Making oxygen and fuel on Mars

Also known as Mars in-situ resource utilisation, Mars ISRU, MOXIE, Mars propellant production

NASA's MOXIE instrument on the Perseverance rover made oxygen from the Martian atmosphere over 16 runs, reaching 12 grams per hour at its most efficient. Making methane fuel and extracting water on Mars, and scaling oxygen production for crewed missions, remain at model, proposal or concept stage.

Research and prototypesDemonstratedReviewed 22 Oct 2026
AI illustration · Illustrative scene, not a photograph of the specific device or experiment.

In plain language

Almost everything a crew needs on Mars would have to be launched from Earth, which is expensive. In-situ resource utilisation, or ISRU, means making some of those supplies from local material instead. On Mars the obvious raw materials are the thin atmosphere, which is mostly carbon dioxide, and water, which is locked up as ice underground or in minerals.

The first real test came with MOXIE, a small box on NASA's Perseverance rover that pulled in Martian air, separated the carbon dioxide and released oxygen. Over 16 runs it made 122 grams of oxygen in total. A full crewed mission would need far more, and it would also need fuel for the trip home, which is a much bigger demand. The idea of making rocket fuel from Martian carbon dioxide and water is still a plan, not a working plant.

Technical explanation

MOXIE uses a solid oxide electrolyser. It takes in Martian air, electrochemically separates the carbon dioxide, and measures the oxygen it releases for quantity and purity. NASA reports that the instrument reached 12 grams of oxygen per hour at its most efficient, with purity of 98% or better. The first seven runs, to the end of 2021, were reported to meet a target of six grams per hour in each run across different atmospheric conditions, including day and night and different seasons. The published account of those runs is a Science Advances study whose full text was not reviewed for this profile, so the numbers here come from the MIT news release describing it.

Fuel production needs a different chemistry. The Sabatier reaction combines carbon dioxide with hydrogen over a catalyst to make methane and water. Water electrolysis can supply the hydrogen. An engineering model by NASA Kennedy Space Center engineers in 2017 studied a Sabatier reactor with a ruthenium catalyst on alumina pellets, fed by cryocooler-captured carbon dioxide. Its authors examined long-duration operation and detected catalyst failure, but the abstract does not give a methane output rate. A 2025 NASA-hosted feasibility study modelled full-scale plants producing 30 or 300 tonnes of oxygen and methane propellant per Martian year, in a mixture of 3.5 parts oxygen to 1 part methane by mass. That work is an extended abstract with a single expert technical review, so it has not had full peer review.

The main engineering problem is that these steps have never been run together as a continuous plant on Mars. Each stage needs power, heat management, dust tolerance, storage at cryogenic temperatures and a water supply that is itself proven on site. None of these has been shown at the scale needed for a crewed mission.

What the evidence says

Each statement is typed by the strength of its evidence and linked to the source that supports it.

Demonstrated Shown in a published experiment, test or shipping product.

MOXIE on NASA's Perseverance rover made 122 grams of oxygen over 16 runs, and NASA reports a peak rate of 12 grams per hour at its most efficient, with purity of 98% or better.

Sources: [1] · Mission summary

In its first seven runs, to the end of 2021, MOXIE reached its target of six grams of oxygen per hour in each run, across varied atmospheric conditions including day and night and different seasons.

Sources: [3] · Press release body, reported from the Science Advances study · medium confidence

The MIT account of the Science Advances study reports that running MOXIE at dawn or dusk, when temperatures change quickly, has not been demonstrated, and that repeated on and off cycling is a thermal stress that can degrade the system.

Sources: [3] · Quoted statements from the project lead

A 2017 NASA Kennedy Space Center engineering model of a Sabatier reactor combines carbon dioxide captured by cryocoolers with hydrogen to make methane and water, and its authors examined long-duration operation and detected catalyst failure.

Sources: [4] · Abstract · medium confidence

Proposed A design or plan that has not yet been built or tested.

NASA's Jet Propulsion Laboratory release describes the next step as a full-scale system that would include a way to liquefy and store oxygen. The release does not say whether MOXIE's own oxygen was vented or stored.

Sources: [1] · Closing paragraphs · medium confidence

NASA estimated in 2020 that a full-scale oxygen system for a rocket return would need to produce about 200 times MOXIE's oxygen, and said it might be a bit larger than a household stove and weigh about 1,000 kilograms.

Sources: [2] · Body text on scale-up · medium confidence

An extended-abstract feasibility study models full-scale Martian ISRU plants producing 30 or 300 metric tonnes of propellant per Martian year, as a 3.5 to 1 oxygen to methane mixture by mass, and has had a single expert technical review rather than full peer review.

Sources: [5] · Study purpose and modelled production targets · medium confidence

A NASA-funded project at the Colorado School of Mines proposes protonic ceramic cells that would combine water electrolysis and carbon dioxide methanation in one reactor, but the project page reports no experimental results.

