Currently, SpaceX has developed the Merlin 1 family (1B (Falcon 1), 1C (Falcon 9 v1.0), 1D (Falcon 9 v1.1/F9-R/Falcon Heavy), vacuum versions and sealevel versions) which are LOX/RP1 based. (75-205Klbs thrust)

They developed (and abandoned) the Kestrel engine used on the second stage of the Falcon 1, also LOX/RP1. (6.9Klbs thrust)

They developed the Draco and SuperDraco engines using hypergolics (mono-methyl hydrazine fuel and nitrogen tetroxide oxidizer). (90 lbs and 15K lbs thrust)

The next engine on their list is the Raptor, which is expected to be Methane (CH4) and LOX based. (660Klbs thrust)

Do we know whey they chose methane over hydrogen? Performance wise, LH2 is usually the go to propellant.

  • 8
    $\begingroup$ Musk probably said "No LH2" at the initial blank-sheet design meetings. Musk/SpaceX have demonstrated a philosophy of simple, robust design. They want to minimize surprises and control costs to maintain a fast tempo for development, testing, and operations. H2 is anathema to that philosophy. It requires special materials and processes. LH2 imposed untold headaches & delays on the STS. For example, the ET insulation had to be foamed with helium; if foamed with air or N2, the foaming gas would liquefy and foam collapse. Welds that are impermeable to any other fuel leak H2. Ad nauseum. $\endgroup$
    – Kengineer
    Commented Aug 8, 2016 at 23:43
  • 1
    $\begingroup$ @Kengineer I don't think the blowing agent was helium. If you have a reference to prove otherwise, I'd be fascinated to see it. $\endgroup$ Commented Apr 6, 2018 at 21:35

10 Answers 10


Methane has the benefit of being easier to store than hydrogen. Mostly passive cooling can suffice to keep it cryogenic, whereas hydrogen needs active cooling, and will still vent over time. Which makes Methane much closer to 'storable' than hydrogen can be. This would make it useful for deep space missions, with long mission durations.

Methane is less bulky than hydrogen. Which means tankage is smaller for the same mission. (The Shuttle external tank is mostly hydrogen tanks with a small oxygen tank (at the top?)).

Methane should be simpler to use in an engine due to its higher density than hydrogen, less needs to be pumped by volume.

Methane is potentially manufacturable on Mars. With imported Hydrogen (or native water), CO2 (Carbon dioxide) can be converted to CH4 reasonably straight forwardly.

There are ideas for ISRU (In-Situ Resource Utilization) and demonstrations on Mars. (Robert Zubrin's model is launch the return vehicle that uses ISRU to fill its fuel tanks, and do not launch the manned mission until the return vehicle is fully fueled and ready to go. Then you launch the manned mission, along with a second return vehicle, that uses ISRU over the duration of the surface mission to fuel itself).

SpaceX is focused on developing re-usability technology for their rocket lines. Traditional rocket-grade kerosene produces residue (a process known as "coking") when it burns. Methane fuel burns cleaner so there is no residue build-up which means engines can be re-used more times without refurbishment.

  • 7
    $\begingroup$ What is the maximum achievable Isp for CH4/O2 comared to H2/02? $\endgroup$
    – Ingo
    Commented Apr 24, 2014 at 13:13
  • 12
    $\begingroup$ @RussellBorogove Elon Musk has stated time and again that Mars is his goal. Mars missions are deep space. He has already started development of the Raptor engine in 2013/2014. They seem very serious about Mars, and Mars sooner than later. $\endgroup$
    – geoffc
    Commented May 2, 2014 at 17:45
  • 3
    $\begingroup$ @Ingo Numbers being floated are 363s vacuum / 321s sea-level. Compare to other historic Lox/LH2 engines of about 450s vacuum / 370s sea-level. $\endgroup$
    – AlanSE
    Commented May 15, 2014 at 14:53
  • 1
    $\begingroup$ @AlanSe - Those are what they managed to achive, the goal is at least 200TWR with 381Sec vaccum and 300atm of presure $\endgroup$ Commented Oct 3, 2017 at 3:51
  • 1
    $\begingroup$ Because storing H2 is hard. Storing water is not terrible. Using H2 is much harder, since it leaks, and keeping it during an entire trip, or wait the year to refuel is harder. The tanks for H2 are much much bigger than the same tank for CH4. Many reasons, all adding ip. $\endgroup$
    – geoffc
    Commented Aug 19, 2021 at 15:23

