Space radiation measurements during the Artemis I lunar mission

The paper addresses the critical hazard posed by space radiation to long-duration human spaceflight, including increased risks of cancer, cataracts, degenerative diseases and acute tissue reactions. Space radiation arises from galactic cosmic rays (GCRs), Earth's trapped radiation belts, and solar-particle events (SPEs). Prior human-spaceflight radiation data predominantly came from low-Earth orbit (LEO) platforms such as the Space Shuttle and the ISS, which benefit from substantial geomagnetic shielding, and from lightly shielded robotic interplanetary probes. Limited Apollo-era beyond-LEO human data and ground-based simulations exist but with caveats. The purpose of this study is to report and analyze radiation measurements inside the heavily shielded Orion spacecraft during the uncrewed Artemis I lunar mission, to quantify dose rates across differing shielding environments, assess the effect of spacecraft orientation, and benchmark GCR dose equivalent relative to past observations. These results inform crew protection, validate Orion’s shielding performance for future crewed exploration, and guide mission design and operations to reduce exposure.

The authors contextualize Artemis I measurements with past datasets: LEO measurements from the ISS and Space Shuttle under significant geomagnetic and structural shielding; interplanetary cruise data from lightly shielded spacecraft like Mars Science Laboratory (MSL/RAD) and Lunar Reconnaissance Orbiter/CRaTER; sparse Apollo-era trans-lunar human dosimetry; and ground-based mixed-field simulation efforts. They highlight model developments for radiation belts (AE9/AP9-IRENE), GCR spectra (Badhwar–O’Neill), and transport codes (HZETRN, Geant4). Prior works reported GCR dose equivalent rates around 1.58 mSv/d (MSL/RAD during similar solar cycle phase), ~1.55 mSv/d (CRaTER in lunar orbit), and ~1.40 mSv/d extrapolated to free space from ISS high-latitude data. These references establish benchmarks against which Orion’s heavily shielded environment can be compared.

Radiation in the Orion crew cabin during Artemis I was monitored using multiple instruments placed at distinct shielding locations and within anthropomorphic phantoms: NASA’s Hybrid Electronic Radiation Assessor (HERA), ESA’s Active Dosimeter (EAD), DLR’s M-42, and NASA’s Crew Active Dosimeter (CAD). HERA and EAD provide both absorbed dose and dose equivalent estimations; M-42 and CAD measure absorbed dose. The Matroshka AstroRad Radiation Experiment (MARE) deployed instrumented life-size female phantoms (Helga and Zohar) with detectors on the skin and in internal organ sites to capture organ-specific doses. Detectors were hard-mounted on Orion structure (e.g., HERA HSU1 in the ‘storm shelter’, HERA HSU2 and EAD units on more exposed walls) and within/on the phantoms (M-42 and CAD) to sample varying local shielding. The mission profile included launch, passages through Earth’s inner (proton-dominated) and outer (electron-dominated) radiation belts, an interplanetary GCR-dominated phase, two close lunar fly-bys (providing transient GCR shielding), and Earth re-entry. Time-resolved measurements captured belt transits and interplanetary exposure. An unplanned 90° attitude change during the inner-belt transit enabled assessment of orientation effects on dose rate. Modeling tools complemented measurements: AP9-IRENE provided trapped proton flux estimates; detailed Orion CAD-based mass models supported transport simulations with HZETRN and Geant4, propagating the Badhwar–O’Neill GCR spectrum through local shielding to compare with measured dose rates and spectra. Spectral characterizations included energy-deposition spectra (M-42) and LET spectra (HERA), with ICRP60 quality factor computation to derive dose equivalent from absorbed dose. Cumulative doses were partitioned by flight phase (inner/outer belts, GCR interplanetary).

