SUCCESS AND FAILURE OF THE RANDALL-WILKINS MODEL FOR THERMOLUMINESCENCE IN LIF(TLD)

Computerised glow curve analysis has been used to determine the trappi ng parameters of the main peaks of LiF(TLD-100,TLD-600,TLD-700) origin ating from 21 different production batches. There are striking differe nces in sensitivity between the batches. However, the trapping paramet ers of peaks 2, 3, 4 and 5 are (with a few exceptions) fairly similar. In two batches an extra peak (5a) has been found which is not present in the other batches investigated. It appears that the Randall-Wilkin s (RW) model yields a glow peak shape which perfectly fits experimenta l measured glow curves of LiF(TLD). However, experimentally determined fading rates of peak 2 (in ten batches) disagree by a factor 24 with the rates calculated from the E and s values from the glow curve fitti ng. Furthermore, the RW model fails to explain why peak areas and acti vation energies (determined by first-order glow curve analysis) increa se as the heating rate increases from 0.2 K.s-1 up to 6 K.s-1. These d iscrepancies are ascribed to defect reactions during readout. An inter acting defect model is introduced. This model gives an explanation of why the fitting parameters involved in first order 'glow curve analysi s are not the physically correct parameters describing the TL process involved. This means that first order kinetics glow curve analysis of LiF(TLD) gives us a phenomenological description of the peaks but no i mmediate insight and understanding.