After the 9/11 incident, illegal smuggling of nuclear materials by terrorists has become a chief concern for homeland security. Currently, plastic scintillators serve as the first line of defense in detecting radiation. However, even the most sophisticated plastic scintillator can not distinguish between Chinese clay pots and dangerous nuclear materials. There is a critical need for development of a material with good spectroscopic capabilities and low cost that can be used on a large scale to detect and identify radioactive materials which can be used in a nuclear weapon.

High Pressure Xenon (HPXe) is ideally suited for rugged, field deployable radiation detectors for the purpose of a variety of security applications, because it offers extremely high resolution (an intrinsic 0.6% at 662 keV), no temperature sensitivity, excellent performance at or even significantly above room temperature, insusceptibility to neutron bombardment, and fast response due to high electron mobility. Moreover, HPXe requires no cooling system and is non-toxic.

Current limitations. While a number of groups have developed detectors using HPXe, none of the systems has come close in attaining the true intrinsic resolution of Xe [1,2,3]. The best published resolution at 662 keV for cylindrical HPXe systems is 2%, which corresponds to a roughly four-fold degradation of Xe’s intrinsic capability. Such systems utilize a Frisch grid in their designs, which corrects for the difference in signal response due to the cylindrical geometry [1,2,5]. However, the use of a Frisch grid substantially increases detector capacitance, which degrades energy resolution. It also makes the system highly sensitive to acoustic effects, which makes the detector unsuitable for rugged field use [1,4]. Thus, currently available HPXe systems are not competitive with other spectrometers, such as HPGe and NaI.

New detector design. We propose a cylindrical ionization detector design, as shown in the schematic below, that incorporates a light transparent window (BK-7), allowing measurement of both the primary and secondary scintillation light generated in xenon. Light signals are digitized, and an algorithm is employed that generates a complete record of the drift time of all electronic charges inside the detector. Using published data on electron drift speed in high pressure xenon, the radial distribution of charge is deduced. We propose that this technique reveals the complex distribution of primary charge produced by Compton and photoelectron tracks, and xenon fluorescence conversions. The charge distribution is used to correct the corresponding pulse height spectrum, and obtain near optimal energy resolution in the detector. Proportional Technologies, Inc has been recently awarded a patent (#6,486,468) for this unique method.

Prototype development & testing. A pilot detector, shown on the right, was constructed and tested to evaluate the technique. It consisted of a stainless steel (SS) tube, 50 mm in diameter (48 mm ID), containing xenon gas, with 0.3% H₂, at a pressure of 30 atm (0.20 g/cc). The anode diameter was 1.52 mm and the active length was 160 mm. One end of the detector incorporated a MgF₂ crystal window for light transmission. Signals induced in the anode were read with a carefully designed charge sensitive amplifier, matched to the low 7 pF capacitance of the detector. Light signals were sensed with a photomultiplier tube. Both charge and light signals were digitized with a 14-bit, 100 million samples/s data acquisition board.

Using the scintillation light technique outlined above, we estimated the radius of each photon interaction, then plotted it against the corresponding charge pulse height, as shown in Figure 1 below (left column). Both ²²Na and ⁶⁵Zn results obtained in separate acquisitions are shown together. A clear correlation is seen between energy and radius, in the form of narrow bands corresponding to the characteristic 511 keV and 1116 keV gammas emitted by each source. The Compton edge and fluorescence escape bands are also clearly visible.

Using a specially developed algorithm, the pulse height of each individual event was corrected, producing the plots shown in the right column panels. The projected energy spectra are shown for each source in Figure 2. The top panels show the uncorrected spectra, exhibiting the expected poor energy resolution. The middle panels show the light corrected spectra after an additional empirical correlation correction was applied to remove the slight curvature in the bands of Figure 1 (right column). The bottom panels show spectra for an approximately 50% subset of events having radii greater than 1.75 cm. The corrected energy resolution at 511 keV is 2.3% (2.1% for the radius limited data). The predicted ideal resolution, accounting for statistical and amplifier noise, is 1.5%. For ⁶⁵Zn (1116 keV), the corrected spectrum resolution is 1.5% for all events, and 1.35% for events limited to large radii. The predicted ideal resolution is 0.92%.

Figure 1. (left column) Mean radius to pulse height correlation for Zn-65 and Na-22  gamma
interactions in the prototype detector; (right column) Same as left, after pulse height correction,
as described in the text.

Figure 2. Uncorrected (top) and corrected (middle and bottom) energy spectra for a Na-22
(left) and Zn-65 (right) gamma source, collected in the prototype detector.

Xenon purification. We have constructed a gas purification system in order to satisfy the extremely stringent gas purity requirements for the detector. The system is used to purify xenon gas through the reduction of contaminants to parts-per-billion (PPB) levels, and to mix xenon and hydrogen to the required ratio inside the detector. The system (see picture below) mainly includes a large stainless steel reservoir to store xenon gas, a getter pump for xenon purification, a turbo vacuum pump, a small stainless steel reservoir for the pre-filling of purified HPXe before filling the detector, and a HALO H₂O gas analyzer.