A MODEL FOR THE INFLUENCE OF DEFECT INTERACTIONS DURING HEATING ON THERMOLUMINESCENCE IN LIFMG, TI (TLD-100)

Thermoluminescence in LiF:Mg,Ti is frequently described with the Randa ll-Wilkins model which assumes first-order kinetics. This model fits a measured single glow curve (light output as a function of temperature during heating up the sample) very well. If this model describes the thermoluminescence phenomenon correctly, one expects that the model pa rameters are not dependent on the experimental conditions. However, th e model parameters show a strong dependence on read-out heating rate a nd annealing procedure. To explain this dependence it is assumed that the defects involved in the thermoluminescence process are also involv ed in reactions with other defects. In this article, defect reactions are incorporated into the rate equations for thermoluminescence. Glow curves are simulated at different read-out heating rates for the simpl e case in which ons defect reaction occurs simultaneously with a trap emptying process. It appears that the simulated glow curves can be ver y well fitted by the Randall-Wilkins model and that the dependence of the parameters of these simulated curves on the heating rate during re ad-out is comparable with measured data of LiF: Mg,Ti. Moreover, the o ccurrence of defect reactions during read-out also gives an explanatio n for (i) the dependence of the Randall-Wilkins parameters on the anne aling procedure and impurity concentration, (ii) the too low values of the fading rates predicted by the Randall-Wilkins parameters and (iii ) the unexpected high E and s values of peak 5 of LiF:Mg,Ti.