On the intrinsic accuracy and precision of luminescence dating techniques for fired ceramics

Abstract

Luminescence dating techniques have been used extensively for archaeological and geological samples. Such techniques are based on thermally or optically stimulated signals. This paper presents simulations of several luminescence techniques for equivalent dose (ED) estimation for ceramic materials containing quartz. The simulations are carried out using a recently published comprehensive kinetic model for quartz, consisting of 11 electron and hole traps and centers. The complete sequence of the experimental protocols for several thermoluminescence (TL) and optically stimulated luminescence (OSL) techniques are simulated using the same set of kinetic parameters. The specific simulated protocols are: additive dose TL protocol, predose technique (both additive and multiple activation versions), phototransfer protocol, single aliquot regenerative optically stimulated luminescence (SAR-OSL) protocol, and SAR thermoluminescence protocol (SAR-TL). One hundred random variants of the natural samples were generated by keeping the transition probabilities between energy levels fixed, while allowing simultaneous random variations of the concentrations of the 11 energy levels. The relative intrinsic accuracy and precision of the protocols are simulated by calculating the equivalent dose (ED) within the model, for a given natural burial dose of the sample. The intrinsic accuracy of these techniques is estimated by simulating natural irradiation of the samples with a known burial dose, followed by simulation of the luminescence method used to recover the estimated dose ED. The percent difference between the burial dose and the ED value represents the simulated accuracy of the luminescence technique. The relative intrinsic precision of these techniques is estimated by fitting Gaussian probability functions to the ED values obtained with the 100 model variants. It is found that the various techniques can reproduce natural paleodoses in the range 10 mGy–10 Gy with a typical intrinsic accuracy of +1 to 10%. Techniques based on single aliquot protocols were found in general to be more precise than techniques requiring the use of multiple aliquots. In addition, techniques based on interpolation of experimental data were found to be consistently both more precise and accurate than those based on extrapolation of experimental data.