The Moon has no global magnetic field today, but returned lunar samples carry a natural remanent magnetization recording ancient fields: a long-lived field of at least 1 μT at 4.2 Ga (troctolite 76535; revised upward to ~20–40 μT by the authors' 2017 restudy), surface intensities of roughly 20–110 μT between 4.25 and 3.56 Ga, a decline to a few μT after about 3.1 Ga, weak fields persisting to at least 2.0–2.5 Ga, and final cessation between about 1.92 and 0.80 Ga. Two intertwined questions remain open. First, existence: whole-rock paleointensity data support a long-lived core dynamo, while single-crystal paleointensity measurements return null fields over 4.36–3.2 Ga and impact-plasma amplification offers a non-dynamo route to strong magnetizations. Second, energetics: the inferred early intensities (~20–110 μT in the record compiled below; quoted as 40–120 μT by Evans and Tikoo 2022) exceed what a ~330 km-radius lunar core can conventionally power by orders of magnitude, and no proposed source — thermal or compositional convection, precession-driven mechanical stirring, inner-core precession dissipation, or episodic mantle-overturn heat extraction — is accepted as explaining the full intensity-time history. The problem is to determine the origin, power source, and quantitative history of the ancient lunar magnetic field in a way that is physically self-consistent and compatible with all paleomagnetic, geophysical, and thermal-evolution constraints.

Background

The Moon today has no global magnetic field, but Apollo-era measurements found that lunar rocks and the lunar crust carry natural remanent magnetization (NRM) — magnetization acquired when the material cooled and locked in the ambient magnetic field of its time. In igneous rocks that cooled through the Curie temperature of their magnetic minerals, this is a thermoremanent magnetization (TRM), whose strength scales with the ambient field, which allows the ancient field intensity (the paleointensity) to be estimated. Lunar paleointensity work is harder than its terrestrial counterpart because the magnetic carriers are metallic iron-nickel phases (kamacite and taenite) that often have multidomain behavior (large grains whose internal magnetic domains reorganize easily, degrading the fidelity of the recorded field), and because many samples were shocked, brecciated (crushed and re-welded by impacts), or reheated after formation. A global field persisting for millions of years requires an internal dynamo — self-generation of magnetic field by motion of electrically conducting fluid in a metallic core — so lunar magnetism has been used since the 1970s to probe whether the Moon differentiated a metallic core and how that core cooled.

Garrick-Bethell, Weiss, Shuster and Buz (2009) studied troctolite 76535 (a coarse-grained plutonic rock composed mainly of olivine and plagioclase), the oldest known essentially unshocked lunar rock (4.2–4.3 Ga — Ga denotes billions of years ago — by the U/Pb, Th/Pb, Sm/Nd and 40^{40}Ar/39^{39}Ar radiometric chronometers). By resolving the NRM into components that demagnetize at different alternating fields (coercivity components, a measure of how robust each remanence fraction is) and combining this with 40^{40}Ar/39^{39}Ar thermochronology, they inferred a field of at least ~1 μT (a microtesla, roughly two percent of Earth's present surface field) that persisted for millions of years at 4.2 Ga. Slow cooling ruled out transient impact-generated fields, leaving an early core dynamo as the favored source. At the time, the Moon's core had not been detected, Apollo-era paleointensities as high as ~100 μT sat uneasily beside this modest minimum, and no energy source was known that could sustain a dynamo in a body as small as the Moon (radius 1737 km), whose core should have cooled quickly.

The empirical picture has since transformed. Seismic reanalysis of Apollo data detected a fluid outer core (~330 km radius) overlying a solid inner core (Weber et al. 2011), establishing the core's existence and partially solidified state by seismic inference. A paleointensity record assembled from Apollo basalts and breccias (Shea et al. 2012; Tikoo et al. 2014; Strauss et al. 2021; Garrick-Bethell et al. 2017) indicates surface fields of ~20–110 μT between at least 4.25 and 3.56 Ga — comparable to Earth's present surface field — followed by a decline to no more than a few μT by 3.1–3.19 Ga, weak fields of a few μT persisting to at least 2.0–2.5 Ga (Tikoo et al. 2017; Cai et al. 2025), a rebound to ~5–21 μT at ~2.8 Ga recorded by Chang'e-6 farside basalt (Cai et al. 2024), and final cessation between ~1.92 and 0.80 Ga (Mighani et al. 2020).

