LH2, Love It or Hate It?

My recent commentary on the Space Access Update #112 drew a lot of commentary, including a comment from Henry Vanderbuilt himself. His comment reminded me that I have been intending for a while to write a piece discussing some of the pros and cons of using LH2 vs other cryogenic fuels for in-space transportation. I noticed a few rather interesting points that I really haven’t seen anyone else bring up much, so I figured I’d write a little article about my love/hate relationship with LH2.

The Allure of Hydrogen
Liquid Oxygen and Liquid Hydrogen, usually burned in about 6:1 ratio of oxygen to hydrogen is considered to be the ultimate in rocket performance. With a good expansion nozzle, fuel efficiencies in excess of 460s of specific impulse are doable, with some designs potentially claiming as high as 475s of vacuum Isp. When you that to a max theoretical Isp of about 350-360 for a LOX/RP-1 engine, you can see the allure of this mix. NASA in particular has been very fond of this mixture. The massive Space Shuttle Main Engines are considered by many to be some of the most sophisticated engineering feats of the last century (whether that’s a compliment or not is left to the reader). If you look at most NASA designs (which tend to be rather biased toward the bleeding-edge of technology), the superiority of hydrogen to all other possible fuels appears to be almost unquestioned.

Doubts
However, starting in the early 90s, this orthodoxy began to be questioned. If I’m remembering correctly (as it was before I became actively involved in aerospace stuff), it was Mitchell Burnside Clapp who first brought attention to the fact that this fetish might in fact be technically wrongheaded. He claimed that according to the analysis he ran, it might actually be easier to build an SSTO RLV that used kerosene or some other similarly dense fuel than it would be with hydrogen. Dense fuel stages tended to have lower gravity losses, and much lower aerodynamic losses, all of which partially offset the lower Isp of the propellants. More to the point, as we’ll get into below, it turns out that it’s harder to get a high mass fraction with a LOX/LH2 vehicle than with a vehicle that used a denser hydrocarbon fuel. [Ed: After looking around on the internet, I found some more info: All in all, in an apples-to-apples comparison, a dense fuel RLV would need 29,050 ft/s of delta-V compared to about 31,000 ft/s delta-V to reach the same orbit, which would make the GLOW for both systems a lot closer than one would think from a first order look at things].

Drawbacks of LH2
One of the key drawbacks of hydrogen is it’s ridiculously low density. Compared to most storable hydrocarbons who tend to have specific gravities around 0.7-0.8, hydrogen’s specific gravity is a measly 0.07! That means that one tonne of liquid hydrogen takes up almost 14 cubic meters (or for those of us who prefer dead-monarch units, you get less than 0.5lb of the stuff per gallon). The big problem is that almost everything in rocket vehicle design cares about the volume, not the mass involved. Tanks mass scales almost linearly with volume. Pumps pump volume, not mass. Feedlines have to be sized for the volumetric flow rate of the fluid. As Henry brings up in his comment:

By my hasty back-of-the-envelope numbers, the ET LOX tank masses less than 1% of the LOX it carries, the ET LH2 tank masses greater than 12% of its LH2 content.

Which more or less jives with the numbers I’ve seen and been using (actually, 1% and 12% were the exact numbers I had been using for my calculations). Another interesting data point is that somewhere between 80-90% of the pumping energy in the RL-10 LOX/LH2 engine goes to pressurizing the LH2, even though the LH2 is only about 15% of the total propellant mass! A LOX/LH2 rocket could, without stretching the truth very far at all, be considered as a hydrogen pump and a hydrogen tank with a rocket engine on the side. Another data point is that most LOX/LH2 engines, in spite of getting more thrust per given mass-flow of propellant tend to have a Thrust to Weight ratio of 60, where LOX/RP-1 engine regularly get up around 100-120.

