermal neutron absorption cross section (significantly higher than lithium-6, though the density is going to be lower) so repurposing existing tritium breeder systems shouldn't be too hard... just change the plumbing a bit.
Lithium is somewhat hard to come by
It'll be available somewhere between 2 and 5 parts per million in the asteroid belt. Practically that means that some asteroids will be much richer in it, and will therefore be sensible things to mine.
Given the likely need for other light elements for your space program (beryllium, and especially boron) you'll be wanting a significant asteroid mining program anyway, and clearly your society has sufficient spacefaring chops to accomplish this.
