Na-rich feldspar as a luminescence dosimeter in infrared stimulated luminescence (IRSL) dating

One of the challenges in dating rock surfaces is the choice of the luminescence mineral. Although quartz is the preferred dosimeter in sediment dating, it is often not sufficiently sensitive when extracted from solid rocks. The intensity of signals from feldspars tends to be much less dependent on geological origin and erosion history, but the dosimetry of K-rich feldspar grains extracted from rocks is complicated because the internal dose rate is very dependent on the original feldspar grain size. The in situ grain size information is lost during the crushing process used to separate the grains for measurement. This latter problem does not apply to Na-rich feldspar because of the absence of internal radioactivity.The potential application of Na-rich feldspar as a luminescence dosimeter for the IRSL dating of rock surfaces is investigated using a variety of sediment samples from different geological settings for which independent age control is available. The blue and yellow luminescence emissions are measured for IR stimulation at 50 °C (IR50), and post-IR IR stimulation at 290 °C (pIRIR290). Thermal stability experiments imply that the corresponding signals in both emissions have comparable thermal stabilities and that all signals have similar recombination kinetics and are thermally stable over geological timescales. The IR50 doses measured using blue and yellow emissions are similar to or lower than quartz doses while pIRIR290 blue doses are higher than those from yellow emission and quartz doses. The fading rates measured for the IR50 signals are ∼3%/decade larger than those measured for the pIRIR290 signals in both yellow and blue emissions. Furthermore the average fading rates of both yellow signals are ∼3%/decade higher than the corresponding fading rates of the blue signals. However, there is no detectable correlation between fading rates and the measured De values. The residual doses measured from the laboratory-bleached samples and a modern analogue suggest that the IR50 signals in both blue and yellow emissions bleach to the same degree, as do the corresponding pIRIR290 signals, and that there is no significant naturally-unbleachable residual dose observed using these signals. Neither anomalous fading nor incomplete bleaching explains the observed dose discrepancy between the two emissions. Eight uncorrected and fading-corrected ages are calculated for each sample based on all four signals, using the dose rate relevant to Na-rich feldspar extracts (i.e. ∼3% K). The IR50 and pIRIR290 blue ages were also calculated assuming a dose rate based on 12.5% internal K (i.e. assuming that the blue signals were mainly derived from contamination by K-rich feldspar). The latter pIRIR290 blue ages are in agreement with the expected age control, raising the possibility that this signal originates mainly from K-rich feldspar contamination in our Na-rich fractions, and thus is not so useful in the luminescence dating of rock surfaces. On the other hand, the pIRIR290 fading-corrected ages based on the yellow emission are consistent with the independent age controls; higher preheat and stimulation temperatures may result in more stable yellow signals from Na-rich feldspar extracts from rocks, and so reduce the size of the fading correction. We conclude that, because this signal avoids the dosimetry difficulties of K-rich feldspar extracts, it has considerable potential in the IRSL dating of rock surfaces.