Sources: [6] · Project description and phases · low confidence

Limitations and open problems

  • MOXIE is a small instrument that runs in short bursts. NASA's 2020 estimate is that a full-scale oxygen system for a rocket return would need about 200 times MOXIE's output and might be a bit larger than a household stove, weighing about 1,000 kilograms. That is an estimate, not a tested design.
  • The MIT account of the Science Advances study says that operation at dawn or dusk, when temperatures change quickly, has not been demonstrated.
  • Repeated switching on and off creates thermal stress that can degrade the system over time, and the service life of a continuously running plant is not known from the evidence reviewed.
  • The 2017 Sabatier engineering model detected catalyst failure during long-duration tests, and its abstract gives no methane output rate. The sources reviewed show no methane production on Mars.
  • The 2025 full-scale plant model is an extended abstract with a single expert review, so its mass, power and volume figures should be treated as provisional until a full peer-reviewed paper is available.
  • The sources reviewed here do not show water being extracted from Martian ground ice at the scale a plant would need. Orbital maps of ice are not reserve estimates.

Possible future uses

Speculative

Ideas, not predictions. None of these has been demonstrated.

  • A Mars ascent vehicle fuelled with oxygen and methane made locally from Martian carbon dioxide and water, reducing the mass that must be launched from Earth.
  • A continuously running oxygen plant that stores its output ahead of a crew's arrival, supplying breathing oxygen as well as propellant.
  • A single protonic ceramic reactor that electrolyses water and converts carbon dioxide to methane in one unit, as proposed in a NASA-funded project at the Colorado School of Mines.

Timeline

  1. Dec 2021MOXIE completes seven experimental runs under varied atmospheric conditions, each meeting a target of six grams of oxygen per hour. Sources: [3]
  2. Aug 2022Science Advances publishes the early MOXIE results, as reported in an MIT news release. Sources: [3]
  3. 6 Sept 2023NASA's Jet Propulsion Laboratory announces that MOXIE has made its sixteenth and final run, bringing the total to 122 grams of oxygen. Sources: [1]
  4. Jul 2025Andersen and colleagues present an extended-abstract model of full-scale Martian propellant plants producing 30 and 300 tonnes per Martian year. Sources: [5]

Futures this could help bring about

AI illustration
How far away?

A self-sustaining settlement on Mars

Science unproven · 3 of 9 steps done

AI illustration
How far away?

Humans landing on Mars

Engineering prototypes · 2 of 7 steps done

Related breakthroughs

Glossary

In-situ resource utilisation (ISRU)
Making useful materials, such as oxygen, water or fuel, from resources found at the destination rather than launching them from Earth.
Sabatier reaction
A chemical reaction in which carbon dioxide and hydrogen react over a catalyst to form methane and water.
Propellant
The fuel and oxidiser, usually oxygen and methane here, that a rocket burns to lift off.
Catalyst
A material that speeds up a chemical reaction without being used up by it. Catalysts can also degrade over time, which limits how long a reactor lasts.

Sources

  1. government report
    NASA's Oxygen-Generating Experiment MOXIE Completes Mars Mission
    NASA Jet Propulsion Laboratory, 6 Sept 2023. Accessed 22 Oct 2026.
  2. government report
    MOXIE Could Help Future Rockets Launch Off Mars
    NASA Jet Propulsion Laboratory, 24 Nov 2020. Accessed 22 Oct 2026.
  3. university press release
    MOXIE experiment reliably produces oxygen on Mars
    Jennifer Chu. ScienceDaily (republishing MIT News), 31 Aug 2022. Accessed 22 Oct 2026.
  4. conference paper
    Mars Atmospheric Conversion to Methane and Water: An Engineering Model of the Sabatier Reactor with Characterization of Ru/Al2O3 for Long Duration Use on Mars
    Anne J. Meier, Malay G. Shah, Paul E. Hintze et al.. International Conference on Environmental Systems (ICES), Jul 2017. Accessed 22 Oct 2026.
  5. conference paper
    Assessing the Feasibility of a Martian ISRU Propellant Plant
    Noah L. Andersen, Avery L. Carlson, Kyle L. Ostendorp et al.. NASA Technical Reports Server (ASCEND 2025), Jul 2025. Accessed 22 Oct 2026.
  6. grant
    Producing Methane and Oxygen on Mars Using Proton-Conducting Ceramics
    Duc Nguyen. NASA Space Technology Research Grants, 12 Sept 2017. Accessed 22 Oct 2026.

Revision history

3 changes since this profile was created.

Changes on 22 Oct 2026
Published on owner instruction (owner requested space, food and water topics)dc4cd89
Sources and claims re-checked against originals409337d
Initial profile drafted from primary sourceseaac9eb