Paleointensity estimates derived from non-SD recording materials: Testing the IZZI technique on MD slag and a new bootstrap procedure

Experimental techniques to determine paleomagnetic field intensity are based on a theoretical framework that is valid only for single-domain (SD) ferromagnetic particles. Yet, most of the available materials exhibit distinctly non-SD properties. Hence, designing a robust paleointensity methodology for non-SD is a fundamental challenge in paleomagnetism. Here we test the IZZI Thellier absolute paleointensity technique on slag material that has properties of small MD. The test has two purposes: 1) to describe the characteristic non-SD patterns occurring in Arai plots, and 2) to identify the optimal approach in interpreting non-SD behavior. We produced 4 synthetic re-melted slag samples with identical magneto-mineralogical properties under field intensities of 20, 40, 60 and 80 μT. We ran three batches of IZZI paleointensity experiments on 10 specimens per samples. The IZZI experiments were designed to identify the dependency of the results on the ratio and the angle between the field used in the paleointensity experiment (B_TRM) and the field in which the NRM was acquired (B_NRM). The results demonstrate different types of Arai plots, systematically dependent on the angle and the proportion between B_TRM and B_NRM. Straight-line Arai plots occur when the two fields are parallel and equal, and seem to always give the 'true' slope. Convex curves occur when B_TRM is parallel and significantly stronger than B_NRM. Concave curves occur in all the other cases and yield two end-case slopes that are always different from the 'true' slope. The zigzags of the Arai plot increase with the proportion between B_TRM and B_NRM, and as the angle between B_TRM and B_NRM approaches 90°. We test the accuracy of the commonly used data analysis approach of best fitting line through a chosen segment of the Arai plot. We conclude that best fitting line in curved plots cannot provide robust paleointensity estimates. However, the two 'end-case' slopes in concave curves can provide adequate constraints for the 'true' value. We introduce a new bootstrap procedure to calculate a 95% confidence interval of the result. We substantiate the new procedure by conducting two independent tests. The first uses synthetic re-melted slag produced under known field intensities - 3 SD samples and 4 non-SD samples. The second compares paleointensity determinations from archaeological slag samples of the same age - 34 SD samples and 10 non-SD samples. The two tests suggest that the bootstrap technique is an optimal approach for non-ideal dataset.