Starship On-orbit refueling - Options and Discussion

[Navier–Stokes]

https://twitter.com/SpaceX/status/1879406151508648236

I don't recall seeing this picture before.

[InterestedEngineer]

https://twitter.com/SpaceX/status/1879406151508648236

I don't recall seeing this picture before.

That tanker does not look like a Starship V3, or even V2.  Too fat.

[Navier–Stokes]

That tanker does not look like a Starship V3, or even V2.  Too fat.

I think it's the same diameter but much longer. Interestingly, there appears to be four struts linking the two vehicles together with the propellant transfer occuring at the rear of the vehicles using a QD panel-type connector.

[InterestedEngineer]

That tanker does not look like a Starship V3, or even V2.  Too fat.

I think it's the same diameter but much longer. Interestingly, there appears to be four struts linking the two vehicles together with the propellant transfer occuring at the rear of the vehicles using a QD panel-type connector.

4 struts?  Someone has been playing too much KSP

[Navier–Stokes]

It looks like both vehicles have those four connectors but of dissimilar design (one set for receiving perhaps). All of them appear to have been folded rearward during ascent.

[Overtone]

That tanker does not look like a Starship V3, or even V2.  Too fat.

I think it's the same diameter but much longer. Interestingly, there appears to be four struts linking the two vehicles together with the propellant transfer occuring at the rear of the vehicles using a QD panel-type connector.

Looks to me like a standard 9 m diameter starship, plus about half a meter of external spray-on insulation. They need that to reduce boil-off. Expect the actual vehicle to have big indents in the insulation here and there for the starlink antennas, the QD, maneuvering thrusters, etc.

[Greg Hullender]

And...  Wouldn't an orbital repair facility be called something other than a dry dock? Open the bay doors, run the ship in, seal it up and let in some air. It's an air dock.

We can do better than that! It's a space dock! :-)

[rsdavis9]

That tanker does not look like a Starship V3, or even V2.  Too fat.

I think it's the same diameter but much longer. Interestingly, there appears to be four struts linking the two vehicles together with the propellant transfer occuring at the rear of the vehicles using a QD panel-type connector.

Looks to me like a standard 9 m diameter starship, plus about half a meter of external spray-on insulation. They need that to reduce boil-off. Expect the actual vehicle to have big indents in the insulation here and there for the starlink antennas, the QD, maneuvering thrusters, etc.

So does anybody know how to spray on insulation in orbit?
Any tests? Products? etc?

It has to be different than the "great stuff" orange foam I use to seal cracks at 1 atm.

[Twark_Main]

That tanker does not look like a Starship V3, or even V2.  Too fat.

I think it's the same diameter but much longer. Interestingly, there appears to be four struts linking the two vehicles together with the propellant transfer occuring at the rear of the vehicles using a QD panel-type connector.

Looks to me like a standard 9 m diameter starship, plus about half a meter of external spray-on insulation. They need that to reduce boil-off. Expect the actual vehicle to have big indents in the insulation here and there for the starlink antennas, the QD, maneuvering thrusters, etc.

So does anybody know how to spray on insulation in orbit?
Any tests? Products? etc?

It has to be different than the "great stuff" orange foam I use to seal cracks at 1 atm.

More than you ever wanted to know about NASA's SOFI (Spray-On Foam Insulation): https://ntrs.nasa.gov/api/citations/20110014400/downloads/20110014400.pdf

Naturally this is installed on the ground, not in orbit.

[TheRadicalModerate]

Looks to me like a standard 9 m diameter starship, plus about half a meter of external spray-on insulation. They need that to reduce boil-off. Expect the actual vehicle to have big indents in the insulation here and there for the starlink antennas, the QD, maneuvering thrusters, etc.

Insulation is good for stuff sitting on the pad and experiencing aeroheating during launch, where the bulk of heat comes from convective effects, but all it does is slow down heat transport, which doesn't help as much in equilibrium.  It still helps in vacuum, because slower heat transport causes temperature to increase on the outside of the SOFI, which then increases emissions due to Stefan-Boltzmann.  But you're dealing with a fairly low temperature, so you don't get huge advantages from the T⁴ term, and SOFI's emissivity isn't great.

For lowering temperature in vacuum, the trick is to have a material with very high reflectivity in the visible and UV wavelengths, but with high emissivity in the IR.  Then the equilibrium temperature of the whole system will be lower.  SOFI can't do that, but NASA has been working on coatings that can.  (See attached for an overview.)

We also have a statement from one of the Artemis deputy program managers that implies that NASA and SpaceX are collaborating on producing a different type of tile for use with the depot.

