Atmospheric Scoop Divers
Kardashev range: 0.8–1.8
The broader idea of atmospheric scoops is to collect gas and gain momentum from an orbiting craft interacting with atmosphere. These are my very special variations on fast-pass versions.
ABEP is an idea for capturing gas in a low Earth orbit (presumed circular) and ejecting it at higher speed to gain momentum. There is punishingly little engineering margin at the speeds involved and if it does work, it's really hard to make it work as a propellant exporter for other missions.
Actively-Cooled Scoop Diver
What I am positing instead is that it might work in a non-steady-state way. That means using an elliptical orbit with a regular dive phase into atmosphere, followed by a coast phase back out of it.
That offers several advantages. One is active cooling, which can be refreshed in the coast phase. You cannot lose any coolant, so the vehicle needs to remain closed until close to perigee, at which point it can open its mouth and let the bow shock keep the new air and the coolant in place.
The upper levels of Earth's atmosphere are mostly molecular oxygen, and that kind of sucks if what you really want is nitrogen. Oxygen is still useful, and in some architectures it could even be the point, but it brings additional handling and storage challenges and does not solve the big atmosphere-fill problem for space habitats.
To get any significant fraction of nitrogen, the diver has to go much lower. That makes the thermal and drag problem substantially worse, but it is the whole reason this concept matters. If you believe in the O'Neill habitat vision, then you have to accept that large amounts of nitrogen will be needed in the cislunar system. The Moon and most asteroids are critically deficient in nitrogen. Mars would be a good source, but Earth is closer.
So the long-term role here is not just “grab some propellant.” These could be nitrogen factories for orbital colonies, with oxygen or propellant products possibly coming along as byproducts. The drawback is a low collection ratio, but if it ran continuously it could still produce net positive mass.
When I was first writing about this, I kept trying to come up with a means of docking at higher orbit. That line of thought resulted in the L1 diamond stabilizer idea. I tried many other ideas for a mothership or for "jousting," but now I think the scoop diver works best on its own, independently.
There is some angular weirdness involved in thrusting, but I do not currently think that is the real blocker.
While I am fairly sure the actively-cooled scoop diver is the winning design, I believe I have made some legitimate progress on another alternative which looks better but is worse in practice.
Aerocycler
The Earth–Moon system has free energy. So if you've ever asked if a fully passive spacecraft can extract that energy, I'm here for you. Our challenge is to avoid proposing a megastructure, which is basically my MO. Can you, instead, do it with a microsat? Because that should get us excited. And yes. It seems this is possible.
The physical mechanism we will exploit here is lift. In a nearly-polar pass by Earth, we will use the west-east motion of the gas around the equator to nudge the spacecraft to the "side" relative to the direction it is traveling. In layman's terms, the spacecraft grazes the atmosphere, and its wings lift it to the right. This is a direction specifically chosen so that a later Moon pass by will both restore the orbit to its original parameters and add energy to the orbit.
I can't articulate this well in text, or even in a picture. You have to grab and drag to understand the picture.
Explore the aerocycler orbit in 3D
This orbit looks trollishly simple. I personally spent years back-and-forth on exactly how you might be able to construct an aerocycler orbit, or even if you could! I experimented with every possible orbital resonance and backflips (no really, this is in literature). In the end, GPT Astra and Sol found a 1:1 resonance that solved it. This uses an extremely distant pass, which suits us fine. The problem of viability is shot when you require a lunar close pass closer than its radius, which won't work for obvious reasons.
This has a required L/D of 20-something, which comes from basic theory. It relates to dividing the Moon's orbital period by Earth's rotation period (the length of a month is 30-ish days). This is completely out of the question for hypersonics. However, it's not totally disproven to be possible by magnetic ion deflection. The ion population is only something like ~4%. Again, most of the stuff is the atomic oxygen which nobody wants. So you might be able to deflect that 4% of ions by a simple magnetic field at a low deflection angle, and then, well, we can't say this won't work. Not for sure.
Assuming the spacecraft design works for that L/D and you have extra margin to work with, you can collect gases, powered by the rotation of the Earth. Nonetheless, I think required collection area is probably really bad. This is likely inferior to the actively-cooled scoop diver for real industrial use. This could also be called a scoop diver, with electromagnetic wings.
What I don't know
I'm legitimately unsure if this specific form has been described in other papers. I have seen a diagram of collection using elliptical orbits, but they seemed to be incorrectly labeled or else missed something in dissemination.
Also, the basic analysis isn't at all finished. I need to make a calculator for this where the ultimate output is mass return over time.
Finally, the big unknown is materials. Some numbers I can remember looking at (possibly for other variants of this) required about 15x the amount of coolant relative to the amount of gas collected. This means you'll need about 15 passes before doubling that mass, and more for the structural mass payoff. I do think it will be less than this, but even optimistically we would be looking at a dozen passes for payoff. Getting a sense of hypersonic material use with cooling is hard, but Starship might give a good reference very soon!