There’s another annoying problem with LH2–the stuff is so darn cold. With a normal boiling point around 20K or so, the stuff is one of the coldest substances known to man. Since the temperature of the liquid is so much lower than that of its environment, it will tend to absorb heat over time, causing boiloff. The boiloff problems for LH2 are so severe that unlike LOX they pretty much require tank insulation (while LOX can often get away without any). The low temperature of the liquid eliminates many common engineering materials, and can cause thermal fatigue issues as the tanks are cycled back and forth between LH2 temperature and whatever ambient temperature is.

Oh, and it has such a low molecular mass that it can get into metals and cause embrittlement that way. Oh, and it makes sealing tougher. Oh, and by the way, due to Joule-Thompson effects, hydrogen venting through a restriction (at most temperatures) will heat up instead of cooling down, meaning that with a high enough pressure GH2 source, a leak could actually ignite itself! Oh, and it burns with a nearly invisible flame that is several thousand K…

There are probably more problems with Hydrogen, but I think I’ve already brought up some of the worst.

So What are the Alternatives?
Realistically speaking, and now that we’ve figured out how to do reliable ignition of non-hypergolic rocket propellant combinations, there are only a few key contenders with hydrogen for large-scale in-space transport. Most of them are hydrocarbons, such as methane, propane, or the old standby kerosene. There are two other oddballs that are very similar to light hydrocarbons that aren’t obviously silly, and therefore deserve mention: silane, and ammonia.

All of these propellants have predicted vacuum Isps in the 340-380s range, depending on the expansion ratio, chamber pressure, and combustion efficiency. All of them have bulk propellant densities much better than LOX/LH2. Ranging from a bulk density of about 1.03 for LOK/RP-1, down to 0.83 or so for LOX/Methane, as compared to 0.33 or so for LOX/LH2. That means you can get somewhere near 2.5-3x as much propellant into the same volume when compared to LH2. This is important for two things: drylaunch, and tank mass.

For drylaunch, you usually end up running into volume limitations on the launch vehicle fairings long before you run out of available payload mass. For example, the Atlas V, 4.5m PLF has about 180 cubic meters of space in its cylindrical section. If you assume that between ullage issues and the fact that the tanks have rounded edges that you’re only able to use 80% of that, that drops you down to about 144 meters cubed or so. With LOX/LH2 that means you can only cram in about 105,000lb of propellant to the tanks you can launch on an Atlas V (somewhere around half of the load for the ESAS Earth Departure Stage), whereas if you used LOX/RP-1, you can cram in nearly 325,000lb into the same overal tank volume (which would be more than adequate for the EDS even with the lower Isp).

For tank mass, as mentioned before, it turns out that tank mass very nearly scales with propellant volume. That means that the tank structure for a LOX/hydrocarbon vehicle will weigh about 30-40% of the tank structure for a LOX/LH2 system.

Another important thing is boiloff. Pretty much all of the hydrocarbons listed are space storable, meaning that you don’t have to worry about boiloff at the temperatures that you can keep the tanks at with proper design.

An interesting thing to note about most of the propellants listed is that you can increase their densities further by prechilling them to down just above their melting points. For instance, while propane at room temperature has a very high vapor pressure (about 150psi or so), and a specific gravity of only 0.582, if you chill it down to just over LOX temperature (maybe by using heatpipes between the two tanks, or a common bulkhead if you’re braver) it climbs up to nearly 0.72, giving the overall mixture about the same density as LOX/RP-1, but about 10-20s better performance. [Ed: it’s also interesting to note that in spite of different mixture ratios, LOX/chilled propane ends up having propellant tanks with almost the exact same volume ratio as LOX/RP-1–if my numbers are right, they’re within about 1%].

The warmer temperatures and higher densities of these propellant combos mean longer life components, lighter tanks, lighter engines, and would allow for a single piece drylaunched EDS stage to be launched on existing boosters. Not to mention cheaper to design, easier to handle, etc.

