ISRU ascent propellant: making the oxygen to fly home

Study 005 · Draft · Concept study, not a built system.

Study 005 · Draft · First published 2026-10-13 · Revised 2026-10-13

Question

How much oxygen must Mars make to send four people home?

Assumptions

  • Ascent vehicle4 crew; 22.7 t O₂ + 7.0 t CH₄; 480 days of continuous production (Kleinhenz & Paz 2017)
  • ReferenceDRA 5.0 (6 crew): about 24 t ascent O₂, plus about 2 t life-support O₂ and 3.5 t water; CH₄ from Earth; 30 kWe fission
  • MOXIE best12 g/h
  • Energy floor4.9 kWh per kg O₂ (Study 001)

The numbers

Rate:  22,700 kg / (480 d × 24 h) ≈ 1.97 kg/h ≈ 164 × MOXIE (12 g/h)
NASA range:  2.2–3.5 kg/h O₂ for DRA 5.0, O₂-only, 480 or 300 days (Sanders)

Energy floor:  1.97 kg/h × 4.9 kWh/kg ≈ 9.7 kW continuous (thermochemistry only)

Published plant estimates (Kleinhenz & Paz 2017):
  Case 1, O₂ only from CO₂:              ≈ 0.9 t hardware, 34 kW
  Case 2, O₂ + CH₄ from regolith water:   ≈ 1.7 t hardware, 52 kW (17 kW of it regolith heating)
DRA 5.0 daytime-only plan:  66 kWe × 8 h/sol for 300 sols (22 kWe average)

Mass leverage:  22.7 / (22.7 + 7.0) ≈ 76 % of propellant is oxygen

Result

Modeled: making the ascent oxygen on Mars removes about three quarters of the ascent propellant mass from the cargo manifest. The cost is a plant running about 2 kg/h for over a year at roughly 34 kW (oxygen only), or 52 kW if it also makes the methane from regolith water. That is the largest single power load in the architecture.

Trade-offs

RouteFeedstockPlant (Kleinhenz & Paz 2017)Status
Solid-oxide electrolysis (MOXIE-style), O₂ onlyatmospheric CO₂≈ 0.9 t, 34 kWMOXIE: 16 runs, 122 g total
Water electrolysis + Sabatier, O₂ + CH₄mined water + CO₂≈ 1.7 t, 52 kWno Mars demonstration; soil processing ≈ 41–66 kg/h at 3% water (Sanders)
Bring all O₂ from Earth—0 kW+22.7 t landed per crew

Open problems

  • A continuous year-long solid-oxide run at about 160× MOXIE scale.
  • Dust ingestion in compressors; solid-oxide stack degradation.
  • Boil-off of 22.7 t of liquid oxygen held for months (DRA 5.0 budgets about 3 kWe at night to keep propellant liquid).

What would change the answer

  • Run production on daytime power and pause in storms, as DRA 5.0 assumed, if storage and schedule allow.

Sources

  1. Kleinhenz & Paz, “An ISRU Propellant Production System to Fully Fuel a Mars Ascent Vehicle,” AIAA 2017-0423 (2017)
  2. NASA, Human Exploration of Mars Design Reference Architecture 5.0, NASA-SP-2009-566
  3. “Power Requirements for the NASA Mars Design Reference Architecture (DRA) 5.0,” NTRS 20120012929
  4. Sanders et al., “Mars ISRU for Production of Mission Critical Consumables: Options, Recent Studies, and Current State of the Art,” NTRS 20150016009
  5. NASA, “NASA’s Oxygen-Generating Experiment MOXIE Completes Mars Mission” (2023)
  6. Hoffman et al., “Mars Oxygen ISRU Experiment (MOXIE): Preparing for human Mars exploration,” Science Advances (2022)

Status: Sketch is first-order arithmetic that may move by ±50%; Draft has full structure but uses an approximation we name; Study has every input sourced. Figures are rounded. Where a value is our assumption, the table says so. Errors are logged on Corrections; report one on Discord.

← All design studies