Methane (CH4) and RP-1 are roughly equivalent in realizable performance. As previously mentioned by other posters, CH4 has slightly higher impulse – about 370 s in vacuum vs the 360 s – at the same chamber pressure of 7 MPa. But, this is counterbalanced by its lower bulk density of about 830 kg/m3 vs about 1030 kg/m3. Bulk Density is the density of the combined Fuel and Oxidizer load in their appropriate ratios. Even though Methane is "only" 430 kg/m3 it is burned with 3.5 parts oxygen compared to 2.1 parts for RP-1, hence a CH4 rocket will be carrying more oxygen and less fuel by weight. Oxygen is pretty dense at a little over 1140 kg/m3 which is denser in fact than RP-1 (about 810 kg/m3). If we assume that chamber pressures and engine cycle efficiency will be equal, RP-1 outperforms CH4 simply because a 20% larger tank will impose weigh penalties that slightly outweigh the 3% increase in specific impulse. However, the RP-1 advantage is contingent upon operating at an equal chamber pressure which may not be the case. And, Methane (CH4) has additional advantages that are applicable in specific scenarios.

The reasons CH4 is a front runner for SpaceX's Raptor can probably be attributed to four factors:

  1. Methane does not coke (polymerize) at the operating temperatures of a rocket engine – it's coking point is roughly twice as high. This makes it easier to make an engine reusable and re-usability is a key SpaceX objective.

  2. Because Methane does not coke, it is also easier to implement a full-flow stage combustion (FFSC) cycle where all the fuel and oxidizer flow goes through the pre-burner. Compared to contemporary Russian partial flow stage combustion engines higher chamber pressures are attainable resulting in a total impulse advantage of about 30 seconds, or 9%. This eliminates the performance deficiency of CH4 compared to RP-1.

  3. If SpaceX intends to use the same fuel in all the stages, CH4 can be considered a better upper stage fuel and a worse lift-off fuel, even without enabling higher working pressures. This is because upper stages are typically 1/8th to 1/10th the size of the 1st stage, and here impulse is more important than density. Using Methane with the aforementioned FFSC cycle means that SpaceX can potentially get equivalent 1st stage performance and better upper stage performance.

  4. Even though it is, IMHO, somewhat dubious that early Mars mission will use in-situ fuel production. If that ever becomes an applicable practice, Methane can be produced from water and CO2 while RP-1 cannot.

Other than that, there is the non-factor that somewhat favor Methane, such as regular grade Natural Gas being good enough and not having to highly refine the fuel from regular kerosene to RP-1 to achieve low coking characteristics and consistent densities. I say it is a non-factor, because fuel cost is such a negligible part of launch costs that it really doesn't matter if fuel cost a few times more or less. Fuel is typically only about 0.3% of the cost of flying a rocket to orbit, so fuel cost really doesn't matter – Not even when you compare highly expensive propellant combos like Hydrazine/Tetroxide to the relatively cheap Kerosene/Oxygen.

  • $\begingroup$ The density difference is even bigger if the Kerosene is sub-cooled. $\endgroup$
    – uhoh
    Commented Aug 8, 2016 at 14:25
  • 1
    $\begingroup$ Right now cost is of little consideration but even considering the reuse of the first generation BFR with 100 trips each the cost can go from 0.3% cost to 30% cost. So fuel costing half of the other per weigth unit means 15 % lower cost per launch. $\endgroup$ Commented Oct 3, 2017 at 4:14
  • 1
    $\begingroup$ @OuNelsonMangela just for the sake of clarity, 0.3% cost in 100 launches becomes 0.003*100/(0.997+0.003*100) = 23%. And considering there are other maintenance costs for each launch, those 23% become even smaller. $\endgroup$
    – metalim
    Commented Dec 23, 2017 at 10:52
  • 1
    $\begingroup$ @JoeJobs Yes, that's exactly it. It covers the surfaces, interferes with the fluid flow, and degrades the chamber materials. It's less of a problem in expendables, of course...! $\endgroup$
    – SusanW
    Commented Aug 22, 2020 at 17:19
  • 1
    $\begingroup$ And they can practice making methane via Sabatier at industrail scale on Earth where they can debug it, before they ship such a system to Mars. $\endgroup$
    – geoffc
    Commented Dec 17, 2021 at 19:00

Logistically, methane can be easier to work with than hydrogen. Methane's boiling point is about 110K, compared to hydrogen's 20K. This means that both fuel and oxidizer lines can be purged with gaseous nitrogen. Liquid hydrogen lines can only be purged with helium, as hydrogen's boiling point is below the melting point of other inert gases.