Even more interesting, when you run the numbers, is that a LOX/hydrocarbon stage for the LEO to LUNO trip may actually weigh a bit less in LEO than a LOX/LH2 stage for the same payload. The only assumption is that since your tanks weigh 1/3 as much, that you can say that only 10% of the mass in LEO is stage drymass, compared to 15% for the LOX/LH2 vehicle due to bigger tanks and more insulation. Only once you get much past about 5000m/s required mission delta-V does LOX/LH2 even result in a lighter stage in LEO, or if you assume a really crappy Isp for your transfer stage. [Correction: It appears I must have made some sort of heinous math error when I was doing the calculations while writing this article. Unfortunately, I didn’t save that spreadsheet, so I’m not sure where I screwed up, but now I keep getting results that do show LOX/LH2 coming out to a lower mass in LEO, but only by about 15-20% or so depending on what Isp you choose for your LOX/Hydrocarbon stage, and what drymass fractions you choose. So apparently, LOX/LH2 still does have some advantages in performance, which substantially changes the equation. Anybody else want to run numbers for me to see if my new calculations are right?]

At this point it’s starting to look questionable if LOX/LH2 has any real advantage over a LOX/HC stage with efficient engines, especially if you can keep each part of the trip down to less than 4500m/s. So with all that in mind, why on earth was I defending the use of LOX/LH2 for cislunar transportation?

LH2: What’s there to Love?
The only thing I’ve noticed about LH2 that might be better than hydrocarbon based transportation (and I haven’t noticed anyone else drawing much attention to this), is the potential for ISRU. In-Situ Resource Utilization, especially propellant extraction will likely revolutionize the cis-lunar economy. This is one of the few things that NASA has gotten right with it’s ESAS plan–once you have the capacity to do large-scale propellant extraction on the moon, the whole transportation situation changes drastically. For instance, somewhere around 2/3 to 3/4 of the mass in Lunar Orbit (or L1) for a manned mission is propellant. Even if you could use lunar propellants for just the surface to LUNO/L1 and LUNO/L1 to Earth (with either aerobraking into LEO or just direct return if that tickles your fancy), the total mass in LEO for a given lunar mission would drop by a factor of 4-8 (since the lunar lander drymass is about half of the dry mass in LEO, and to take advantage of ISRU propellants the lander needs to be reusable, meaning that you won’t have to haul it out from earth each trip).

There’s one big problem. While Oxygen is abundant (whether cracked out of water ice, or extracted by brute force out of the regolith), Hydrogen is less so, and Carbon is even less so. Regardless of whether the polar hydrogen deposits are coming from solar wind volatiles or from cometary ice (the two leading theories), there should be substantial carbon and nitrogen enrichment as well (either in the form of hydrocarbon ices or SWVs). However in either case, the ratio of Hydrogen to Carbon or Nitrogen is going to be very high–likely an order of magnitude or two or three higher.

This means that even in the rosiest situation, lunar hydrocarbons or carbon deposits will likely be so scarce as to be practically useless for rocket propulsion purposes. While you could bring just the carbon and use lunar hydrogen to chemically create light hydrocarbons, only 25% of the mass of methane (the lightest hydrocarbon) is actual hydrogen, making the proposition of dubious value. Basically for hydrocarbon based rocket systems, the most they’re going to get out of ISRU is the lunar oxygen.

And that is the second problem. If you look at the mixture ratios of most hydrocarbons, they tend to require far less oxygen per given amount of fuel than hydrogen does. For LOX/LH2, the ratio is usually 6:1, whereas for LOX/Methane it is only 3.4:1, 3.1:1 for LOX/propane, and only 2.7:1 for LOX/RP-1. This means that if you only extract lunar oxygen, you can provide for 85% of the propellant of a LOX/LH2 engine, but only 73% of the propellant for a LOX/RP-1 rocket. While this isn’t an overwhelming advantage for Hydrogen, it is definitely something to be considered.

