Difficulties encountered with some intermediate-atomic-number radiation-protection dosimeters irradiated on-phantom...

During a measurement-assurance study done with a particular type of radiation-protection dosimeter, we obtained unexpectedly poor results when several of the dosimeters were irradiated simultaneously on-phantom in a low-energy bremsstrahlung beam. The dosimeters incorporated low- and intermediate-atomic-number (Z) thermoluminescence dosimeter elements with the intermediate-Z elements under Pb filters. The response of the low-Z elements was in good agreement with the predicted response regardless of dosimeter location on the phantom, while agreement for the intermediate-Z elements was poor for some locations not coinciding with the phantom center. We have since studied the behavior of this dosimeter type and of four other types in detail by irradiating them, one at a time, on- and off-phantom, with their geometric centers displaced up to 10 cm along the vertical and horizontal axes through the center of the beam cross section. Two low-energy bremsstrahlung beams and a 137Cs gamma-ray beam were used. Variation of response to low-energy photons with on-phantom location was observed to some degree in all intermediate-Z dosimeter elements covered with high-Z filters. The reason for the observed dependence of response upon on-phantom location seems to be insufficient side shielding against phantom and filter albedo, as proved in one case, in which added shielding removed this dependence for the element immediately adjacent to the added shield. Since the effective dose equivalent to the human body is lower for lateral incidence of low-energy photons than for the same photons incident from the front or back, dosimeters incorporating high-Z filters without side shielding (and thus expected to exhibit the phenomenon) do not lend themselves to the determination of the effective dose equivalent or any quantity proportional to it by currently employed methods. A detailed study of the angular dependence of dosimeter response performed on a suitable phantom should aid in eliminating such inadequate dosimeters from practical use.