Abstract Paleomagnetism relies4 on stable remanent magnetizations held in rocks to reconstruct the ancient geomagnetic field direction and intensity. However, rocks may carry secondary overprints that obscure or completely destroy the original signal. To study the stability of the magnetic vector(s), laboratories routinely apply static alternating field demagnetization along three orthogonal axes (AFD₃), which is fast, non-destructive, and easy to automate. Here, we present a multiparticle model that shows AFD₃ can deviate the natural remanent magnetization (NRM). Deviations can be avoided when fulfilling two conditions: (i) the NRM was acquired in a weak field where magnetization intensity varies linearly with field strength, and (ii) the sample is magnetically isotropic. The first condition is generally satisfied for rocks holding thermal or detrital remanent magnetizations, but not those affected by an isothermal remanence (e.g., lightning), even though AFD₃ is often used to remove them. In rocks with an anisotropic particle orientation distribution, stepwise AFD₃ progressively removes different coercivity subpopulations as a function of grain orientation so the effective remanence anisotropy of the surviving carriers changes during demagnetization. The anisotropy-driven deflection therefore evolves with AF step, producing curvilinear demagnetization trajectories. Our theoretical results argue for caution when applying AFD₃ to anisotropic samples or those with isothermal overprints. Undesired NRM rotation can be avoided by tumble demagnetization or mitigated by increasing the number of alternating field axis orientations.
Finn et al. (Sun,) studied this question.