This record poses a theoretical problem: the Moon's small core cannot plausibly power such fields by Earth-like thermal convection alone. The early intensities of 40–120 μT quoted by Evans and Tikoo (2022) as a summary of the whole-rock literature — the sample-by-sample record compiled above spans ~20–110 μT, including the ~20–40 μT of the 2017 restudy of troctolite 76535 — 'are unsustainably high and cannot be reconciled with the small lunar core size, requiring three orders of magnitude more energy than deemed possible' (Evans and Tikoo 2022). Candidate power sources include mechanical stirring of the core by mantle precession (Dwyer et al. 2011), thermal convection powered by dissipation of differential inner-core precession (Stys and Dumberry 2020), episodic enhancement of core heat flow by foundering of dense titanium-rich magma-ocean cumulates (mineral layers that crystallized and settled out of the Moon's early magma ocean; Evans and Tikoo 2022), and compositional convection from core crystallization (Zhai et al. 2024). Moreover, as of 2024 the very existence of a long-lived dynamo is contested: single-crystal paleointensity (SCP) measurements — which avoid multidomain-metal artifacts by using millimeter-scale plagioclase crystals — yield null fields for samples spanning 4.36–3.2 Ga, constraining any core dynamo to the Moon's first ~140 million years (Zhou et al. 2024; Cottrell et al. 2024), and self-consistent impact-magnetohydrodynamic simulations (simulations that solve the impact plasma dynamics and the magnetic field together) show that plasmas from basin-forming impacts can amplify a weak 2 μT dipole to ~43 μT near the basin antipode, providing a non-dynamo origin for strong magnetizations (Narrett et al. 2025). The field is thus in an unusual state: the observational record is richer than the 2009 authors could have hoped, yet its central interpretation is more contested than ever.

Problem Statement

Determine the origin, power source, and quantitative intensity-time history of the ancient lunar magnetic field. The current conflict between the whole-rock paleointensity record — surface fields of ~20–110 μT between at least 4.25 and 3.56 Ga, decline to no more than a few μT by 3.1–3.19 Ga, a rebound to ~5–21 μT at ~2.8 Ga, weak fields of ~2–5 μT persisting to at least 2.0–2.5 Ga, and cessation between ~1.92 and ~0.80 Ga — and the single-crystal paleointensity null fields over 4.36–3.2 Ga that constrain any core dynamo to the Moon's first ~140 million years is resolved only by identifying, epoch by epoch, the responsible mechanism(s) — dynamo and/or non-dynamo: energy source(s) and dynamo mechanism(s) able to sustain, within the Moon's small (~330 km-radius), partially solidified core and its thermal budget, the inferred field intensities and durations of the epochs treated as dynamo-driven, and non-dynamo mechanisms, such as impact-plasma amplification of a weak field, accounting for the sample magnetizations and crustal magnetic anomalies attributed to non-dynamo processes rather than to a dynamo — reproducing the full history including the decline, any rebound, and the final termination. The origin, power source, and history sought here are constrained by the seismically inferred core structure and by the independent thermal-evolution, geochemical, and orbital constraints recorded in the literature cited here, and by the measured paleointensity data, accounted for by method and epoch: the records treated as faithful are reproduced quantitatively, and those attributed to artifacts or to non-dynamo magnetization have an identified physical origin.

