Characterising and correcting for a previously unconsidered source of scatter in measurements of equivalent dose

Single aliquot methods make it feasible to routinely determine multiple estimates of equivalent dose (De) for a sample. The scatter in De values observed within a sample has been used previously to make inferences about the resetting of the luminescence signal being measured prior to the event dated, about post-depositional processes that may have affected the sample, and about spatial variations in environmental dose rate on a small (millimetre) scale. However, in addition to processes associated with deposition and burial, scatter in the De values generated may also arise if the dose rate from the beta source used as part of the De determination method varies between grains or aliquots. Spatial variation in the beta dose arising from 90Sr plaque beta sources has been known for more than 20 years. What has not previously been described is variation in beta dose rate between different sample positions on the Risø TL/OSL instruments, one of the most widely used commercial systems used for luminescence dating. The variation arises from misalignment of the carousel used to hold samples with respect to the lid upon which the beta source is mounted. For one instrument in this study, this misalignment led to variations in apparent dose between different carousel positions of 14%. For most instruments the variability was lower (0.15–∼2%), but still detectable. A method for assessing the severity of this problem for individual instruments is described, along with a procedure for correcting for the effect and thus reducing the impact of this hitherto undescribed source of scatter in De.