Limits of detection for special nuclear materials with an advanced high-pressure xenon detector and robust fitting analysis

The limits of detection for a special nuclear material of particular concern, 239Pu, using an advanced high pressure xenon detector (HPXe) and robust fitting analysis (RFA) were experimentally determined for a set of spectra with two different source strengths relative to the background, and for collection times ranging from 8 minutes to 7.5 seconds. The analysis of this data provides complementary information to recent results in which the same source was attenuated with varying thicknesses of lead shielding ranging from zero to four millimeters. The objective of these studies is to better understand the advantages and limitations of using this combination of technologies in nuclear portal monitoring systems to directly detect illicit trafficking of sensitive nuclear materials. HPXe detectors of the type used in this experiment are attractive candidates for portal monitoring systems because of their relative insensitivity to temperature changes and rough treatment, their full-energy peak efficiency over a wide range of energies, and their excellent energy resolution (1.9% at 662 keV with 3.0% efficiency and 1.3% at 1332 keV with 0.7% efficiency) without the need for cryo cooling. A disadvantage of HPXe detectors for field use has been their susceptibility to degraded resolution when subjected to acoustic influences, but significant progress recently has been made with HPXe vibrostability showing negligible degradation in resolution with acoustic noise up to 70 dB. The cylindrical HPXe detector used in the experiment was modeled after the one described in, and differed only in that it had one, not three, rectangular cells. It contained 1.73 liters of xenon gas with 0.3% H2 at a density of 0.25 g/cm3. In summary, the combination of a single 1.73-liter HPXe detector unit and an automated RFA analysis procedure appears to provide a reasonably reliable means by which unshielded amounts of 239Pu on the order of 1 gram can be detected with good confidence at a reasonable distance and collection time. A maximum reduced chi-square of 2.20 is an effective discriminator for reliable analyses and is rarely exceeded with surveys of 2 minutes or more, and infrequently exceeded with surveys of 7.5 seconds or more. If the reduced chi-square exceeds this threshold, the inspector need only continue the survey until a spectrum with sufficient chi-square statistics is obtained. At a distance of 73 cm with survey times as low as 15 seconds, the 1-gram plutonium source was consistently detected with a significance exceeding 4.0 in cases in which the chi-square criterion was met. At a distance of 200 cm with survey times as low as 2 minutes, with the exception of one of the six 2-minute spectra, the 1-gram plutonium source was consistently detected with a significance exceeding 4.0 in cases in which the chi-square criterion was met. While this work provides encouraging results for the detection of 239Pu in portal monitoring situations, further work is needed to understand the likelihood of false detection of 239Pu in spectra known not to contain 239Pu