1. Introduction

During underground nuclear tests, most solid radioactive products are contained within the explosive cavity, and only the produced radioactive inert gases can leak into the atmosphere ^[1,2]. ^131mXe, ¹³³Xe, ^133mXe, and ¹³⁵Xe have been chosen as gases to monitor nuclear testing ^[2]. The International Monitoring System has been established in various locations to detect and measure atmospheric concentrations of xenon radioisotopes (^131mXe, ¹³³Xe, ^133mXe, and ¹³⁵Xe) for nuclear weapon tests ^[3,4]. Table 1 lists the characteristic energies for the decay of xenon radioisotopes ^[5].

The beta–gamma coincidence technique is a well-known method to detect radioactive xenon radioisotopes ^[4]. During the last forty years, several radioxenon detectors have been designed, such as the Swedish Automatic Unit for Noble gas Acquisition (SAUNA) ^[6], the SPALAX unit ^[7], and the Automated Radioxenon Sampler and Analyzer (ARSA) ^[8,9]. Only the SPALAX unit uses a high-purity germanium detector. The SAUNA and the ARSA unit are both based on the beta–gamma coincidence technique.

In a typical detection system, a beta–gamma coincidence spectrum is constructed to locate three X-ray/gamma-ray energy lines, at 30, 81, and 250 keV ^[8]. It has been shown that using beta–gamma coincidence counting provides a 10³–10⁴-fold background reduction over standard gamma-ray spectroscopy ^[9].

A phoswich (phosphor sandwich) detector is a combination of more than two different scintillators that are optically coupled, and their scintillation light should be collected by a single photomultiplier tube so that different types of radiation can be detected simultaneously ^[10-15] and distinguished using the pulse shape discrimination method ^[16]. The Canadian CTBT (Comprehensive Test Ban Treaty) xenon radioisotope laboratory’s PhosWatch detector product package, PW5, consists of a phoswich detector, and Zhang et al. have performed some simulations and tests on it ^[17].

In this work, a semi-well-type phoswich detector was designed to obtain a higher detection efficiency. The Geant4 software was also used to analyze this detector in order to determine the optimal parameters.

2. Phoswich detector

The phoswich detector, shown in Fig.1, consists of a hollow plastic scintillator (BC-400) to detect beta and conversion electrons and a NaI:Tl crystal surrounding the BC-400 cylinder for gamma-ray and X-ray measurement. BC-400 is a plastic scintillator, so it is highly sensitive to beta radiation and much less sensitive to gamma or X radiation. Gamma rays and X-rays can penetrate BC-400 without considerable absorption. The thickness and diameter of the NaI:Tl crystal are both 5.00 cm. The well in the NaI:Tl crystal is ∅1.00 cm × 3.00 cm, and the wall of this well is made of quartz glass with a thickness of 0.20 cm. The BC-400 is a ∅1.00 cm × 3.00 cm cylinder with a ∅0.60 cm × 2.60 cm cavum in the center. This cavum is filled with radioactive carrier gas at a pressure of 1 atm. Table 2 shows the physical properties of the scintillators used in this phoswich detector. For this detector, the hollow BC-400 is filled with radioactive gas, so the detector has a very large solid angle of nearly 4π. Compared with the phoswich detector designed by the researchers at the University of Missouri-Columbia ^[18], this phoswich detector would have higher detection efficiencies.

Fig. 1Full-screen View

Schematic diagram of the phoswich detector discussed in this work

3. Geant4 analysis

Geant4 is a toolkit for simulating the passage of particles through matter. Its areas of application include high-energy, nuclear, and accelerator physics, as well as studies in medical and space science ^[19].

In this work, Geant4 analyses were used to simulate electron and gamma interactions in each scintillator layer of this phoswich detector. First, the designed semi-well-type phoswich detector was analyzed so that we could obtain the absolute efficiency of each detector layer. Second, BC-400 with a very thin Al layer on its inner walls was analyzed to determine the influence of the Al layer. Third, we also analyzed several NaI:Tl crystals with different thicknesses. Finally, the self-absorption effects of the carrier gas were analyzed in order to determine the carrier gas’s influence.

In the simulation, the doping materials are not modeled, as their concentrations are very low, and they have little effect on the simulation overall. The simulations typically model 10 million source photons or source electrons, so the relative errors in significant energy bins are 1% or less. Further, the entire beta spectrum is not described, just that of mono-energetic electrons.

4. Result

Gamma rays and X-rays can penetrate BC-400 without considerable absorption, their interactions are nearly all Compton scattering, so there will be only a Compton plateau and no full-energy peak. However, when gamma rays are detected using a NaI crystal, the photoelectric effect can occur and all of the energy can be deposited, generating the full-energy peak. Thus, when simulating gamma interactions in BC-400, we assume that the efficiencies are those of all of the events we detected, but we assume that in the NaI:Tl crystal, the efficiencies are only those of events in the gamma peaks. Figure 2 shows the gamma source peak efficiencies as a function of source photon energy in both the BC-400 and NaI:Tl crystals. It demonstrates that BC-400 is insensitive to gamma rays, as the efficiencies are all below 5%. The NaI:Tl crystal is most efficient (nearly 90%) for photons whose energies are approximately 100 keV. As the energy becomes larger, the gamma efficiency decreases rapidly, to 32% at 400 keV and 10% at 1000 keV. We think that the main cause of this decrease is an increase in the half-absorption thickness. As the photon energy becomes larger, more photons penetrate the NaI:Tl crystal without energy loss. A function has been given to describe the efficiencies. It is shown as follows:

Fig. 2Full-screen View

(Color online) Scintillator gamma ray source peak efficiency for BC-400 and NaI:Tl crystal

where E is the energy in keV, and NaI:Tl_efficiency is the source peak efficiency of the NaI:Tl crystal. If accurate gamma measurement can be obtained, we can use this function in reverse to obtain the activity.

