A method for improving the measurement of low-field magnetic susceptibility anisotropy in weak samples

Many minerals and rocks have low susceptibilities and magnetic anisotropies on the order of the noise level of the measuring instrument. Anisotropy is often not significant in these samples when using the standard measurement procedure. We propose a method that uses stacking of data to improve the signal-to-noise ratio, thus extending the dynamic range for measurement and allowing for assessment of the data quality. The method makes it possible to obtain consistent directions of the principal axes for samples with an anisotropy on the order of, or even slightly below, the noise level of the instrument. For noisy datasets, the stacking procedure makes it easier to recover correct directions. However, the degree P and shape U of the anisotropy ellipsoid show large variations. Large values of P, in combination with a badly defined U, may indicate noisy data rather than a large anisotropy. The stacking procedure is especially useful for determining the magnetic anisotropy of single crystals that often have a low susceptibility but must be measured with high accuracy.

► AMS is computed from averages of repeated susceptibility measurements.
► Averaging (i.e., stacking) procedure markedly improves the data quality.
► Stacking method yields meaningful results when standard procedures do not.
► AMS is determined reliably for samples with low susceptibility and weak anisotropy.
► Directions of AMS principal axes are determined better than the ellipsoid shape.