I was wondering if the propellant required to leave Earth orbit, accelerate to Mars' orbital velocity and then perform orbital insertion will leave the starship with enough propellant to perform a safe landing. I am confused that I have read that Mars’ atmosphere is substantial enough not to ignore, but not substantial enough for effective aerobraking. Are there any “worst case scenario” plans for even partial refueling in Mars orbit before EDL? Refuelers in Earth-Mars Hohmann transfer orbit perhaps?

  • $\begingroup$ Mars' atmosphere is enough for aerocapture, and aero-deorbit (I don't know if there is a technical term for that), but it's not enough for aero-soft-landing, which may be what you are thinking of when you say aerobraking in your question. $\endgroup$ Commented Aug 15, 2023 at 3:58
  • 1
    $\begingroup$ The martian atmosphere will provide a substantial aerobreaking effect at high velocity and mars gravity is only 0.38g, but there will still need to be a long landing burn. I imagine that SpaceX have done the calculations in outline to show its possible but that's probably about as it goes. It's far too early to be working through the fine detail. I doubt they have seriously considered refueling in mars orbit yet. Again it's too far ahead. They will probably only look at it if there are problems with plan A. And sending a tanker to Mars would also require additional work to prevent boil off. $\endgroup$
    – Slarty
    Commented Aug 15, 2023 at 7:02
  • 1
    $\begingroup$ I've seen that confusion before... Mars has plenty of air for aerobraking (i.e. using the atmosphere to reduce the fuel cost for slowing down from solar orbit to planetary orbit) but not enough for parachutes to handle a landing on their own. $\endgroup$ Commented Aug 15, 2023 at 14:15
  • $\begingroup$ The only answer to "if the propellant required... will leave enough propellant to perform a safe landing" is "that depends on your payload". The interesting question is whether you end up with a useful final payload to the Martian surface when you run the numbers. $\endgroup$
    – Cadence
    Commented Aug 15, 2023 at 14:25
  • $\begingroup$ Note that Starship does the bulk of the work in reaching orbit. Like the Falcon 9, the booster stages relatively early, which is a big part of what makes it practical to recover and reuse it. This means that when Starship reaches orbit, it has big empty propellant tanks that give it ~7 km/s of delta-v once reloaded with propellant in orbit, well above the minimum needed to reach Mars. $\endgroup$ Commented Aug 15, 2023 at 23:49

1 Answer 1


The problem has not been that the atmosphere isn't effective for braking, but that it's very difficult to purely aerodynamically brake to subsonic speeds before encountering the surface. This matters because Mars probes have traditionally used large heat shields and parachutes to get the vehicle subsonic before starting the rocket phase of deceleration, due to inability to model how rockets would perform when braking at supersonic speeds. Parachutes are effective for small landers, but do not scale up well, and supersonic parachutes are also complicated and difficult to design (see the many issues ExoMars has had with them).

However, SpaceX already had to deal with this issue for Falcon 9 booster recovery, and every landing booster does a supersonic retro-burn during reentry. In fact, they do so at an altitude where Earth's atmosphere is similar to what would be encountered at Mars. Starship is to do a similar landing burn when it arrives at Mars, replacing the parachute stage entirely.

The exact amount of velocity Starship will be able to shed at Mars will depend on the lift, drag coefficient, and mass the Mars vehicles end up having, but they should easily be able to shed the great majority of their velocity aerodynamically, braking to less than 1 km/s of delta-v for the final braking and landing from an entry velocity of over 7 km/s. It would require far more propellant to enter orbit at Mars instead.

To address the "worst case scenario": it would be prohibitively expensive to send tankers along just to cover the scenario where a Starship has insufficient propellant to land. However:

  • Passenger ships are unlikely to be traveling alone. In an emergency, it may be practical to transfer propellant from a cargo ship in the same fleet, or to transfer passengers to another passenger ship for the landing.
  • Passengers are very low density cargoes. It is likely a passenger ship won't be anywhere near its maximum payload mass. The exact numbers will likely change, but about 70 t is required to give 1 km/s of landing delta-v with 100 t of payload, but only about 50 t is required with 40 t of payload. Meaning if you only need to land 40 t, and dedicate the other 60 t to additional landing propellant, you have the mass budget for 220% of the propellant required to land. Passenger Starships could plausibly have expanded or double-redundant header tanks.

However, it's likely that propellant leaks won't be the biggest risk on such a trip, and that the mass would better be devoted to something else. For example, water radiation shielding, or additional life support consumables. (Though an extra reserve of LOX would also be a reserve of breathing oxygen, which was one of the points in Starship's favor for the Artemis HLS.)


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.