Year
2008
Abstract
The interdiction of illicit radioactive materials relies on detecting the emissions from the material when passed in front of or through a radiation monitor. These portal monitors must be sensitive enough to detect and identify a specified (ANSI N42.38) quantity of material. The ability to identify, rather than just detect changes in activity, is important in order to differentiate between benign materials and threat materials. Clearing benign materials has a high cost in time and money. The identification depends on the signal-tonoise ratio, not the gross efficiency, of the detector. In low- or no-resolution detectors the background (noise) is high resulting in a low signal-to-noise ratio and the actual background can be affected by the object in the portal causing real increases in the background to go undetected. In a high-resolution monitor, the background is determined on the actual object spectrum, avoiding this problem. Portal monitors typically have 5 to 45 seconds to collect the gamma-ray spectrum and then 2 to 5 seconds to analyze the spectrum and give a result. The determination if a gamma-ray peak is present in the spectrum is usually done by calculating a quality factor and then comparing this quality factor with a threshold. The quality factor depends on the net peak area and the background under the peak in some way. The net peak area and the background depend on the full-energy peak efficiency and the detector resolution. The short collection time means there are few counts in the spectrum and the variation in the calculated values is high. The quality factor threshold for positive identification must be set so that the number of false identifications is less than 1 in 1000 measurements (ANSI N42.38). This threshold value also determines the minimum identifiable activity for the given background level. To measure this threshold, a pedestrian portal was constructed and the quality factor was measured for 10000 occupancies for several different background levels. The quality factor was also measured for 10000 occupancies with a small 133Ba source. The results show the expected distribution for high background, but have some granularity at the lowest background. The results also show that the threshold must be set higher than expected for the lowest backgrounds in order to accomplish the 1 in 1000 false rate. The minimum identifiable activity for 133Ba in the different backgrounds at the 1:1000 FAR (False alarm rate) setting is shown and is significantly lower than the limit defined in the standard. It is clear from these results that the high signal-to-noise ratio is more important than gross efficiency for radiation portal monitors required to interdict one group of radionuclides while clearing a second group for onward transportation.