Meta-Stable Magnetic Domain States That Prevent Reliable Absolute Palaeointensity Experiments Revealed By Magnetic Force Microscopy

Obtaining reliable estimates of the absolute palaeointensity of the Earth’s magnetic field is notoriously difficult. Many methods to obtain paleointensities from suitable records such as lavas and archeological artifacts involve heating the samples. These heating steps are believed to induce ‘magnetic alteration’ – a process that is still poorly understood but prevents obtaining correct paleointensity estimates. To observe this magnetic alteration directly we imaged the magnetic domain state of titanomagnetite particles – a common carrier of the magnetic remanence in samples used for paleointensity studies. We selected samples from the 1971-flow of Mt. Etna from a site that systematically yields underestimates of the known intensity of the paleofield - in spite of rigorous testing by various groups.

Magnetic Force Microscope images were taken before and after a heating step typically used in absolute palaeointensity experiments. Before heating, the samples feature distinct, blocky domains that sometimes seem to resemble a classical magnetite domain structure. After imparting a partial thermo-remanent magnetization at a temperature often critical to paleointensity experiments (250 °C) the domain state of the same titanomagnetite grains changes into curvier, wavy domains. Furthermore, these structures appeared to be unstable over time: after one-year storage in a magnetic field-free environment the domain states evolved into a viscous remanent magnetization state.

Our observations may qualitatively explain reported underestimates from technically successful paleointensity experiments for this site and other sites reported previously. Furthermore the occurrence of intriguing observations such as ‘the drawer storage effect’ by Shaar et al (EPSL, 2011), and viscous magnetizations observed by Muxworthy and Williams (JGR, 2006) may be (partially) explained by our observations. The major implications of our study for all palaeointensity methods involving heating may be evident.