The chemistry is old and well understood. Electrolysis splits water into hydrogen and oxygen. The Sabatier reaction combines that hydrogen with atmospheric CO2 to produce methane and more water, which is recycled back through the electrolyser. The outputs are methane and oxygen — the propellant combination modern Mars vehicles are designed around — plus surplus oxygen and water for the crew.
The oxygen half has flown. MOXIE aboard Perseverance produced oxygen directly from Martian atmospheric CO2 across 16 runs between 2021 and 2023, up to 12 grams per hour, 122 grams in total. It is a small number that settles a large question: consumables can be manufactured on Mars from Martian material.
The economics are decisive. Returning a crewed vehicle from the Martian surface requires on the order of a thousand tonnes of propellant. Launching that from Earth and landing it on Mars is prohibitive at any launch cost anyone projects. Manufacturing it in the two years between cargo arrival and crew arrival converts an impossible mission architecture into a routine one — which is why every serious Mars plan since Zubrin's Mars Direct has been built around ISRU.
What has not been demonstrated is the integrated plant. Mining ice from frozen regolith, purifying it, running electrolysis and Sabatier continuously for two years, liquefying and storing cryogenic methane and oxygen through Martian winters, all unattended and with no maintenance crew — that full chain has never operated anywhere, and it is on the critical path of every crewed Mars mission.
MOXIE manufactured oxygen from Martian air on the surface of Mars. The principle is no longer theoretical.
Roughly a thousand tonnes of propellant cannot be landed from Earth at plausible cost. Making it locally is the difference between a mission and a fantasy.
Electrolysis and the Sabatier reaction are textbook industrial processes. The engineering challenge is autonomy and duty cycle, not discovery.
The same reactor chain that produces propellant produces breathing oxygen and potable water, so life support and departure share infrastructure.
Ice mining, purification, electrolysis, Sabatier and cryogenic storage operating together, unattended, for two years is unproven end to end — and it is on the critical path.
Melting frozen regolith and running electrolysis at propellant scale needs continuous power well beyond anything yet demonstrated on a planetary surface.
Liquid methane and oxygen must be kept cold for two years with no crew and minimal boil-off, in an environment with dust, thermal cycling and no maintenance.
If the plant underproduces, the return vehicle does not fly. This single-point dependency shapes every mission rule around it.
| Propellant | ~1,000 t per return | Methane + oxygen |
| Production window | ~2 yr | Between cargo and crew arrival |
| Water input | Mined ice | Excavated and melted |
| Power | Continuous, 100s kW | Melting and electrolysis dominate |
In-situ resource utilisation: manufacturing what a mission needs from Martian material rather than shipping it. On Mars that means turning atmospheric carbon dioxide and subsurface water ice into oxygen, drinking water and methane propellant.
Electrolyse mined water into hydrogen and oxygen, then run the Sabatier reaction to combine the hydrogen with atmospheric CO2, producing methane and more water. The result is methane and liquid oxygen — the propellant pair modern Mars vehicles use.
On the order of a thousand tonnes, which is why manufacturing it on Mars during the roughly two years between cargo landing and crew arrival is central to every credible mission architecture.
