Interpretation of single grain De distributions and calculation of De

Abstract

Recent development of an instrument for measuring the optically stimulated luminescence signal from individual mineral grains has made it practicable to measure the equivalent dose (De) from many hundreds or thousands of single mineral grains from a sample. Such measurements can potentially be used to address issues such as sample integrity, and to make it possible to obtain ages from samples that consist of mixtures of grains, enlarging the range of materials to which luminescence dating can be applied. However, for reliable ages to be obtained, the characteristics of the equipment and the sample being analysed need to be understood. Using sensitised sedimentary quartz grains, the instrumental uncertainty in repeated optically stimulated luminescence measurements made using a single grain laser luminescence unit attached to a conventional luminescence reader was evaluated; a value of 1.2% was obtained. Grains from this sample were then used to investigate the uncertainty in a measured dose distribution obtained using the single aliquot measurement protocol on each grain that had previously received a known laboratory dose; after systematic rejection of grains that did not pass defined acceptance criteria, over dispersion of 7% was found. Additional spread in data was found when uniform Aeolian sands were examined, resulting in over dispersion of ~ 12%; this was attributed to a combination of factors relating to differences in field and laboratory conditions. A similar value was found for an archaeological horizon below this sand. For another sample from the same section, a significantly larger value was found, ~ 29%; on this basis the finite mixture model was applied to obtain the likely dose components. The paper demonstrates the importance of correct assessment of error terms when analysing single grain De distributions and a number of rejection criteria that are vital to avoid the inclusion of data that could lead to misinterpretation of the degree of scatter present.