This could be done by leading tapped-off exhaust gases through heat-exchangers in the tanks (being careful that the exhaust gases don't freeze and block the pipe). I'm imagining this as an alternative to carrying separate pressurization gases in pressure-fed engines.


I will clarify a bit. I'm aware both of expander cycles, and of leading heated propellants back into the tank. This is similar, but I'm interested in the specific case of leading exhaust gases through fuel tanks. I couldn't find any previous description of this, so that's why I'm asking.

Being more specific, there are at least two options:

  1. In a first variant, some amount of exhaust gas is tapped from the combustion chamber and driven through heat exchangers in the propellant tanks. The cooled gas could be injected along the nozzle walls to protect them (ala F-1), or otherwise expelled.
  2. In a second variant, one could have the combustion chamber at the top of the rocket, and lead the exhaust gas through the propellant tanks, before expelling it through a diverging nozzle at the bottom.

I expect that there is there some reason for a considerable loss of energy here compared to other (turbomachinery-free) engine cycles, but I am too inexperienced to realize where. Hopefully someone thinks this is interesting and can offer an educational take on it.

EDIT 2: Tightened the wording in the heading.

  • $\begingroup$ This is already done, for example in the Space Shuttle Main Engine. Hydrogen is tapped from the cooling circuit and fed back to the tank. Oxygen is tapped off, run through a heat exchanger, and fed back to the tank. $\endgroup$ Apr 13, 2016 at 20:15
  • $\begingroup$ @OrganicMarble That's not really what I meant. See my clarifications! $\endgroup$
    – spookysys
    Apr 13, 2016 at 22:19
  • 1
    $\begingroup$ It makes far more sense to draw a bit of liquid oxygen from existing oxidizer line, heat it up to -100C and inject the vapour back into the tank, than to route 1000C combustion lines all over the place, especially as a liquid absorbs so much energy when boiled. That's why they do it that way in the space shuttle. The space shuttle (or any large rocket vehicle) also needs turbopumps rather than a pressure fed system because combustion chamber pressure is very high, and if you tried to pressurize the tanks to that pressure they would burst. Strengthening tanks would cost too much extra weight. $\endgroup$ Apr 13, 2016 at 22:43
  • $\begingroup$ The emphasis here is on low complexity. I think it makes more sense to compare against basic pressure feed engines than the space shuttle main engines. To use heated propellant for pressurization would need a turbopump (as well as regenerative cooling), as far as I can tell. With regards to the lines for the hot exhaust gas, keep in mind these would go through the cold propellant which would cool the wall of the tube. $\endgroup$
    – spookysys
    Apr 14, 2016 at 17:52
  • $\begingroup$ Regarding tank vs chamber pressure, this is one of the things that makes more sense when you compare against pressure fed engines than something like the SSME: In a pressure-feed engine, the tanks are at a higher pressure than the combustion chamber. Finally, with regards to energy spent on vaporizing the liquid propellants before use. The SSME does this as well, it shouldn't make a difference energy-wise whether you do it in the tank or just before entering the chamber. $\endgroup$
    – spookysys
    Apr 14, 2016 at 18:02

2 Answers 2


The first stage (Emeraude) of the 1960s French launch vehicle Diamant used a cycle somewhat similar to what you're describing. It had a gas generator that burnt a powdered solid fuel. Its exhaust gases were cooled with water and injected into the first stage propellant tanks (the stage ran on nitric acid and turpentine oil).

Another source, in French though.

  • $\begingroup$ The PDF link is broken. $\endgroup$
    – Vikki
    Jun 25, 2020 at 20:28
  • 2
    $\begingroup$ Flight Global has taken their archive offline after the company was bought. The new owner said the archive would come back eventually, but that it would take months. $\endgroup$
    – Hobbes
    Jun 26, 2020 at 6:22
  • 2
    $\begingroup$ For the time being I'm going to leave the FlightGlobal link in, in the hope they'll reinstate the archive. Added another source. $\endgroup$
    – Hobbes
    Jun 27, 2020 at 14:04

I understand your idea and it sounds elegant at first, but it requires a pressure fed cryogenic rocket. All the pressure fed rocket engines I know are hypergolic or inert gas (non-cryogenic) rockets. They are pressure fed because they are auxiliary methods of providing thrust, in post-launch systems such as RCS, the Apollo Service module (lunar departure burn) or the Lunar Module or the Space Shuttle Orbiter OMS.

Boiling hypergolic rocket propellants is probably not a good idea.

Another issue that comes to mind is if you want to put the heat exchangers in the propellant tanks, where would you locate them? In the top, middle, or bottom of the tank? The bottom would be the only place where the heat exchanger would have access to liquid propellant for the entire duration of the burn, but if you had a simple coil of pipe carrying exhaust gas surrounded by the entire mass of the propellant, I worry about the gas generation rate.

I think the current schemes, where the heat exchanger is close to the engine and taps off the propellant flow that feeds the engine, with a pipe carrying the heated (but not hot) gas back to the top of the propellant flow, is simpler and more effective.


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.