Characterization and calibration of an n-type HPGe detector for in situ low energy gamma spectrometry

In case of an incident or accident with deposited or dispersed radioactive material, e.g. after a release from a nuclear power plant accident, it is important to obtain measurements results fast. This can be offered already at the site using in situ gamma spectrometry with e.g. HPGe detectors. In order to ensure that the measurement system delivers correct results, within measurement uncertainty, the detector needs to be calibrated for the specific energy range and sample geometry. In the semi-empirical calibration, which has been used in this work, data from measurements are used in combination with Monte Carlo calculations of a detector model. Monte Carlo calculations require knowledge of detector dimensions, e.g. the distance between crystal and detector window. Manufacturers commonly supply dimensions for a given detector type. However, individual detector variations are usually not readily available. To obtain a more accurate model, dimensions inside of HPGe detectors have, in previous studies, been measured by using X-ray imaging, mainly with a room tempered detector. To our knowledge, there has been no thorough investigation of the effect of cooling the detector to operation temperatures, which may help to explain the discrepancies between studies involving Monte Carlo detector models and empirical data. This report describes the characterization and semi-empirical calibration of an n-type HPGe detector which will be used for in situ gamma spectrometry, and, more specifically, for measurements of low energy photons. Monte Carlo models and calculations were made using GESPECOR and MCNP. Computed Tomography (CT) high-resolution imaging was used to acquire a more accurate detector model and to investigate if detector parameters change when HPGe detectors are cooled to operation temperature (77-100 K). It was found that the distance between crystal and detector window increases when HPGe detectors are cooled; (0.376±0.070) mm for an n-type detector and (0.40±0.15) mm for a p-type. Monte Carlo calculations indicate that the crystal-to-window differences have a significant impact on the detector efficiency for close geometries. For the optimized MCNP detector model to fit as well as possible (within the uncertainty) with empirical data, parameters were changed e.g. the inner dead layer was increased by 0.8 mm and the crystal-to-window distance was increased by 2.5 mm compared to manufacturer specifications.