How could we make fuel on Mars?

Why make things on Mars?

A crew on Mars would need oxygen to breathe, and a crew that wants to return home would need rocket propellant. Carrying all of that from Earth adds mass to every launch. In-situ resource utilisation, usually shortened to ISRU, means producing useful materials such as oxygen, water or fuel from what is already at the destination. In a 2020 NASA article, a MOXIE systems engineer at NASA's Jet Propulsion Laboratory said that a full-scale system built to make propellant for the flight home would need to scale up oxygen production by about 200 times what MOXIE would create. MOXIE is a small experiment, described below (NASA, 2020).

How it works

In plain terms, the machine takes in local material and turns part of it into something useful. The two resources discussed here are carbon dioxide from the atmosphere and water.

The first job is making oxygen. MOXIE is a small oxygen generator on NASA's Perseverance rover. It splits carbon dioxide from the Martian atmosphere electrochemically, separating one oxygen atom from each molecule (JPL).

The second job is making fuel. In the Sabatier reaction, carbon dioxide and hydrogen react over a catalyst to form methane and water (NTRS, 2017). A NASA engineering model captured carbon dioxide from a simulated Martian atmosphere with cryocoolers and fed it to a packed bed of ruthenium catalyst pellets (NTRS, 2017). A NASA-funded university project, described on a 2017 NASA page, proposes carrying out water electrolysis and carbon dioxide methanation in one reactor (Colorado School of Mines project).

What has been demonstrated

MOXIE extracted oxygen 16 times on Mars and made 122 grams of oxygen in total. Its best rate was 12 grams per hour, with purity of 98% or better. NASA's Jet Propulsion Laboratory says that peak was twice NASA's original goal for the instrument. MOXIE finished its final run in August 2023 (JPL).

In its first seven runs, to the end of 2021, MOXIE reached its target of six grams of oxygen per hour in each run. The runs covered day and night and different seasons. These results were published in the journal Science Advances (ScienceDaily, reporting MIT News). The same August 2022 report notes two limits. At that time, running at dawn or dusk, when temperatures change quickly, had not been demonstrated. Repeated on and off cycling is a thermal stress that can degrade the system over time (ScienceDaily).

A 2017 NASA engineering model of a Sabatier reactor examined long-duration use. Its abstract reports that catalyst failure was detected, but gives no methane output rate (NTRS, 2017).

What is only reported

These items come from plans, models and proposals. None has been built and run as a working plant on Mars.

Agency plans. NASA's Jet Propulsion Laboratory says the next step would be a full-scale system that includes an oxygen generator like MOXIE and a way to liquefy and store the oxygen (JPL). MOXIE itself did not store its oxygen: NASA's 2020 article says all the gases were to be vented back into the atmosphere after each run. The same article reports an estimate by the project lead, Michael Hecht, that a full-scale system might be a bit larger than a household stove and weigh about 1,000 kilograms, and says work was ongoing to develop a prototype (NASA, 2020).

Modelled plants. A 2025 extended abstract models full-scale plants producing 30 or 300 metric tonnes of propellant per Martian year, as a mixture of 3.5 parts oxygen to 1 part methane by mass. The modelled plant would take carbon dioxide from the atmosphere and water from ice mined below the surface. It was presented at ASCEND 2025. The record lists it as an extended abstract with a single expert technical review (NTRS, 2025).

Proposed reactors. The Colorado School of Mines project page describes planned work on ceramic cells, and reports no experimental results (project page).

What is speculative

Speculative: a crew flying home on propellant made on Mars. This is a goal of the work, not a result. NASA's 2020 article describes a full-scale system as a future step, with a prototype then in development (NASA, 2020). No source reviewed for this article reports methane production on Mars, or a crewed mission relying on local propellant.

Limits and open problems

• Scale. A full-scale propellant system would need about 200 times MOXIE's oxygen output, according to a JPL engineer quoted by NASA (NASA, 2020).
• Durability. The 2017 reactor model detected catalyst failure during long-duration tests (NTRS, 2017).
• Continuous running. MOXIE starts and stops with each run. The MIT report says that if MOXIE can operate despite this repeated cycling, it would suggest that a full-scale system designed to run continuously could do so for thousands of hours. This is a projection and remains untested (ScienceDaily).
• Power, heat and water. A plant would also need power, heat control and a steady water supply. None of the sources reviewed here shows water extracted from Martian ground ice at plant scale.

What to watch next

• Whether NASA's full-scale system, with oxygen liquefaction and storage, moves from plan to hardware (JPL).
• Independent, peer-reviewed analysis of full-scale plant designs such as the one in the 2025 extended abstract (NTRS, 2025).
• Demonstrations of operation at dawn or dusk, and of longer cycling (ScienceDaily).
• Methane output from a reactor. The Colorado School of Mines project planned to integrate Sabatier catalysts into its ceramic cells in a second phase to maximise methane formation (project page).