1. Introduction

The market for radiation-based inspection systems is rapidly growing and is increasing by approximately 7% annually [1]. Moreover, related technologies have shown a rapid pace of advancement for applications in the fields of logistics, anti-terrorism, military, and security inspection [2–4]. Following the September 11 attacks in the United States in 2001, the paradigms of cargo inspection systems have changed from revenuers to anti-terrorism. Therefore, the importance of a technique that can determine the material shape and also acquire material information is growing [3–5]. While X-ray images supply material appearance or distinguish between organic or non-organic material [6], neutron images supply the material ingredient information for detecting drugs, explosives, or nuclear materials. The principle of radiation inspection systems is as follows: the generated radiation passes through the objective, and the image sensor pixel detects the information-containing radiation. The entire image can be acquired using the synthetically processed signal from each radiation sensor [1]. The current X-ray and neutron multi-radiation inspection system comprises two separate radiation generators and two imaging systems, as shown in Fig. 1(a). Each imaging module consists of a radiation sensor composed of a scintillator and photodetector (indicated in green color), as shown in the inset of Fig. 1(a). In this case, a single-sided silicon photodetector (SSSP) is coupled with the scintillator [7–11]. If two radiation generators and two imaging modules are merged into one generator and one imaging system, the fabrication cost and system volume could be significantly reduced, as shown in Fig. 1(b). For this merged system, a simultaneous X-ray and neutron detecting radiation sensor module can be developed [12]. The concept of a simultaneous detecting radiation sensor is shown in the inset of Fig. 1(b), which is composed of a double-sided silicon photodetector (DSSP) and different scintillators on the opposite side of the photodetector (indicated in green color). The X-ray scintillators are coupled to the upside photodetector, and neutron scintillators are coupled to the bottom side photodetector. The detection timing should coincide with the X-ray/neutron pulse generator [13] that can separately detect each type of radiation. This technology requires a double-sided photodetector fabrication process technology with one silicon wafer for each scintillator coupling.

 

Fig. 1. (a) Current X-ray/neutron multi-radiation inspection system comprising two separate radiation generators and two imaging systems. Inset of Fig. 1(a) shows the radiation imaging module that is composed of scintillator and single-side silicon photodetector (SSSP). (b) X-ray/neutron simultaneous detecting radiation inspection system. The inset of Fig. 1(b) shows the radiation imaging module composed of double-sided silicon photodetector (DSSP) and different scintillators on the opposite side of the photodetector (indicated in green).

Download Full Size | PDF

In this study, a simultaneous X-ray and neutron detecting radiation sensor module was developed based on a DSSP coupled with a scintillator for each radiation. Each coupled sensor was mounted on a specially designed printed circuit board (PCB), and an irradiation test with different materials was conducted. The X-ray imaging module feasibility was confirmed by an X-ray transmission test by comparing the calculated value with the experimental value for different materials. The neutron imaging module feasibility was confirmed by counting the neutron energy peaks. Our results confirm the successful combination of two imaging modules into one imaging system, which can help implement a cost-efficient and lightweight portable multi-radiation security inspection.

The remainder of this paper is organized as follows. In Section 2, the fabrication process is described. In Section 3, the electrical and optical characteristics of the fabricated DSSP and the X-ray/neutron irradiation test result are described. Finally, the summary and a future work plan are presented in Section 4.

2. Experimental

The DSSP was designed by combining the PN junction photodetector formed on both sides of the silicon wafer for different types of scintillator couplings. The upside photodetector is used for X-ray scintillator coupling and the bottom side for neutron scintillator coupling. The active area of the photodetector was designed as an array type for minimizing the reverse leakage current and capacitance [1]. Each sensor chip is designed as a four-channel array photodetector. The total dimension of the four-channel array silicon photodetector is 35 mm ×10 mm, and one cell dimension is 35 mm × 2.5 mm. Each channel has edge protection and a guard ring around the active area that reduces surface leakage [14].