No clue what SpaceX is going to do.  Some possibilities:

1) Nothing (i.e., just the vanilla TPS tiles and bare metal).
2) I guess SOFI should be on the list, but I don't think it helps much.
3) High-reflectance, high-emissivity coatings on bare metal.
4) High-reflectance, high-emissivity tiles.
5) Some kind of deployable sunshade made out of MLI.

[TheRadicalModerate]

4 struts?  Someone has been playing too much KSP

If they have, they've been playing it in the art department at SpaceX.  They're not known for being fanciful: it's a lot easier to render from CAD files.  It is necessarily a snapshot, and SpaceX tears up and replaces designs at a furious rate, but I'd bet it's at least something they were designing at one point in the not-too-distant past.

[Twark_Main]

Looks to me like a standard 9 m diameter starship, plus about half a meter of external spray-on insulation. They need that to reduce boil-off. Expect the actual vehicle to have big indents in the insulation here and there for the starlink antennas, the QD, maneuvering thrusters, etc.

Insulation is good for stuff sitting on the pad and experiencing aeroheating during launch, where the bulk of heat comes from convective effects, but all it does is slow down heat transport, which doesn't help as much in equilibrium.  It still helps in vacuum, because slower heat transport causes temperature to increase on the outside of the SOFI, which then increases emissions due to Stefan-Boltzmann.  But you're dealing with a fairly low temperature, so you don't get huge advantages from the T⁴ term, and SOFI's emissivity isn't great.

For lowering temperature in vacuum, the trick is to have a material with very high reflectivity in the visible and UV wavelengths, but with high emissivity in the IR.  Then the equilibrium temperature of the whole system will be lower.  SOFI can't do that, but NASA has been working on coatings that can.  (See attached for an overview.)

Since it isn't orange, we can presume they are already specc'ing SOFI covered in some sort of space-rated high-emissivity high-albedo paint.

These coatings are already widely available, even going back to the first Dragon launch:

https://phys.org/news/2013-03-white-coating-spacex-dragon-trunk.html

https://www.aztechnology.com/product/1/az-93

...
2) I guess SOFI should be on the list, but I don't think it helps much.

SOFI still does a ton, because the equilibrium temperature of the coating is still above cryogenic temperature. With no insulation layer the heat just strolls right in, but adding insulation reduces the heat ingress rate (inverse linear) and correspondingly increases the equilibrium temperature of the coating.

Remember that the first Shuttle tanks were painted white (they only removed it to save weight), so this is hardly cutting-edge tech. Nevertheless I expect SpaceX will make use of (sensible, cost-optimizing) advances in materials since then.

This SOFI + coating is probably the only thing that will work on the lunar surface, because in that environment you're getting a lot of re-radiated IR from the surface over ~half the view sphere. LEO has the same problem but not quite as bad, because the Earth's surface is cooler. So IMO there's little reason for SpaceX to gate their timelines on NASA finishing their in-progress "super-coatings" and eliminate SOFI entirely, since that's only thermodynamically workable in deep space anyway.

[TheRadicalModerate]

I am struggling with why a stock Starship-3 is not the ideal depot. Put heat shedding tiles on the backside (though I favor solar power + heat exchangers), face the idea side of the tiles towards the sun (or rotate them slowly), call it good.

Some reasons:

1) This is kinda trivial, but it's obviously not stock if the QD is gender-bent--unless SpaceX commits to completely redesigning the QD to be androgynous, which isn't out of the question.

2) You want more prop capacity for any kind of tanker, or the usable prop to LEO will be somewhat less than non-prop payload, due to mass ratio effects.

3) Once you've committed to moving the domes forward, you might as well move them as far forward as possible for both the tanker and depot, even if you can't launch the tanker completely full.  (This also has the nice property that you can fill the tanks fuller as engine performance improves, without having to change the structure.)

4) Black EDL TPS wants to maximize emissivity at very high temperatures.  TPS for a depot wants to have this weird compromise of high IR emissivity (albeit at fairly low temperatures, so the T⁴ term doesn't help as much) with high reflectivity for visible and UV wavelengths.  How you deploy such a coating is a very good question.  Seems like SpaceX and NASA are collaborating on some kind of tile, but that obviously raises a lot of questions on its own.

5) A lot depends on how much boiloff can be tolerated.  If we have a v3-sized LSS and depot both completely filled in VLEO, boiloff rates can be extremely high:  With some arm-waves about what percentage of average LEO boiloff occurs in HEEO, NRHO, and the polar surface, I can make about 13.5t/day work.  But that's probably close to 25 tankers' worth of prop to LEO.

