The effect of cooling rate on the intensity of thermoremanent magnetization (TRM) acquired by assemblages of pseudo-single domain, multidomain and interacting single-domain grains

Abstract

Experiments designed to measure the absolute palaeointensity of the geomagnetic field generally do so by comparing the ancient thermoremanent magnetization (TRM) retained by an igneous rock with a new TRM imparted in the laboratory. One problem with this procedure is that the relative magnitudes of the ancient and laboratory TRMs may be influenced, not only by the external field intensities at the time the two coolings took place, but also by the rate at which the coolings themselves occurred. Here, we present new measurements of this ‘cooling rate effect’ obtained from treatments in the laboratory differing in cooling rate by a factor of ∼200. Synthetic samples containing sized ferrimagnetic grains were used in the experiments. Theoretical considerations and previous experiments have indicated the cooling rate effect to be dependent on domain state. Increases in TRM magnitude of more than 7 per cent per order of magnitude decrease in cooling rate have been reported for assemblages of non-interacting single-domain (SD) grains. Here, we focus on magnetite grains in the less well-studied pseudo-single domain (PSD) and multidomain (MD) states using a range of applied field intensities to impart the TRMs. For the first time, we also measure the cooling rate effect in grains of titanomagnetite that have been oxyexsolved so that they contain strongly interacting SD lamellae. In all cases, the cooling rate effect measured was in the same sense as already observed in ideal magnetically non-interacting SD grains but was considerably weaker. On average, the effect did not exceed ∼3 per cent increase in TRM per order of magnitude decrease in cooling rate and did not show any systematic dependence on applied field intensity. In some samples containing coarser grains, the cooling rate effect was not distinguishable from zero. The sense and magnitude of the cooling rate effect remain uncertain in truly MD grains as different studies have produced discrepant results. For the more practically relevant case of PSD and interacting SD grains, which commonly dominate the TRM in igneous rocks, however, it appears that we can be more confident in our assertions. The cooling rate effect in such materials is in the same sense as in non-interacting SD grains but smaller: a consequence of long-range ordering. In lavas and small intrusions containing these, it is unlikely to exceed 10 per cent. Although a correction should always be attempted, the results of palaeointensity studies based upon such samples will generally not be severely biased.