Ramifications?
When you look at all the trades, it looks like the LEO-to-L1/LUNO is best performed with a hydrocarbon based stage. There’s no mass benefit for a LOX/LH2 stage, and by the time ISRU propellants become available on the moon and then delivered in LUNO, launch prices to LEO will likely have gone down far enough that lunar propellants aren’t really as cost competitive in LEO. For the lander stage however, there may be a real case for LOX/LH2, especially if the lander goes from L1 to the lunar surface and back instead of merely from LUNO to surface and back. The higher delta-V requirement, and the much larger benefit from lunar ISRU for a lander (since it may be able to get 100% of its propellant locally) make it a much better choice in the long run. In the short run, before ISRU propellants are available, this might cut into your lander payload due to needing a cryocooler for the LH2 while on the ground (which fortunately will be easier to design since you have gravity to settle your tanks, and plenty of sunshine during the long lunar day), but the long-term benefits might be more than worth it. Ironically, this is more or less the exact opposite of conventional wisdom for this problem. [Ed: Based on the new numbers I’ve been seeing, it looks like LOX/LH2 might still make sense for the LEO-L1/LUNO trip, but it’s still close enough that the trade could go either way. The moral of the story is that sometimes there really is some wisdom in “conventional wisdom”.]

Thoughts, comments, flames?

The following two tabs change content below.
Jonathan Goff

Jonathan Goff

President/CEO at Altius Space Machines
Jonathan Goff is a space technologist, inventor, and serial space entrepreneur who created the Selenian Boondocks blog. Jon was a co-founder of Masten Space Systems, and is the founder and CEO of Altius Space Machines, a space robotics startup in Broomfield, CO. His family includes his wife, Tiffany, and five boys: Jarom (deceased), Jonathan, James, Peter, and Andrew. Jon has a BS in Manufacturing Engineering (1999) and an MS in Mechanical Engineering (2007) from Brigham Young University, and served an LDS proselytizing mission in Olongapo, Philippines from 2000-2002.
This entry was posted in Uncategorized. Bookmark the permalink.

15 Responses to LH2, Love It or Hate It?

  1. murphydyne says:

    Nice post Jon.

    I do have a question, though, and please excuse the brutal ignorance of it, but why can’t we store the hydrogen in the form of water and crack it when we need it?

    It just seems like such a handy currency, given it’s role in fuel cells, human use, etc.

    Sure, more fuel cells or a nuke plant would be needed. And? We can’t launch more fuel cells?

    Just curious.
    Ken

  2. murphydyne says:

    Uh, that would be solar cells.

    (It was a long day doing public space outreach at the IMAX theatre)

  3. Jon Goff says:

    Ken,
    Thanks. Storing LOX/LH2 in the form of water sounds like a great idea, but the devil is in the details. First off, as mentioned H2O has two hydrogens for every oxygen. Since Oxygen has an amu of 16g/mole, and Hydrogen about 1g/mole, that means that mass wise you have about 8:1 Oxygen to hydrogen ratio. However in a rocket, you usually run rich at 6:1, which means that you couldn’t store it all as water (you’d have quite a bit of left over hydrogen).

    More importantly, electrolysis takes a ton of energy, which with the size of solar cell you could reasonably bring with a given mission would mean taking days or weeks to crack all the water back to LOX/LH2. Once there’s a base, it might make sense to store some hydrogen tied up in water, and some in cryo form, especially if you’re using regenerative fuel cells, but for early missions it probably is a nonstarter. In an emergency, before you have a settlement, you’d really like to be able to head up from the surface ASAP instead of having to wait a day or two while the tanks fill themselves.
    ~Jon

  4. Ed says:

    Fuel extracted from the moon wouldn’t need to be a hydrocarbon. Aluminum can be burned as rocket fuel.

  5. Anonymous says:

    I have seen references to very lean operating rockets. Perhaps a 14/1 mixture ratio of LLOX with LH2 would leverage resources better.

