Accurate efficiency calibration and properties of semiconductor detectors for low-energy photons

Abstract

Methods are given for investigating the properties and determining precision detection efficiencies of semiconductor Si(Li), Ge(Li), and intrinsic Ge photon detectors in the energy range 3 to 100 keV. The accuracy with which the efficiency curves can be established depends in part on the individual detector; an accuracy of ≈±3% between 6 and 60 keV could be achieved for the intrinsic Ge detector due to the relative simplicity of its sensitive volume and its low tail/peak ratio. In general, accuracies of 3 to 5% between 5 and 100 keV are achieved. Four typical detectors have been investigated extensively and a theoretical physical model developed for calculation of detection efficiencies to assist in the interpolation between the experimental points. For each detector, the procedure involves three separate steps: (1) accurate experimental measurement of the absolute detection efficiency at a standard geometry with calibrated photon-emitting radioactive standards; (2) accurate determination of the parameters required to predict from the theoretical physical model the absolute detection efficiency, such as source-to-detector distance, thickness of the gold and the Si or Ge deadlayers, the sensitive depletion depth, the sensitive detection area, collimation geometry, and charge collection efficiency; and (3) a fit of the experimental points with the theoretical model to minimize errors in interpolating between the points and to reveal anomalies. Results of narrow-beam photon scanning across the face of the detectors are presented and tail/peak ratios are given and discussed. A useful compilation is included of the photon energies and latest values of the emission rates per decay for the following radioactive standards: 241Am, 57Co, 60Co, 137Cs, 203Hg, 54Mn, 22Na, 88Y and 182Ta. In addition to these sources, others were calibrated by a new multiwire proportional counter technique or by application of the conversion electron-X-ray coincidence (CEX) technique. Rather substantial changes (≈20%) in efficiency and in tail/peak ratios were observed in some detectors with ageing, or, in the case of a Si(Li) detector, after a one-hour warmup cycle. It is concluded that each detector requires an individual determination of efficiency, as no two detectors are alike in all of their properties even though their nominal specifications may be identical, and that such efficiency determinations must be repeated, since changes occur with time, temperature, and other treatments. The mean life of cooled semiconductor X-ray detectors against failure appears to be approximately 18 months.