IR-RF dating of sand-sized K-feldspar extracts: A test of accuracy

In this paper we use a recently developed radioluminescence (RL) attachment to the Risø TL/OSL reader to test the InfraRed-RadioFluorescence (IR-RF) dating method applied to K-feldspar rich extracts from our known-age archive samples. We present experiments to characterise the instrument performance and to test the reproducibility of IR-RF measurements. These experiments illustrate the high sensitivity and dose rate of our RL system, the negligible influence of the turntable movement on IR-RF signals and the effectiveness of the built in 395 nm LED at bleaching IR-RF signals. We measure IR-RF ages on a set of samples with independent age control using a robust analytical method, which is able to detect any possible sensitivity change. Our IR-RF ages do not agree well with the independent age control; the ages of the younger samples (20-45 ka) are significantly over-estimated while the ages of the older samples (∼130 ka) are significantly under-estimated. Experiments are undertaken to investigate this disagreement and our results indicate that they can most likely be explained by 1) the difficulty of defining the correct bleaching level prior to regeneration measurements, 2) signal instability, 3) sensitivity changes between the additive dose and regenerative dose measurements, or a combination of these three factors.