Figure 3 shows the ratio of the BC-400 efficiency to that of NaI:Tl for various gamma energies. The ratio is approximately 5% for gamma energies below 200 keV, and it increases linearly for gamma energies between 200 and 1000 keV. We give the following function to describe this:

Fig. 3Full-screen View

(Color online) BC-400 to NaI:Tl efficiency and energy deposition ratios for various gamma energies

where E is the energy in keV, and NaI:Tl_ratio is the ratio of the BC-400 gamma efficiency to that of NaI:Tl. When the ratio increases, it is mainly because the gamma efficiency decreases rapidly as the photon energies become larger. The BC-400 to NaI:Tl energy deposition ratio for various energies is also shown in Fig.3. All the values are below 5%, indicating that BC-400 captures very little gamma energy and would not cause significant gamma attenuation.

Figure 4 shows the source detection efficiencies for electron interaction in BC-400 and the NaI:Tl crystal. For BC-400, the efficiency is nearly 100% when the electron energy is above 30 keV. However, below 30 keV, there is a sharp decrease as the electron energies become lower. This decrease is caused mainly by the self-absorption of the carrier gas. For the NaI:Tl crystal, the detection efficiencies are low at energies below 600 keV: 0 at 50 keV and 1.88% at 600 keV. However, when the electron energies are above 600 keV, there is a notable increase, to 11.37% at 700 keV and 56.50% at 1000 keV. This is because the BC-400 cannot absorb all the electron energy and the energy resulting from Bremsstrahlung radiation. Figure 5 shows the energy deposition ratios for electron interaction in the BC-400 and NaI:Tl crystal. This figure shows that when the electron energy is between 100 and 600 keV, all the energy is deposited in the BC-400, but when the energy is greater than 600 keV, some of the energy begins to be deposited in the NaI:Tl crystal. This just corresponds to the result in Fig.4.

Fig. 4Full-screen View

(Color online) Electron source efficiency for BC-400 and NaI:Tl crystal

Fig. 5Full-screen View

(Color online) Electron energy deposition for BC-400 and NaI:Tl crystal

During measurement, xenon will be absorbed on the inner surface of the BC-400, and this absorption will increase the detection background for subsequent gas samples. A thin aluminum layer can effectively reduce this effect. However, an aluminum layer would stop many low-energy electrons. Thus, a series of aluminum layers with different thicknesses ranging from 100 to 500 nm were simulated. Figure 6 shows the electron source efficiency for BC-400 with an aluminum layer, and Figure 7 shows the electron energy deposition for BC-400 with an aluminum layer. It can be seen that electrons with energies exceeding 50 keV would not be significantly absorbed by the aluminum layer. Further, as the electron energy decreases, the absorption becomes much more noticeable. Figure 8 shows the electron source efficiency for BC-400 with a 500 nm aluminum layer and without an aluminum layer after events below 50 keV are discarded. There is very little difference in these electron efficiencies. Thus, if we discard events whose energies are below 50 keV, the aluminum layer would not affect the electron detection.

Fig. 6Full-screen View

(Color online) Electron source efficiency for BC-400 with aluminum layer

Fig. 7Full-screen View

(Color online) Electron energy deposition for BC-400 with aluminum layer

Fig. 8Full-screen View

(Color online) Electron source efficiency for BC-400 after events below 50 keV are discarded

Changing the NaI:Tl thickness can affect only the phoswich detector’s gamma efficiency. Thus, NaI:Tl crystals with different thicknesses were analyzed. Figure 9 shows the photon efficiency of NaI:Tl crystals with different thicknesses. Beyond 200 keV, the gamma efficiency increases with increasing NaI:Tl thickness owing to the exponential nature of attenuation.

Fig. 9Full-screen View

(Color online) Photon efficiency of NaI:Tl crystals with different thicknesses

As low-energy electrons penetrate the carrier gas, a large portion of them would be stopped in it. Reducing the carrier gas density can reduce this loss, so different carrier gas pressures were simulated. Figure 10 shows the BC-400 electron efficiency under different carrier gas pressures. It can be seen that as the pressure decreases, the electron efficiency increases for electron energies below 40 keV. Beyond 40 keV, the efficiency is stable at 100%. Thus, the carrier gas can stop only electrons with energies below 40 keV. Further, if we discard events below 50 keV, it will not affect the efficiency.

Fig. 10Full-screen View

(Color online) BC-400 electron efficiency under different carrier gas pressures

5. Conclusion

This paper analyzed a double crystal phoswich detector that is used for simultaneous detection of beta and gamma radiation. A Geant4 model was established to obtain the beta and gamma efficiencies for different crystal thicknesses and particle energies. The carrier gas’s self-absorption effect and the absorption of electrons by an inner aluminum layer were also analyzed. In further research, this detector will be tested in order to determine the real features for beta–gamma coincidence detection.