References

  1. Garrick-Bethell I., Weiss B.P., Shuster D.L., Buz J. (2009). Early Lunar Magnetism. Science 323(5912), 356–359. doi:10.1126/science.1166804. https://doi.org/10.1126/science.1166804
  2. Weber R.C., Lin P., Garnero E.J., Williams Q., Lognonné P. (2011). Seismic Detection of the Lunar Core. Science 331(6015), 309–312. doi:10.1126/science.1199375. https://doi.org/10.1126/science.1199375
  3. Dwyer C.A., Stevenson D.J., Nimmo F. (2011). A long-lived lunar dynamo driven by continuous mechanical stirring. Nature 479(7372), 212–214. doi:10.1038/nature10564. https://doi.org/10.1038/nature10564
  4. Shea E.K., Weiss B.P., Cassata W.S., Shuster D.L., Tikoo S.M., Gattacceca J., Grove T.L., Fuller M.D. (2012). A Long-Lived Lunar Core Dynamo. Science 335(6067), 453–456. doi:10.1126/science.1215359. https://doi.org/10.1126/science.1215359
  5. Weiss B.P., Tikoo S.M. (2014). The lunar dynamo. Science 346(6214), 1246753. doi:10.1126/science.1246753. https://doi.org/10.1126/science.1246753
  6. Tikoo S.M., Weiss B.P., Cassata W.S., Shuster D.L., Gattacceca J., Lima E.A., Suavet C., Nimmo F., Fuller M.D. (2014). Decline of the lunar core dynamo. Earth and Planetary Science Letters 404, 89–97. doi:10.1016/j.epsl.2014.07.010. https://doi.org/10.1016/j.epsl.2014.07.010
  7. Garrick-Bethell I., Weiss B.P., Shuster D.L., Tikoo S.M., Tremblay M.M. (2017). Further evidence for early lunar magnetism from troctolite 76535. Journal of Geophysical Research: Planets 122(1), 76–93. doi:10.1002/2016JE005154. https://doi.org/10.1002/2016JE005154
  8. Tikoo S.M., Weiss B.P., Shuster D.L., Suavet C., Wang H., Grove T.L. (2017). A two-billion-year history for the lunar dynamo. Science Advances 3(8), e1700207. doi:10.1126/sciadv.1700207. https://doi.org/10.1126/sciadv.1700207
  9. Mighani S., Wang H., Shuster D.L., Borlina C.S., Nichols C.I.O., Weiss B.P. (2020). The end of the lunar dynamo. Science Advances 6(1), eaax0883. doi:10.1126/sciadv.aax0883. https://doi.org/10.1126/sciadv.aax0883
  10. Stys C., Dumberry M. (2020). A Past Lunar Dynamo Thermally Driven by the Precession of Its Inner Core. Journal of Geophysical Research: Planets 125(7), e2020JE006396. doi:10.1029/2020JE006396. https://doi.org/10.1029/2020JE006396
  11. Strauss B.E., Tikoo S.M., Gross J., Setera J.B., Turrin B. (2021). Constraining the Decline of the Lunar Dynamo Field at ≈3.1 Ga Through Paleomagnetic Analyses of Apollo 12 Mare Basalts. Journal of Geophysical Research: Planets 126(3), e2020JE006715. doi:10.1029/2020JE006715. https://doi.org/10.1029/2020JE006715
  12. Evans A.J., Tikoo S.M. (2022). An episodic high-intensity lunar core dynamo. Nature Astronomy 6, 325–330. doi:10.1038/s41550-021-01574-y. https://doi.org/10.1038/s41550-021-01574-y
  13. Zhou T., Tarduno J.A., Cottrell R.D., Neal C.R., Nimmo F., Blackman E.G., Ibañez-Mejia M. (2024). A lunar core dynamo limited to the Moon's first ~140 million years. Communications Earth & Environment 5, 456 (2024). doi:10.1038/s43247-024-01551-z. https://doi.org/10.1038/s43247-024-01551-z
  14. Cottrell R.D., Zhou T., Tarduno J.A. (2024). Dataset of replicate Apollo sample magnetizations bearing on impacts and absence of a long-lived lunar dynamo. Scientific Data 11, 807 (2024). doi:10.1038/s41597-024-03626-0. https://doi.org/10.1038/s41597-024-03626-0
  15. Cai S., Qi K., Yang S., Fang J., Shi P., Shen Z., Zhang M., Qin H., Zhang C., Li X., Chen F., Chen Y., Li J., He H., Deng C., Li C., Pan Y., Zhu R. (2024). A reinforced lunar dynamo recorded by Chang'e-6 farside basalt. Nature 643(8071), 361–365 (published online 19 December 2024; print issue 2025). doi:10.1038/s41586-024-08526-2. https://doi.org/10.1038/s41586-024-08526-2
  16. Zhai K., Yin Y., Zhai S. (2024). Thermal and Dynamo Evolution of the Lunar Core Based on the Transport Properties of Fe-S-P Alloys. Geophysical Research Letters 51(14), e2024GL108131 (2024). doi:10.1029/2024GL108131. https://doi.org/10.1029/2024GL108131
  17. Cai S., Qin H., Wang H., Deng C., Yang S., Xu Y., Zhang C., Tang X., Gu L., Li X., Shen Z., Zhang M., He K., Qi K., Fan Y., Dong L., Hou Y., Shi P., Liu S., Su F., Chen Y., Li Q., Li J., Mitchell R.N., He H., Li C., Pan Y., Zhu R. (2025). Persistent but weak magnetic field at the Moon's midstage revealed by Chang'e-5 basalt. Science Advances 11(1), eadp3333 (2025). doi:10.1126/sciadv.adp3333. https://doi.org/10.1126/sciadv.adp3333
  18. Narrett I.S., Oran R., Chen Y., Miljković K., Tóth G., Mansbach E.N., Weiss B.P. (2025). Impact plasma amplification of the ancient lunar dynamo. Science Advances 11(21), eadr7401. doi:10.1126/sciadv.adr7401. https://doi.org/10.1126/sciadv.adr7401
  19. Kalman B., Yong W., Secco R.A. (2026). High pressure melting of Fe-Si alloys with applications to the lunar core composition and dynamo processes. High Pressure Research (2026), 1–15 (published online 30 May 2026). doi:10.1080/08957959.2026.2680425. https://doi.org/10.1080/08957959.2026.2680425