Reproducibility study of (TLD-100) dosimeters using an automated and manual TLD reader for radiotherapy applications

The precision of the response of Harshow thermoluminescent dosimeters (TLD-100) is investigated for two Harshow TLD readers. The mean reproducibility for chips pre-readout annealed at 100 deg. C for 15 min, evaluated with the manual planchet reader 4500, is 0.61% (1 standard deviation) using pre-readout method. When chips are evaluated with the automated hot-gas reader 4500, reproducibility values are undoubtedly worse, mean reproducibility for numerically stabilized dosimeters being equal to 2.12%(1 standard deviation) using pre-readout method also. The pre-readout method was used for elimination of lower temperature peaks in order to obtain more stability. These results indicate that the automated hot-gas 4500 reader, or, at least, the instrument used for the present measurements, is not adequate for chips evaluation, if precise and accurate dosimetry is required. The difference in precision for manual planchet reader is apparently due to geometry inconsistencies in the orientation of the planchet to imperfect chips faces during readout, requiring careful and manual reproducible arrangement of the selected chips faces in contact with the manual reader planchet. In order to elucidate this point, the planchet support face of all chips dosimeters was identified with a small point made by pencil, reducing in this way geometrical variations