As I stated in the answer to another question, LH2 suffers some serious drawbacks compared to other fuels.

Off the top of my head, they are

  • Extremely low density, resulting in:

    • Lower mass-fraction because of high tank mass
    • High aerodynamic drag due to tank volume
    • Larger and heavier vehicle structure due to tank volume
  • Boiloff and tank seepage, resulting in:

    • Even higher tank mass because of the need for insulation
    • More complicated tank arrangements because of the inability to butt LH2 tanks against relatively warm cryopropellants like LOX.
    • Long-term storage issues from propellant boiloff

The only possible advantages I can think of are

  • Marginally higher specific impulse
  • Producing only water as a byproduct, keeping environmentalists happy.

In light of all this, what rationale is there for a LH2/LOX Rocket?

  • $\begingroup$ LH2 is extremely cold, a lot of materials become very brittle at those temperatures. Construction of valves and hoses are difficult. LH2 can't build deposits in the cooling channels of the rocket engine, it can't polymerize or even carbonize like kerosene. $\endgroup$
    – Uwe
    Commented Oct 21, 2016 at 11:35
  • $\begingroup$ Con: GH2 can embrittle metals causing them to lose strength over time. Pro: H2/O2 combustion is a very quick and simple chemical process (very much unlike hydrocarbon combustion). It is usually very stable. Pro: The specific impulse is not marginally higher, it's a lot higher, and in the end that's what counts. $\endgroup$ Commented Jan 20, 2017 at 12:49

2 Answers 2


It's not a "marginally higher" specific impulse. High performance hydrogen engines typically have a vacuum Isp of around 420-450 seconds, compared to 310-350 for hypergolics or kerosene. That's about 30% delta v advantage, ton for ton, which more than offsets the structural volume penalty.

The drag penalty is mostly irrelevant for upper stages as long as their diameter is no larger than lower stages.

Another drawback to hydrogen is that low propellant density yields lower thrust from a similar chamber size; again not a big problem for upper stages.

The clean exhaust is nice, but not a huge advantage over kerosene. It is much more attractive than toxic hypergolics, of course.

Consider two 22-ton upper stages each pushing a 5 ton payload.

  • Stage H is 20 tons hydrogen-LOX, 2 tons dry mass, 450s Isp.
  • Stage K is 21 tons kerosene-LOX, 1 ton dry mass, 350s Isp.

H delivers 5956 m/s of delta-v versus K's 5162 m/s -- a 15% improvement despite hauling twice the dry structural mass.

  • 2
    $\begingroup$ It depends on the timeframe under consideration. Hydrogen boil off rate for a stage like Centaur is about 4% per day, so it's not a major issue for typical LEO/GEO insertion operations, but a definite concern for lunar and a showstopper for interplanetary operations, which is why you see hypergolics used for orbital insertion on those kinds of missions. LOX boil off is about 2% per day, so kerolox has similar limitations. $\endgroup$ Commented Jul 29, 2016 at 15:53
  • 12
    $\begingroup$ Saturn/Apollo is a terrific example of the tradeoffs in fuel selection. Kerosene first stage to meet the colossal liftoff thrust requirement and keep the stage diameter somewhat sane; hydrogen second and third stages to make the delta-v needed for orbital and translunar injection; hypergolics on the CSM and LM to store fuel for a two-week mission and start reliably, repeatedly, over many short burns. $\endgroup$ Commented Jul 29, 2016 at 16:01
  • 1
    $\begingroup$ Does hydrolox = LOX/LH2? $\endgroup$
    – uhoh
    Commented Aug 3, 2016 at 12:02
  • 3
    $\begingroup$ Yes. Kerolox = kerosene/LOX, methalox = methane/LOX. In casual rocket-engine-discussion use, these are synonymous with the name of the fuel by itself (because other oxidizers are rarely used with those fuels). $\endgroup$ Commented Aug 3, 2016 at 14:41
  • 1
    $\begingroup$ @UIDAlexD: Thermal nuclear rockets may work on pretty much damn everything as long as its boiling temperature is less than the chamber melting temperature. They are way more efficient on hydrogen than anything else, but they won't suffer too badly if other, not too heavy elements are introduced into the mix. $\endgroup$
    – SF.
    Commented Jan 19, 2017 at 12:54

Another advantage (relating to efficiency) is that it's easier to build a staged combustion rocket engine using LOX/LH2. Specifically, you can use LH2 for a fuel-rich staged combustion engine, like the Space Shuttle main engines (and the Energia main engines). Most other common rocket fuels don't work for fuel-rich combustion, I believe due to the risk of coking. Oxidizer-rich cycles don't have that problem, but require very advanced metallurgy to safely handle the (extremely corrosive) hot oxidizer. Russia / the USSR figured out how to do that in the 60s, but the USA never built a working oxygen-rich (or full-flow) staged combustion engine until the 2000s.

In the end, efficiency is efficiency. Even if, hypothetically, some new propellants could get 600s ISP with a good TWR but could only be used in a gas generator cycle (or something similarly wasteful), it would still be more efficient than the best staged combustion engines available today. With the currently available propellants, though, the fact that LH2 makes it easier to build higher-efficiency rockets (not just because of the inherent suitability of the propellants for high ISP) is - or at least was - significant. (It still is for rockets that need maximum efficiency, but some focus on other desiderata now, like storage, transportation, cost, in-situ generation off earth, etc.)

  • 2
    $\begingroup$ fuel-rich combustion with methane seems like it has some advantages over the alternatives. $\endgroup$
    – user8269
    Commented Jun 17, 2018 at 18:13
  • $\begingroup$ What propellants are "new" and could get more than 500 s of ISP? The only thing that I'm familiar with getting anywhere near that high are ultra-dangerous (often fluorine) propellants that were explored in the 60s but never actually used. $\endgroup$
    – ikrase
    Commented Jun 4, 2022 at 7:58
  • $\begingroup$ @ikrase I'm not aware of any; that was a hypothetical example of why overall efficiency is all that matters and efficiency from any particular source only matters to the extent that it modifies the final number. Editing to make that clearer... $\endgroup$
    – CBHacking
    Commented Jun 4, 2022 at 8:16

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.