A widely cited 15% agreement between Artemis I radiation detectors applies to dose measurements below 100 keV/μm of linear energy transfer in water. Above that threshold, lithium-fluoride detectors overestimated the comparison results by up to 30% for dose and about 15% for quality factor. The 2026 study compared detectors at matching positions on a mannequin inside Orion, not radiation exposure at the lunar surface or an individual astronaut’s health risk.
What the Artemis I study measured
NASA’s uncrewed Artemis I carried the Matroshka AstroRad Radiation Experiment, or MARE, inside the Orion spacecraft. A 2026 paper in Radiation Measurements compared two kinds of passive detector at matching positions on the Zohar anthropomorphic phantom, a mannequin used to measure radiation exposure. The study describes its detector method and results.
One detector was a lithium-fluoride fluorescent nuclear track detector (LiF FNTD). Radiation leaves tracks in the crystal that researchers analyze optically. The other was a conventional plastic nuclear track detector (PADC PNTD), which was chemically etched before analysis.
The comparison uses linear energy transfer, or LET, to sort particle tracks by how much energy they deposit over a distance in water. This matters because the agreement between detector types changes with LET.
Where the detectors agree, and where they diverge
For LET below 100 keV/μm in water, the study found that LiF FNTD dose estimates agreed with PADC PNTD results within 15%. Their quality-factor results, which account for radiation type when calculating dose equivalent, agreed within a few percent.
Above 100 keV/μm, LiF FNTDs overestimated the PADC PNTD results. The dose overestimate reached 30%, while the quality-factor overestimate was about 15%. So the 15% figure describes the lower-LET dose comparison, not the full LET range.
Readings beneath the vest on the Zohar phantom
At the same position beneath Zohar’s vest, the detectors produced different estimates. The table gives the study’s cumulative measurements for LET above 10 keV/μm and, separately, for the restricted 10–100 keV/μm interval. The reported uncertainties reflect counting statistics only; they exclude calibration and correction contributions.
| LET interval in water | Detector | Absorbed dose | Dose equivalent |
| Above 10 keV/μm | LiF FNTD | 3.1 ± 0.5 mGy | 48.5 ± 9.3 mSv |
| Above 10 keV/μm | PADC PNTD | 2.4 ± 0.1 mGy | 32.7 ± 1.3 mSv |
| 10–100 keV/μm | LiF FNTD | 1.4 ± 0.2 mGy | 13.4 ± 2.5 mSv |
| 10–100 keV/μm | PADC PNTD | 1.6 ± 0.1 mGy | 16.6 ± 0.7 mSv |
Absorbed dose, measured in milligrays (mGy), describes energy deposited per unit mass. Dose equivalent, measured in millisieverts (mSv), adjusts for radiation type. They are different quantities, so the figures should be read in their respective columns.
The full-range readings are higher for the LiF detector at this position, while the narrower 10–100 keV/μm results differ in the other direction for absorbed dose. That is consistent with the study’s finding that the methods’ agreement depends on LET. The values describe detectors beneath a vest on a phantom, not a personal dose or an individual health-risk estimate.
What separate Orion and ISS measurements add
A separate 2024 study reported galactic cosmic ray dose-equivalent rates of 0.96–1.24 mSv per day inside Orion during interplanetary cruise. It listed 0.68 mSv per day for the International Space Station (ISS) and 1.40 mSv per day for an ISS value extrapolated to free space. The last figure is an extrapolation, not an in-space ISS reading. These distinct conditions do not form a single direct Moon-versus-Earth-orbit comparison.
The 2026 paper’s detector comparison is anchored to phantom positions aboard Orion. Its central result is conditional: below 100 keV/μm LET in water, dose estimates agree within 15%; above that threshold, the LiF method reads higher than the plastic-detector comparator.