Another disadvantage to hydrogen is that it requires advanced metallurgy to prevent hydrogen embrittlement, where more common alloys tend to become prone to fracture and fatigue in high hydrogen environments.


Methane would allow for around 380 second specific impulse (~3.8km/s exhaust velocity), depending on the chamber pressure, expansion ratio, and other design parameters for the engine, while LH2/LOX engines have demonstrated ~450 second specific impulse (~4.5km/s exhaust velocity).

Despite this lower efficiency though, methane has a couple of major advantages. It is significantly higher density as a liquid than LH2 (0.42 g/cc, vs 0.07 for liquid hydrogen), so it requires far less tank volume and smaller piping for the same mass of propellant. It also doesn't need to be stored as cold as liquid hydrogen, which reduces the insulation and cooling requirements.

SpaceX has traditionally favored dense, easy to handle propellant (LOX/RP1) and simple engine designs (the Merlin is a simple gas generator cycle, rather than the more efficient [but more complex] staged combustion design used by most other modern rockets). As such, it makes sense that they would go for the easier to handle and simpler solution of a methane rocket rather than the high performance but difficult and complicated liquid hydrogen motor, so long as the methane provides them with the performance that they require (which will remain to be seen).

  • $\begingroup$ Methane really is a good choice. It is easy to handle. It performs well. It ought to be available on Mars. Why no one else has gone this route before is the more interesting question! $\endgroup$
    – geoffc
    Commented Feb 3, 2014 at 23:04
  • $\begingroup$ Can you add any references to support your claims? No doubting you, it is just good form. $\endgroup$ Commented Feb 3, 2014 at 23:56
  • $\begingroup$ Which claims in particular? Most of what I included in that post is either common knowledge or easily found, but I'd be happy to expand on any specific point (including references) if you'd like... $\endgroup$
    – Chris
    Commented Feb 6, 2014 at 22:47
  • $\begingroup$ Some references: wikipedia article on liquid rocket propellants cites specific impulses of 4462 m/s (455 s) for LOX/LH2, 3615 m/s (368.6 s) for LOX/methane. Per this paper the specific impulse for LOX/methane is 368.9 s. The Space Shuttle main engine (RS-25) had an Isp of 452.3 seconds. $\endgroup$ Commented Apr 21, 2014 at 10:25
  • 1
    $\begingroup$ SpaceX isn't using LH2. They're using RP1, which is kerosene, and/or hydrazine. The transition to Methane (CH4) is a significant improvement. $\endgroup$
    – aramis
    Commented May 14, 2014 at 16:55

A good question. In pre-EELV studies, NASA and the U.S. Air Force looked at LOX/methane. EELV resulted in the LOX/kerosene Atlas V and the LOX/hydrogen Delta IV.

At the 4th International Conference on Launcher Technology in 2002, Burkhardt et al. compared a reusable LOX/kerosene launch vehicle using the RD-180 type engine from the Atlas V with a LOX/methane vehicle using a possible engine of the same efficient staged combustion cycle:

The LOX/methane engine had about 3% higher specific impulse but that advantage was outweighed by the lower density of the liquid methane compared to kerosene.

LOX/kerosene was slightly better performing overall in terms of payload and expected to be lower cost to build and operate, the same result as the pre-EELV studies.

The reason LOX/hydrogen is comparable to or better than LOX/kerosene is that the specific impulse is much higher overcoming the even lower density problem. For the Space Shuttle the main engines operated from ground to orbit so the higher specific impulse of hydrogen at higher altitude was the reason for its use.

For a first stage that only operates to low altitud followed by a LOX/hydrogen second stage as in the Atlas and Delta, kerosene has comparable payload performance and may be lower cost because of vehicle size. For the Delta IV another advantage is commonality with the upper stage propellants.

Methane is not currently supplied at launch sites so a major facility investment would be needed.

Lack of long experience with operation is another negative for methane.

If the Raptor were to be used in space as in a Mars mission then the fact that both LOX and liquid methane are relatively easy to store in space compared to hydrogen or kerosene would be an advantage.