6) There are lots of different environment swhere large propellant stores need to be maintained:

a) VLEO, in the depot, which is very hot and can't shield the tanks with attitude, since the Sun and the Earth as sometimes on opposite sides.

b) NRHO (about 70,000x3000km), while the LSS-HLS waits for Orion or whatever replaces it.  Here, attitude control largely works, because the LSS is effectively in deep space for most of the orbit, and the periapse is over the pole, where only half of the 2.8º solid angle subtended by the Moon is lit.  (For comparison, the Earth subtends a solid angle of 87º for a depot in a 300x300km VLEO.)

c) The lunar surface at the poles, which you can model as half of the LSS directly illuminated by the Sun (but not the same half all the time) plus full illumination by a Lambertian surface illuminated at ~85º from zenith, which should be about 9% of a solar constant.

I'm inclined to think that putting in at least some work in to reduce vacuum boiloff in LEO is going to be needed.

[Greg Hullender]

I'm inclined to think that putting in at least some work in to reduce vacuum boiloff in LEO is going to be needed.

I think life gets a lot easier if they just bite the bullet and invest in a cryo-cooler for the VLEO depot.

[Twark_Main]

3) Once you've committed to moving the domes forward, you might as well move them as far forward as possible for both the tanker and depot

I can't see why that would be true. Unnecessary dry mass is unnecessary.

I've noticed that this forum tends to overestimate the advantages of common hardware. SpaceX has shown us that they're not shy about building different variations.

4) Black EDL TPS wants to maximize emissivity at very high temperatures.  TPS for a depot wants to have this weird compromise of high IR emissivity (albeit at fairly low temperatures, so the T⁴ term doesn't help as much) with high reflectivity for visible and UV wavelengths.

Again, this isn't weird or unusual at all. See my link above: 85% reflectivity combined with 91% emissivity, and that's with off-the-shelf coatings that have been around since Apollo. I have no doubt SpaceX can improve those numbers if they really try.

Incidentally, you don't actually need to make the surface white. You can even make it black (which the world's astronomers would appreciate), but to achieve the same performance you have to make the insulation thicker.

[Greg Hullender]

I could imagine that the depot has actual Solar White tiles while the tanker only has Solar White paint. The paint would burn off during reentry, of course, but it might make a material difference in VLEO.

[TheRadicalModerate]

3) Once you've committed to moving the domes forward, you might as well move them as far forward as possible for both the tanker and depot

I can't see why that would be true. Unnecessary dry mass is unnecessary.

Moving the domes doesn't increase dry mass.  Same number of ring segments, just arranged differently.

Wait, I take it back:  you have to add another meter or two to the LCH4 downcomer.  That's... what?  another 400kg?  Same thing for hot-gas piping to the LCH4 ullage space, but that's very light.

I've noticed that this forum tends to overestimate the advantages of common hardware. SpaceX has shown us that they're not shy about building different variations.

I think tankers and depots are different from vanilla, and different from the LSS.  But I don't see a good reason why they're different from each other.

As for the cost of variants, rearranging domes probably implies new variants of the ring assemblies.  Those are decent-sized assemblies to manage in inventory.  So once you've committed to managing them, it's nice if you only have to manage the one variant, rather than two.

4) Black EDL TPS wants to maximize emissivity at very high temperatures.  TPS for a depot wants to have this weird compromise of high IR emissivity (albeit at fairly low temperatures, so the T⁴ term doesn't help as much) with high reflectivity for visible and UV wavelengths.

Again, this isn't weird or unusual at all. See my link above: 85% reflectivity combined with 91% emissivity, and that's with off-the-shelf coatings that have been around since Apollo. I have no doubt SpaceX can improve those numbers if they really try.

Incidentally, you don't actually need to make the surface white. You can even make it black (which the world's astronomers would appreciate), but to achieve the same performance you have to make the insulation thicker.

I find everything to do with emissivity/reflectivity/absorptance weird and unusual, especially when you're playing different games with different parts of the spectrum.  But that's just me.

As for insulation thickness:  once the heat is past the emitting layer(s), it doesn't matter how thick the insulation is, because the heat will eventually conduct into the tank.  The time scale we're dealing with is multiple months, so everything will be in equilibrium one way or another.

[Twark_Main]

3) Once you've committed to moving the domes forward, you might as well move them as far forward as possible for both the tanker and depot

I can't see why that would be true. Unnecessary dry mass is unnecessary.

Moving the domes doesn't increase dry mass.  Same number of ring segments, just arranged differently.

Wait, I take it back:  you have to add another meter or two to the LCH4 downcomer.  That's... what?  another 400kg?  Same thing for hot-gas piping to the LCH4 ullage space, but that's very light.

This assumes the total vehicle length must be the same, but of course that's not necessarily true. On a dedicated tanker the payload bay is entirely vestigial. As Musk has said, the mass-optimized tanker would look "weird."

At first, the tanker will just be a ship with no payload. Down the road, we will build a dedicated tanker that will have an extremely high full to empty mass ratio (warning: it will look kinda weird).