    Redneck

  6. murphydyne says:

    Thanks for the answer Jon. I wasn’t thinking mission, though, I was thinking a permanent fuel depot. That’s why I’m less concerned about the overall mass of solar cells. They can be added over time, and water is a nice, non-volatile substance (meaning it should be less of a problem to keep near the space station) that has a myriad of other uses as well. Oh, I forgot radiation shielding in my earlier list.

    Plus, why couldn’t the richness be supplemented by LUNOX?

  7. Jon Goff says:

    Ed,
    Yeah, you can in theory use aluminum, however there are two big obstacles. First off, aluminum is a royal pain to extract from the lunar regolith (not impossible, but probably won’t happen very early on). More importantly finding a way to burn it with LOX in such a way that is actually reusable, reliable, safe, and in expensive is almost impossible. The devil is definitely in the details there. It’s a cute idea, but when all is said and done, you’d be better off using hydrogen or hauling your own fuel up and down than using aluminum.
    ~Jon

  8. Jon Goff says:

    John,
    Regarding 14:1 O/F ratios for a lunar LOX/LH2 engine. It’s probably possible, but a bit scary. Running lean risks igniting your engine chamber or doing other scary things, and with a higher O:F ratio, you more or less would have to cool part of the chamber with LOX since you no longer have as much hydrogen available for coolant. It’s not entirely a nonstarter, and will definitely give you better T/W ratios on the engines, and better bulk densities (and hence lighter tanks). However your Isp takes a big hit, and the engine development task gets tougher…at which point it’s no longer an obvious win to go with hydrogen at all.

    In other words, it’s an interesting thought, and not inherently wrong, but probably depends a lot on the details.
    ~Jon

  9. Anonymous says:

    Has anyone looked into storing lunar oxygen as water and then electrolyzing it into hydrogen peroxide for oxidizer instead of going all the way to LOX? I would imagine that the energy cost of going from H2O to H2O2 would be less than stripping off all of the hydrogen to get O2.

    A LH2O2/LH2 rocket would be stochiometric at a ratio of 17:1. And it might be easier to run lean than LOX. H2O2 can even be used as a monopropellant which is the extreme case of leanness. So maybe a 1:1 volume ratio (20:1 mass ratio) could be used.

    It seems strange to think of using a lower performance oxidiser with the highest performance fuel, but in this specific situation (ISRU where H and O are available, but C is not) it might make sense.

  10. Kelly Starks says:

    Cool little summary. And cudos for noticing that by the time you could get a ISRU going on the moon – it could well not be cost competative with fuel shiped out from earth. Folks seem to never realize that upping the flight rates to support a base, would cut costs dramatically.

    Couple bits.

    Mitchell Burnside Clapp who first brought attention to the fact that this fetish might in fact be technically wrongheaded. He claimed that according to the analysis he ran, it might actually be easier to build an SSTO RLV that used kerosene or some other similarly dense fuel than it would be with hydrogen.

    McDonnel Douglas confirmed that during analysis for the DC-X program.

    As for the potential for ISRU with hydrogen. Has anyone checked to see how common hydrocarbons are? I mean the idea is Lunar polar water is the remnants of impacting comets. Comets also have a lot of oil in them – some comet cores look to be as rich in oil as water. Certainly its common in space.

  11. Anonymous says:

    Jon,

    BOTE for the 14/1 ratio seems to be in the 300 range for vac. The Isp hit you mentioned wipes out any even theoretical advantage to that lean mix. Maybe I can do a trade study to salvage the idea.

    Redneck

  12. Anonymous says:

    Kudos on a very interesting and informative post.

    After considering all the facts that support using hydrocarbon fuels over LH2, I’m left wondering why it is that NASA and the big three aerospace contractors insist on using hydrogen wherever possible nowadays. LOX/LH2 is a fuel combination that makes perfect sense wtih ISRU or on upper stages, where gross weight must be kept to a minimum, but what possible advantage does it give the shuttle or Delta IV over denser fuels? It doesn’t make sense to use inefficient fuel combinations, but considering the points you raise in your post I’m having trouble seeing what made hydrogen so alluring in the first place, besides the high isp. Is there another factor here that favors hydrogen first stages?

