PNNL

Abstract

Monitoring changes in atmospheric radioxenon concentrations is a major tool in the detection of an underground nuclear explosion. Ground based systems like the Automated Radioxenon Sampler /Analyzer (ARSA), the Swedish Unattended Noble gas Analyzer (SAUNA) and the Automatic portable radiometer of isotopes Xe (ARIX), can collect and detect several radioxenon isotopes by processing and transferring samples into a high efficiency beta-gamma coincidence detector. The high efficiency beta-gamma coincidence detector makes these systems highly sensitive to the radioxenon isotopes 133Xe, 131mXe, 133mXe and 135Xe. The standard analysis uses regions of interest (ROI) to determine the amount of a particular radioxenon isotope present. The ROI method relies on the peaks of interest falling within energy limits of the ROI. Some potential problems inherent in this method are the reliance on stable detector gains and a fixed resolution for each energy peak. In addition, when a high activity sample is measured there will be more interference among the ROI, in particular within the 133Xe, 133mXe, and 131mXe regions. A solution to some of these problems can be obtained through spectral fitting of the data. Spectral fitting is simply the fitting of the peaks using known functions to determine the number and relative peak positions and widths. By knowing this information it is possible to determine which isotopes are present. Area under each peak can then be used to determine an overall concentration for each isotope. Using the areas of the peaks several key detector characteristics can be determined: efficiency, energy calibration, energy resolution and ratios between interfering isotopes (Radon daughters).