(yes the quote is older, but fundamentally none of the logic has changed)

I've noticed that this forum tends to overestimate the advantages of common hardware. SpaceX has shown us that they're not shy about building different variations.

I think tankers and depots are different from vanilla, and different from the LSS.  But I don't see a good reason why they're different from each other.

As for the cost of variants, rearranging domes probably implies new variants of the ring assemblies.  Those are decent-sized assemblies to manage in inventory. So once you've committed to managing them, it's nice if you only have to manage the one variant, rather than two.

Yep, that's what I'm saying tends to get over-weighted.

Currently everything is bespoke, and honestly it will be that way for a long time. The paper labels read "SN33 common dome," not "Type 1 common dome."

4) Black EDL TPS wants to maximize emissivity at very high temperatures.  TPS for a depot wants to have this weird compromise of high IR emissivity (albeit at fairly low temperatures, so the T⁴ term doesn't help as much) with high reflectivity for visible and UV wavelengths.

Again, this isn't weird or unusual at all. See my link above: 85% reflectivity combined with 91% emissivity, and that's with off-the-shelf coatings that have been around since Apollo. I have no doubt SpaceX can improve those numbers if they really try.

Incidentally, you don't actually need to make the surface white. You can even make it black (which the world's astronomers would appreciate), but to achieve the same performance you have to make the insulation thicker.

I find everything to do with emissivity/reflectivity/absorptance weird and unusual, especially when you're playing different games with different parts of the spectrum.  But that's just me.

Basically there's an absorption (1.0) vs reflection (0.0) curve that's plotted over wavelength. This can also change with temperature, but let's ignore that for now. When incoming light (I'm calling all EM radiation "light" here) hits the surface you multiply the spectral curve of the incoming light by that absorption curve, and that tells you how much light energy is absorbed. Simple enough.

At the same time, the surface is emitting light, and that emission spectrum is a black body curve for the surface temperature multiplied by the absorption curve.

To make the "ideal" selective surface in a general way, you plot both the incoming radiation and the outgoing emission on the same spectral power plot. In the parts of the spectrum where the emission is brighter you make the surface totally black. In the parts of the spectrum where the incoming light is brighter you make it totally white.

In practice, when your energy source is the Sun (T=5800 K, scaled down by a factor of 46240 because of the Sun's solid angle) and your surface is near room temperature (T=300 K), your spectral curves overlap at around 3 micrometers. So at every wavelength longer than 3 micrometers (far IR) you want it to be black, and at every wavelength shorter than 3 micrometers (near IR, visible, ultraviolet) you want it to be white.

To design an ideal absorber it's the same procedure, you just swap 'black' and 'white.'

https://www.nrel.gov/docs/fy02osti/31267.pdf

https://www.researchgate.net/figure/a-Cutoff-wavelength-for-an-ideal-absorber-coating-as-a-function-of-optimization_fig1_40851024

https://pubs.aip.org/avs/jvst/article-pdf/12/5/1010/11428421/1010_1_online.pdf

As for insulation thickness:  once the heat is past the emitting layer(s), it doesn't matter how thick the insulation is, because the heat will eventually conduct into the tank.  The time scale we're dealing with is multiple months, so everything will be in equilibrium one way or another.

Sorry, but this is not right. In practice you run the calculation iteratively, homing in on the correct exterior temperature.

You're seeking to make sure the R-value equation and the net radiant flux (absorption minus emission, as calculated above) give the same heat flux, because at equilibrium they must.

The result is that at equilibrium the outer surface will be hotter (and therefore outgoing emission greater, and therefore net heat flux lower) the thicker you make the insulation. In other words, thicker is better.

[Overtone]

This has been a great discussion about absorbance, emissions, and insulation for depots. Thank you everyone for sharing your insights.

[StraumliBlight]

https://arstechnica.com/space/2025/01/heres-what-nasa-would-like-to-see-spacex-accomplish-with-starship-this-year/

Ars: Which version of the ship is required for the propellant transfer demonstration, and what new features are on that version to enable this test?

Lisa Watson-Morgan: We're looking forward to the Version 3, which is what's coming up later on, sometime in ’25, in the near term, because that's what we need for propellant transfer and the cryo fluid work that is also important to us... There are different systems in the V3 set that will help us with cryo fluid management. Obviously, with those, we have to have the couplers and the quick-disconnects in order for the two systems to have the right guidance, navigation, trajectory, all the control systems needed to hold their station-keeping in order to dock with each other, and then perform the fluid transfer. So all the fluid lines and all that's associated with that, those systems, which we have seen in tests and held pieces of when we've been working with them at their site, we'll get to see those actually in action on orbit.