Reading this question, instead of stacking two tanks, are there boosters where tanks are coaxial, (like a thermos bottle) having a trussed or tridimensional structure between inner and outer tank?

Basically a thick structural skin containing either fuel or oxidizer, and an inner hollow tank containing oxidizer or fuel.

If no, what are the main reasons?


4 Answers 4


An advantage to using a coaxial tank is you don't need to arrange plumbing to get the upper tanks contents to the engine.

A big problem with a coaxial tank is the large surface area between the two tanks, which means your propellants have to have compatible operational temperatures, or you need a lot of insulation.

In practice most rockets use cryogenic liquid oxygen these days, which would freeze kerosene, or would be frozen by liquid hydrogen. Propane, however, does complement LOx in such a setup. Orbex are building a booster using propane/LOx in concentric tanks

By matching the tank diameters to fuel mix, further simplification of a pressure fed engine is possible. This was used in the German Taifun surface-to-air missile and was copied in the Russian R-103 missile

There are sources to suggest the OTRAG rockets may have been designed and flown with coaxial tank configurations, with non-cryogenic diesel and white fuming nitric acid as propellants

  • $\begingroup$ Interesting that the linked article claims it makes for a lighter structure. Not sure about what you say re: the pressurant though - some non-coxial stages use a 'single pressurant' - He. $\endgroup$ Commented Aug 15, 2019 at 11:42
  • 2
    $\begingroup$ by capping the two tanks with a burstable barrier, a single pressurant source can be used (in this case cordite) , without the need for plumbing. $\endgroup$
    – user20636
    Commented Aug 15, 2019 at 13:04
  • 1
    $\begingroup$ Ah! It's not the gas, it's the delivery system. Thanks. $\endgroup$ Commented Aug 15, 2019 at 13:04
  • 1
    $\begingroup$ Why is "careful matching of tank diameters" related to pressuring? Pressure is pressure, regardless of diameter. The matching would be to have the correct areas in the two tanks, so that the fuel load would come to the same height. $\endgroup$ Commented Aug 15, 2019 at 22:13
  • $\begingroup$ @BobJacobsen I interpreted that to mean that the tanks would be sized to deliver the proper mixture ratio. But that is done in every case. $\endgroup$ Commented Aug 15, 2019 at 23:53

For large rockets, you end up with more structure, not less.

If you keep the length and radius of the tankage the same, i.e. don't change the shape of the rocket, you still need the same amount of outside skin you always had, and the same (or more complex, i.e. heavier) ends.

You've replaced a separator between the tanks of area $\pi r^2$ with a cylindrical separator of area $2 \pi r L$, along with extra structural support to allow separate filling draining, etc. Since rockets tend to be much longer than wide, $L >> r$, you end up with more structure, hence weight.

  • $\begingroup$ Ah, are you sure? One requires a certain amount of cylinder to support the payload and hold the propellant in. because the inner tank holds supports part of the payload weight, and holds it's propellant in, the wall of the outer tank can be thinner. It is claimed Orbex's design saves weight - although part of that may be insulation. $\endgroup$
    – user20636
    Commented Aug 15, 2019 at 16:43
  • $\begingroup$ Putting material at a smaller radius (I.e. in an inner wall) generates significantly less buckling strength per unit mass. This is why rockets put material in their outer walls instead of in a central column. $\endgroup$ Commented Aug 15, 2019 at 19:39
  • $\begingroup$ Moreover the pressure on the bottom of the tank increases (as they are now as long as their combined length) and the center of gravity of a half filed vehicle (e.g. mid flight) moves to the back making it harder to "pointy end up, flamey end down". $\endgroup$
    – Christoph
    Commented Aug 16, 2019 at 6:24
  • $\begingroup$ You do not require the same thickness of outside skin, and only require half the number of tank ends. For Propane LOX, inner tank is 82 % of the width of the outer tank for a stoichiometric ratio. Orbex are using concentric tanks because it gives them a significant weight saving, not as you assert a significant weight gain. $\endgroup$
    – user20636
    Commented Aug 16, 2019 at 7:10
  • $\begingroup$ It does increase the CoG shift @Christoph, but all launch vehicles experience, and handle it. Whether Orbex can do so has yet to be seen, however it's unlikely this is something they haven't simulated $\endgroup$
    – user20636
    Commented Aug 16, 2019 at 8:34

Yes, this has been done, more or less:

enter image description here

This is the ESC-A upper stage of the Ariane 5 ECA. This stage reuses the engine and oxygen tank from the Ariane 4, combining it with a new hydrogen tank. The 'Bielles inter-réservoirs' are the trusses between the two. Ariane 5 has a much larger diameter than Ariane 4.

in this Cross-section you can see the hydrogen tank is dome-shaped: enter image description here

This led to a very heavy construction, which makes this stage inefficient.

Coaxial tanks aren't popular because:

  • they are heavier than tanks in tandem (on top of each other) because they contain more structure (instead of a horizontal dome, you have a vertical cylinder)
  • they have more surface area where the two propellants share a wall. That is an issue if the two propellants have to be kept at different temperatures (like LOX/LH2 and LOX/RP-1): all that surface area needs to be insulated.
  • $\begingroup$ does that pair of tanks share a wall at all? even a common bulkhead? $\endgroup$
    – user20636
    Commented Aug 15, 2019 at 13:46
  • 1
    $\begingroup$ no, there's a gap between them. Which is what you want, to improve insulation. $\endgroup$
    – Hobbes
    Commented Aug 15, 2019 at 14:01
  • $\begingroup$ It is claimed Orbex's tanks save weight. $\endgroup$
    – user20636
    Commented Aug 15, 2019 at 16:47

If you count a ring of individual tanks of fuel around a central oxidiser tank, then yes.

This is done in the Proton first stage, due to maximum transport size restrictions. The oxidizer tank was the largest diameter that could be carried on the railway (Baikonur is a long way from the sea.) So an additional six modules comprising one engine and one fuel tank each are prefabricated and installed on site.


NB: Another cluster design was used in the Saturn I, in order to be able to use existing components. An oxidizer tank from a Jupiter rocket was surrounded by 8 tanks from redstone rockets, 4 containing oxidiser and 4 containing fuel.


other than the above, I can think of no good reason to use a concentric tank in tank design for a long, slim rocket. It increases the surface area between the two propellants, which frequently need to be kept at different temperatures.

An approximation to a concentric design has been used in very squat stages, like the Ariane upper stage mentioned in another answer. Spacex have proposed embedding small tanks of propellant for landing inside larger tanks for takeoff propellant in some BFR concepts, but that's not quite the same thing.


Your Answer

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