ESR dating of optically-bleached quartz grains: Evaluating measurement repeatability and reproducibility

The true random uncertainty associated with ESR measurements of quartz grains at cryogenic temperature is currently unknown, as well as its impact on resulting equivalent dose (DE) estimates for ESR dating purpose. We provide here the results from the first quantitative evaluation study based on experimental data. Repeated single aliquot and multiple aliquot measurements of 100-200 mu m quartz grains show that the random uncertainty (1 sigma) on the measurement of an individual aliquot usually lies in the range of 1-2%, 2-3 and 3-6% for the Al, Ti and Ti-H signals, respectively. These values are strongly sample and laboratory dependent, as they directly correlate with the signal-to-noise ratio associated with each signal, and may also be influenced by the stability of the experimental setup. The resulting variability of the DE values obtained from the multiple aliquot additive dose method over repeated measurements is estimated to 9%, 11% and 24% (1 sigma) on average for the Al, Ti and Ti-H signals, respectively. In the light of this outcome, the reliability of a dose estimate derived from one single measurement may be reasonably questioned, and we strongly advocate for the necessity to systematically carry out repeated measurements of a given quartz sample, a procedure that is far from being systematic within the community. Finally, our reproducibility experiments suggest that the experimental setup, measurement temperature or operator do not induce a significant bias in the dose evaluation, provided that repeated measurement are carried out. Our small-scale interlaboratory comparison shows consistent DE results for the Al and Ti signals, giving us some general confidence for the comparison of the dating results obtained by independent laboratories. We nevertheless acknowledge that a larger intercomparison study should be performed within the ESR dating community in order to confirm these initial results.