Palaeomagnetic errors related to sample shape and inhomogeneity

Abstract

When determining the remanent magnetization (RM) of palaeomagnetic rock samples, it is assumed that the samples can be approximated by a magnetic dipole. This assumption greatly simplifies the inverse problem of determining the RM from the measured magnetic field of the sample. The magnetic field of the normally used cylindrical rock samples is however not identical to that of a dipole and care must be taken not to introduce any systematic errors. A numerical test of the effect of cylindrical sample shape on the determination of RM for a spinner-type magnetometer is presented. We find that for a spinner magnetometer the non-ideal sample shape has an insignificant effect (less than 0.3° error), for even the smallest possible sample-sensor distance. Comparing static and spinner-type magnetometers, it is clear that spinner-type magnetometers are less affected by the sample shape. Inhomogeneously magnetized samples are an obvious source of error when determining the RM. We numerically test the error in the determination of RM for the simple case of a laminated cylindrical sample with constant direction, but varying intensity of magnetization between the laminae. For strongly inhomogeneous samples we find an error of ∼4° for typical spinner-type magnetometer and ∼10° for static-type magnetometers.