SpaceX is not shooting at the moon, SpaceX is shooting at Mars. Logistically, I’m not sure there are viable choices besides methane / LOX and Hydrazine / Tetroxide. The return shot requires fuel stored for an unknown time, which means the default conditions are chilly on Mars. RP-1 is a rock hard solid requiring complex heating to liquify it and LH2 is high pressure H2 requiring complex cooling to liquify it. Most fuels are rock hard solids. Performance is important but secondary. Hydrazine and Nitrogen Tetroxide are what I had expected, with perfect handling and storage properties. Methane and LOX, however, are both materiels with an abundant supply of people experienced in handling them, so they can be handled and stored, it’s just harder than hydrazine and nitrogen tetroxide.

Why not LH2 is obvious, the question is why not hydrazine and nitrogen tetroxide. If you were an astronaut, what fuel would you trust to have waiting for you on arrival without leaks so you can come home? To hell with the trip there, get me back and I’ll know we’re ready to go.

  • 4
    $\begingroup$ The problem with hydrazine and nitrogen tetroxide is you have to carry them to Mars with you: They can't be made there. Methane and LOX, on the other hand, can be made on Mars. $\endgroup$ Commented Feb 7, 2018 at 13:46

Excellent question. But twinned to that is also the question of the thrust level which is more than three times greater than the current Merlin engine used in the second stage of the Falcon 9.

I admit this is pure speculation. But based on prior SpaceX practices with the Merlin engine, I believe the Raptor might be intended for use in larger launch vehicles than the Falcon 9. Perhaps as a common engine used in the lower and upper stages of a upgraded Falcon Heavy, where the Raptor replaces the Merlin engines.

But I think the most likely possibility is a major change in the Falcon 9 design (and maybe the Falcon Heavy too) spurred by recent successes with the Grasshopper vertical landing test vehicle.

The Raptor would be a good engine for an oversized upper stage for the Falcon 9 or Falcon Heavy. By transferring more of the effort to reach orbit to the upper stage, the first stage has greater performance margin and also stages at a lower altitude and speed making powered recovery of the first stage to the launch site easier. Recovery and reuse of the 1st stage would save SpaceX a lot of money (particularly for the Falcon Heavy) and permit lower pricing.

  • $\begingroup$ The Raptor is not meant for use on the Falcon 9 family. It is intended for a much larger rocket now known only as BFR. $\endgroup$
    – Hobbes
    Commented Oct 1, 2015 at 8:53
  • $\begingroup$ There is a upper stage project now. $\endgroup$ Commented Jan 8, 2017 at 0:54
  • $\begingroup$ Indeed the thrust is higher and from the beginning the goal was to use on the BFG/ITS. The upper stage never intended to use any other engine As far as I know. There was a small change on the overall size of the vehicle, but they just lost a couple of engines in both upper and lower stages. $\endgroup$ Commented Oct 3, 2017 at 3:55

This article: http://www.nasaspaceflight.com/2014/03/spacex-advances-drive-mars-rocket-raptor-power/

also notes that methane/LOX engines do not suffer coking buildup as LOX/RP1 engines do, and can be run less oxygen-rich, which is easier on the pumps.


One factor nobody else mentioned yet is cost. SpaceX is a for-profit company, so cost matters a lot. Methane has become a lot cheaper recently: https://www.macrotrends.net/2478/natural-gas-prices-historical-chart The price of natural gas declined quite a lot due to technological advances in production (i.e. fracking). This made methane the cheapest rocket fuel. As of 2001, NASA was paying $0.98/gallon for liquid hydrogen, which equates to about \$16/MMBTU, which is much more expensive than LNG nowadays.

  • $\begingroup$ A full fuel load for the Falcon 9 costs on the order of $200k. The launch price is ~$50M. The cost of the fuel is insignificant (esp when SpaceX is already using cheap RP-1, not hydrogen). $\endgroup$
    – Hobbes
    Commented Mar 25, 2019 at 7:35
  • $\begingroup$ Can you add units, substance, and source info to support the chart? $\endgroup$
    – phil1008
    Commented Jan 25, 2022 at 0:59
  • $\begingroup$ The chart is price per MMBTU for natural gas but as Hobbes already noted fuel costs are an insignificant fraction of launch costs. Other important factors Elon mentioned are that hydrogen tends to damage metals and methane has a better specific impulse than RP-1. $\endgroup$ Commented Jan 26, 2022 at 1:58

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