Abstract
In low dose measurements the statistical nature of the light emission and light conversion as well as that of the photomultiplication sets an absolute limit for the accuracy of a particular TL reading. The standard deviation achieved with LiF pellets is greater by a factor of about 2.5 than the optimal value determined by using a light source with variable intensity. The statistical fluctuation of the dark current determines the lower detection limit of the particular read-out instrument. Above a certain dose, in the present experiments about 0.1 Gy, the effect of statistical fluctuations observed in the low dose range is negligible and the precision of the measurements is determined solely by the reproducibility or the statistical uncertainty for a repeated measurement of the same exposure.