The use of feldspar for luminescence dating has been restricted because of anomalous fading. This has made its application to several important geological problems such as volcanic terrains difficult. Presently, two correction procedures are used to correct for anomalous fading. The present study tests these correction procedures using volcanic samples of known ages spanning the time period of 400 ka to 2.2 Ma. These correction procedures provided grossly underestimated ages (up to 60%). The possible causes for the underestimation are discussed.
The thermoluminescence (TL) glow curves from feldspars have been the subject of numerous studies, because of their importance in luminescence dating and dosimetry. This paper presents new experimental TL glow curves in a plagioclase feldspar, measured using the T
stop technique of glow curve analysis. Kinetic analysis of the experimental results is carried out for a freshly irradiated sample, as well as for a sample which has undergone optical treatment using infrared light for 100 s at 50°C. Application of the initial rise method of analysis indicates that the TL signals from both samples can be characterized by a continuous distribution of energy levels. By subtracting the TL glow curves measured at successive T
stop values, a series of TL glow curves is obtained which are analyzed using the empirical general order kinetics. It is found that all TL glow curves obtained by this subtractive procedure can be described accurately by the same general order parameter b ∼1.7. In a second attempt to analyze the same TL glow curves and possibly extract information about the underlying luminescence process, the shape of TL glow curves is analyzed using a recently proposed physical kinetic model which describes localized electronic recombination in donor-acceptor pairs. Within this model, recombination is assumed to take place via the excited state of the donor, and nearest-neighbor recombinations take place within a random distribution of centers. This recent model has been used recently to describe successfully several types of luminescence signals. This paper shows that it is possible to obtain good fits to the experimental data using either one of these two approaches.