On the applicability of the leading edge method to obtain equivalent doses in OSL dating and dosimetry

Abstract

The applicability of an objective methodology for dose distribution analysis proposed by Lepper and McKeever [2002. An objective methodology for dose distribution analysis. Radiat. Prot. Dosimetry 101 (1–4), 349–352] to samples, where available material only allows the measurement of a small number of aliquots, is investigated. The basis for the study are artificially and naturally insufficiently bleached samples for which a large number ( > 90 ) of aliquots were measured. Smaller sets of aliquots are then randomly drawn from this pool and a Gaussian function fitted to the rising limb of the dose distribution. It is found that when the method is based on a three-parameter fit, it is no longer applicable when the number of aliquots is significantly reduced. The percentage of successful fits for data sets made of only 20 aliquots was on average below 20%. In addition for data sets composed of 40 to 60 aliquots, palaeodoses obtained from fitting spread over a large dose range and showed a strongly asymmetric distribution. A significant improvement could be achieved by fixing the center of the Gaussian function to the mode D E bin and only fitting a two-parameter function to the data sets. With this approach an 80% change of successful fitting could be achieved even when using dose distributions constructed from only 20 to 30 aliquots. The median of the palaeodose distributions was approximately independent of the number of aliquots used. The modified method is then applied to samples from a sedimentary record in the northern Oman mountain range and the results compared with the method proposed by Fuchs and Lang [2001. OSL dating of coarse-grain fluvial quartz using single-aliquot protocols on sediments from NE Peleponnese, Greece. Quat. Sci. Rev. 20, 783–787]. It is concluded that the leading edge method proposed by Lepper and McKeever can be useful to derive D E 's from insufficiently bleached sediments even for a small number of aliquots if the Gaussian function is fitted as a two-parameter function.