Testing new kinetic models and calibration methods for Rock Surface Luminescence Exposure dating using controlled experiments

Rock surface luminescence exposure dating (RSLED) is an evolving tool in dating prehistoric rock surfaces. In this study, we explore and compare current and new models and calibration methods using optically stimulated luminescence (OSL) depth profiles from quartz grains and Infrared Stimulated Luminescence (IRSL; IR₅₀, pIRIR₂₂₅, and MET₂₉₀) from feldspar grains found in granite and sandstone samples.

This study shows that, although correct model assumptions improve the quality of exposure age estimates, significant discrepancies between observed and expected fitting parameter values remain, and these discrepancies leads to inaccurate age estimation. This is particularly the case when post-IR signals from feldspar are used. Here, the spectral dependency of luminescence signals is examined to better understand these problems. The demonstrated depth dependency of fitting parameters, previously assumed to be constant with depth, also gives rise to discrepancies in parameter values. The surprising observation that, in rocks, the IR₅₀ signal is apparently more easily bleached than the quartz fast-component OSL signal is explained in terms of the wavelength dependence of light attenuation increasing the effective path length for shorter wavelengths, and so changing the shape of the light spectrum with penetration depth. This complicates parameter estimates in exposure dating even more.

Alternative approaches (rather than parameter estimation) for estimating how long a rock surface has been exposed to light are considered, based on modelling the shape and position of the measured luminescence-depth profile. It is concluded that the most accurate exposure age is derived by interpolating the depth of an unknown profile onto a curve of profile depths from known age profiles (the Exposure Response Curve, or ERC, approach).