The active area shadow-shielding effect on detection efficiency of collimated broad energy germanium detectors

The ISOCS calibration, when utilized for a BEGe detector with a small angled collimator, produces inaccuracies of about 19% for gamma rays with energies greater than 0.4 MeV. Such a discrepancy is caused by the collimator algorithms currently utilized in the ISOCS software which, originally developed for HPGe detectors, are less suited for BEGe detectors. ISOCS's errors are due to the different crystal configurations of broad energy detectors compared to coaxial detectors, i.e. to a different importance of the active area portion obscured by the collimator. This work proposes some solutions for the problem, either using the ISOCS software or implementing a stochastic calibration procedure. In particular, the present work considers a virtual collimator that, maintaining its angular aperture, is capable of continuously enlarging its bottom collimator's aperture cone radius, to expose growing active area portions. In such a way two goals may be achieved: the mathematical characterization of ISOCS' errors and the minimization of observed errors by means of the stochastic calibration procedure. Different reference set-ups are considered in order to test source geometry effects, source materials and different detectors. In particular, a 220 L drum, a 2 m3 box filled with uniformly contaminated cellulose or PVC, and small BE3825 and large BE5030 Canberra detectors are considered. Detection efficiencies calculated by ISOCS software are compared against a completely stochastic MCNPX simulation procedure, that is unaffected by any algorithmic correction. MCNPX simulations demonstrate, when widening collimators' cone but maintaining the angular aperture unchanged, that ISOCS and MCNPX difference percentage between efficiency data points reduces, depending on energy, by more than 50%. This happens as far as the shadow-shielded portion of detector's active area reduces. © 2013 Elsevier B.V.