  13. Jon Goff says:

    Kelly,
    Cool little summary. And cudos for noticing that by the time you could get a ISRU going on the moon – it could well not be cost competative with fuel shiped out from earth. Folks seem to never realize that upping the flight rates to support a base, would cut costs dramatically.

    Actually, that isn’t exactly what I was saying. I was saying that you probably should expect to see LUNOX for sale in LEO. If transportation costs to LEO get where they need to, it would be hard to compete hauling it all the way from the moon. But in L1, or on the Lunar surface, LUNOX will likely be much cheaper than hauling it all the way out from the earth.

    As for the potential for ISRU with hydrogen. Has anyone checked to see how common hydrocarbons are?

    We actually don’t have enough evidence to really know. Odds are pretty high that carbon or hydrocarbons are much less plentify than the hydrogen.

    I mean the idea is Lunar polar water is the remnants of impacting comets.

    That’s one of the two prevailing theories. However, the other theory (that they’re just solar wind volatiles that are concentrated up there due to the cold traps) would have elemental carbon and elemental nitrogen instead of hydrocarbons. If there were hydrocarbons (ie if the polar hydrogen is really coming from comets), it’d probably be of the light hydrocarbon sort, not “oil”.

    But honestly I have no idea how abundant it would be compared to water ice, if the deposits are of cometary origin. Anybody else have an idea?

    ~Jon

  14. Jon Goff says:

    Anonymous,

    After considering all the facts that support using hydrocarbon fuels over LH2, I’m left wondering why it is that NASA and the big three aerospace contractors insist on using hydrogen wherever possible nowadays.

    Well, as I noticed earlier (and posted in my correction), it turns out that LH2 still has a decent performance benefit over hydrocarbons. It turns out that my original analysis must have had some sort of math error or typo in it, because now that I run the numbers again, I’m getting consistently lower mass numbers required in LEO for a given mission compared to hydrocarbon fuels. The volume is smaller for the hydrocarbons, which makes dry launch a lot more feasible, but you’ll have to do more fueling flights.

    Not an unsolvable problem by any stretch of the imagination, especially with high flight rate vehicles, but it at least shows that NASA wasn’t being entirely stupid.

    Is there another factor here that favors hydrogen first stages?

    Not really that I can see. The only places where hydrogen makes any sense is in space. And even then it’s a tradeoff, not a completely clear win (though a better deal than what I saw with my initial flawed calculations).
    ~Jon

  15. Kelly Starks says:

    Actually, that isn’t exactly what I was saying. I was saying that you probably should expect to see LUNOX for sale in LEO. If transportation costs to LEO get where they need to, it would be hard to compete hauling it all the way from the moon. But in L1, or on the Lunar surface, LUNOX will likely be much cheaper than hauling it all the way out from the earth.

    Why specifically? You can buy and launch a lot of fuel for what it takes to operate a remote base of any size?

    As for the potential for ISRU with hydrogen. Has anyone checked to see how common hydrocarbons are?

    We actually don’t have enough evidence to really know. Odds are pretty high that carbon or hydrocarbons are much less plentify than the hydrogen.

    Guess it depends on what evaporates faster.

    😉

    I mean the idea is Lunar polar water is the remnants of impacting comets.

    That’s one of the two prevailing theories. However, the other theory (that they’re just solar wind volatiles that are concentrated up there due to the cold traps) would have elemental carbon and elemental nitrogen instead of hydrocarbons. If there were hydrocarbons (ie if the polar hydrogen is really coming from comets), it’d probably be of the light hydrocarbon sort, not “oil”.

    Well, guess you could always go NEO comet core hunting for fuel. Delta-Vs lighter, and the yields higher – but ya got to like to travel.

    😉

Leave a Reply

Your email address will not be published. Required fields are marked *