Inter-laboratory calibration of low-field magnetic and anhysteretic susceptibility measurements

Inter-laboratory and absolute calibrations of rock magnetic parameters are fundamental for grounding a rock magnetic database and for semi-quantitative estimates about the magnetic mineral assemblage of a natural sample. Even a dimensionless ratio, such as anhysteretic susceptibility normalized by magnetic susceptibility (K-a/K) may be biased by improper calibration of one or both of the two instruments used to measure K-a and K. In addition, the intensity of the anhysteretic remanent magnetization (ARM) of a given sample depends on the experimental process by which the remanence is imparted. We report an inter-laboratory calibration of these two key parameters, using two sets of artificial reference samples: a paramagnetic rare earth salt, Gd2O3 and a commercial "pozzolanico" cement containing oxidized magnetite with grain size of less than 0.1 mum according to hysteresis properties. Using Gd2O3 the 10 Kappabridges magnetic susceptibility meters (AGICO KLY-2 or KLY-3 models) tested prove to be cross-calibrated to within 1%. On the other hand, Kappabridges provide a low-field susceptibility value that is ca. 6% lower than the tabulated value for Gd2O3, while average high-field susceptibility values measured on a range of instruments are indistinguishable from the tabulated value. Therefore, we suggest that Kappabridge values should be multiplied by 1.06 to achieve absolute calibration. Bartington Instruments magnetic susceptibility meters with MS2C sensors produce values that are 2-13% lower than Kappabridge values, with a strong dependence on sample centering within the sensor. The K-a/K ratio of ca. 11, originally obtained on discrete cement samples with a 2G Enterprises superconducting rock magnetometer and a KLY-2, is consistent with reference parameters for magnetites of grain size <0.1 mum. On the other hand, K-a values from a 2G Enterprises magnetometer and K values from a Bartington Instruments MS2C loop sensor for u-channel and discrete cement samples, will produce average K-a/K values that are unrealistically high if not properly corrected for the nominal volume detected by the sensors for these instruments. Inter-laboratory measurements of K and K-a for standard paleomagnetic plastic cubes filled with cement indicate remarkable differences in the intensity of the newly produced ARMs (with a standard deviation of ca. 21%), that are significantly larger than the differences observed from the calibration of the different magnetometers employed in each laboratory. Differences in the alternating field decay rate are likely the major source of these variations, but cannot account for all the observed variability. With such large variations in experimental conditions, classical interpretation of a "King plot" of K-a versus K would imply significant differences in the determination of grain size of magnetite particles on the same material. (C) 2003 Elsevier Science B.V. All rights reserved.