• Inner proton belt: Large spatial variation in dose rates consistent with local shielding. Peak absorbed dose rates differed by up to ~4× between most- and least-shielded locations, e.g., M-42 SN127 (high shielding) 69 µGy/min versus EAD MU01 240 µGy/min and HERA HSU2 287 µGy/min. Simulations reproduced a ~2× difference between HERA HSU1 (storm shelter) and HSU2 (crew cabin) peak rates (134 vs 287 µGy/min); a reference October 1989 SPE simulation indicated ~4× between crew cabin and storm shelter (414 vs 95 µGy/min).
• Interplanetary GCRs: Absorbed dose rates were similar across locations, indicating shielding mass has limited impact on GCR absorbed dose compared to belt exposure. HERA-derived mean quality factors were 2.30, 2.63, and 3.06 (lower in more shielded locations). GCR dose equivalent rates were 0.96–1.24 mSv/d, up to ~60% lower than some prior lightly shielded observations (e.g., MSL/RAD 1.58 mSv/d; CRaTER 1.55 mSv/d; ISS extrapolated free space 1.40 mSv/d).
• Orientation effect: A 90° spacecraft rotation during inner-belt passage reduced measured dose rate by ~50% relative to modelled proton flux, attributed to rotating Orion’s thickest shielding (dorsal airlock/ventral upper stage) into the predominant proton ‘plane’ (particles with pitch angles ≈90° to Earth’s magnetic field).
• Lunar fly-bys: GCR dose rate decreased by roughly one-third during close approaches due to the Moon’s solid-angle shielding.
• Cumulative doses: Mission exposure was dominated by GCRs, though the inner proton belt contributed up to ~23% of total absorbed dose for some locations (e.g., EAD MU01). Whole-mission cumulative absorbed dose reached up to 13.47 mGy (HERA HSU2). Total mission dose equivalents were 26.7–35.4 mSv, with 1.80–3.94 mSv attributable to belt passes.
• Organ dosimetry (MARE Helga): Inner-belt organ doses varied by about 2× among internal organs (e.g., spine lower than lungs), and front skin dose was ~3× spine dose; a ~20% left–right lung difference was observed, consistent with local shielding and field directionality. GCR organ dose variation was only a few percent, with the most shielded back skin showing the highest dose due to secondary particle production.
• Shielding efficacy: An ~80% increase in areal density (median Al-equivalent 25 g/cm² to 45 g/cm²) reduced GCR dose equivalent by ~30%, indicating modest biological-dose reduction for GCRs with substantial shielding mass, in contrast to substantial reductions during belt passes and SPE-like conditions.
• Model validation: HZETRN and Geant4 transport through detailed Orion mass models showed broad agreement with measured dose rates and spectra across flight phases.

Findings demonstrate that Orion’s heavy shielding effectively attenuates trapped-proton exposures and that operational strategies such as selecting spacecraft orientations to place maximal shielding along the dominant particle incidence direction can halve acute dose rates during directional radiation fields (e.g., inner belt or during SPE onset). While absorbed doses from GCRs are relatively insensitive to shielding mass, biologically weighted dose equivalents show modest reductions with substantial shielding, aligning with modern transport predictions. The measurements validate Orion for future crewed missions by showing controlled peak and cumulative exposures in line with design expectations (e.g., storm shelter performance for large SPEs). The broad agreement between detailed transport models and flight data in a heavily shielded interplanetary environment supports the use of these tools for mission planning, shielding optimization, and real-time risk management. Relative to historical observations from lighter, less shielded platforms, Artemis I GCR dose equivalent rates are lower, reinforcing the benefit of compact, massive vehicles. Nonetheless, mission exposures remain contingent on trajectory, shielding distribution, solar cycle modulation, and the occurrence and magnitude of SPEs. Within NASA’s current 600 mSv career dose limit framework, Artemis-class missions do not approach this cumulative limit, and extrapolations suggest that well-shielded longer missions (e.g., Mars transit) could achieve dose equivalents about 30% below past estimates; however, variability in space weather and design choices remain significant determinants.

This study provides the first comprehensive radiation measurements inside the heavily shielded Orion spacecraft beyond LEO, quantifying spatial and operational variability across belt transits, interplanetary GCR exposure, and lunar fly-bys. Key contributions include: (1) verification of Orion’s shielding effectiveness with up to fourfold reductions in proton-belt dose rates between highly and weakly shielded locations and confirmation of storm-shelter performance; (2) demonstration that spacecraft attitude can reduce directional-field dose rates by ~50%; (3) measurement of GCR dose equivalent rates (0.96–1.24 mSv/d) lower than prior lightly shielded observations, with total mission dose equivalents of 26.7–35.4 mSv; and (4) validation of modern radiation transport models using detailed Orion mass representations. These results inform spacecraft design (mass distribution and storm shelters), operational procedures (preferential orientations during belt crossings or SPEs), and mission planning under ALARA principles. Future research should include measurements during actual SPEs, expanded multi-mission datasets across solar cycles, refined directionality-aware operational protocols, and further validation of transport models across varying shielding configurations and trajectories.

• No solar-particle events occurred during Artemis I, limiting direct validation of storm-shelter performance against real SPE spectra and temporal profiles; comparisons relied on simulations and analogy to inner-belt proton environments.
• Results represent a single mission’s geometry, shielding distribution, and a specific phase of the solar cycle; extrapolation to other missions must account for differences in shielding, trajectory, solar modulation, and SPE frequency/severity.
• Orientation effects were observed during a single rotation event in the inner belt; broader characterization across attitudes and belt conditions would strengthen generality.
• Although models showed broad agreement, uncertainties in local mass distribution, particle pitch-angle distributions, secondary production, and detector response may affect detailed comparisons.
• GCR biological dose reductions with shielding were modest and location-dependent; finer-grained mapping of shielding thickness and composition would improve dose-equivalent predictions at organ level.