The fabrication process is illustrated in Fig. 2(a), showing the 675 µm thickness of the double-side-polished n-type silicon substrate with (100) orientation. Following the initial SC1 cleaning, as step 1, 500 nm of a SiO₂ layer was formed by wet oxidation at 950 °C. Owing to double-sided fabrication, all the following processes were conducted upside down sequentially. The wafer process is conducted on one side, whereas the other side of the wafer is protected by the hard-baked photoresist. Following photolithography patterning and oxide removal using a wet etch process, as step 2, phosphorous implantation and 950-°C thermal activation were conducted to form a channel stop region. After the edge protection and opening of the guard ring area using the second photolithography and oxide wet etching, as step 3, a 60-keV ¹¹B implantation was performed for p + deep well structure formation followed by a 1,100-°C thermal activation. The third photolithography was conducted for an open active area preventing oxide etching. For an active area formation, as step 4, BF₂ was implanted using a 40-keV energy and activated at 900 °C. A 500-nm protection and insulation SiO₂ layer was deposited by plasma-enhanced chemical vapor deposition (PECVD). Following the contact hole formation realized by the fourth lithography and the etching process, the fifth metallization lithography was performed. As step 5, Al/Au was deposited using e-beam evaporation equipment, followed by the lift-off process. An active window area was subsequently opened by the sixth lithography, after which as step 6, approximately 40 nm of a SiO₂ anti-reflection and passivation layer was deposited by PECVD [15], followed by post-deposition annealing. Finally, the seventh lithography was conducted to open a metal pad. The photograph of the fabricated DSSP is shown in Fig. 2(b), and the bottom side devices can be seen in a mirror. For the PCB packaging, a wafer-level device was cut into a 2 × 2 array single chip (35 mm × 10 mm) using a dicing machine. Therefore, the total dimension was 70 mm × 20 mm for size matching using the scintillator.

 

Fig. 2. (a) Fabrication process flow, photograph of fabricated DSSP (b) wafer and (c) single chip.

Download Full Size | PDF

3. Results and discussion

The emission peaks of cadmium tungstate (CWO) and BC-430 scintillators are 475 nm and 580 nm [16] for which the penetration depths in silicon are 800 nm and 2.05 µm [17], respectively. The p ⁺-n junction depth should be maintained within 800 nm from the surface of the device. The junction depth was controlled based on the implantation power and thermal activation condition, determining the p ⁺-layer thickness. By controlling the implantation power and the thermal activation process, the p ⁺-layer thickness was ∼200 nm (thickness of Boron atom concentration below 5 × 10⁻¹⁷), as confirmed by secondary ion mass spectroscopy [1].

The electrical measurement of each channel is performed using a semiconductor parameter analyzer with a probe station in a dark box. Figure 3(a) shows the current (I)–voltage (V) characteristics of each channel of the DSSP. The reverse leakage current of each channel is $- $10 nA at $- $5 V, which is slightly larger than that of the SSSP fabricated under the same processed condition with the value of -8 nA at -5 V. Increased leakage current is considered to be the primary cause of the contamination of the active layer owing to photoresist tearing and contamination due to interference with the process equipment on the opposite side during the double-sided manufacturing process. This increased the surface leakage current component of the device. Thus, process equipment modification and process optimization are required to further reduce the leakage current. Figure 3(b) shows the spectral response of the fabricated DSSP using a spectrophotometer. The photodetector yields a responsivity of 0.24 A/W at a CWO scintillator wavelength of 475 nm and 0.30 A/W at a BC430 scintillator wavelength of 580 nm.

 

Fig. 3. (a) Current (I)–voltage (V) characteristics of each channel of DSSP and inset is the top view of the DSSP (b) Spectral response of the fabricated DSSP using spectrophotometer. The photodetector yields a responsivity of 0.24 A/W at a CWO scintillator wavelength of 475 nm and 0.30 A/W at a BC430-scintillator wavelength of 580 nm.

Download Full Size | PDF

After characterization, each DSSP chip was physically and electrically mounted to the specially designed PCB followed by scintillator coupling, constituting one module of the radiation imaging system. The CWO scintillator was coupled to the upper side of the DSSP chips for X-ray detection, and BC-430 was coupled to the lower side of the DSSP chips for neutron detection, as shown in Fig. 4(a)–(d). To verify the performance of the scintillator-coupled DSSP chips as an image sensor, the fabricated sensor was equipped with a readout system [18] and tested under external light via an imaging feasibility test. All the scintillator-coupled DSSPs were covered such that light could not penetrate, except one, as shown in the inset of Fig. 4(e). In Fig. 4(e), the lines represent the images obtained using the scintillator-coupled DSSP chips. The light-induced photo-current generation results in a voltage drop of the readout pulse, and the white line pattern is acquired using a purpose specific imaging software [18]. If all of the scintillator coupled DSSP chips were exposed to light, the line patterns would be all white. In this test, just one DSSP was exposed to light, so one thick white line and side narrow lines (light penetrate to next pixel) were patterned at the software.

 

Fig. 4. Photograph of (a) top view of the physically and electrically mounted DSSP devices to PCB, (b) top view of the CWO scintillator coupled to the upper side of DSSP devices for X-ray detection, (c) top view of the BC-430 scintillator coupled to the lower side of DSSP devices for neutron detection, and (d) side view of the radiation imaging module. (e) Imaging test results of readout connected developed radiation imaging module under external light. The line pattern is noticeably observed using imaging software.

Download Full Size | PDF

To prove the feasibility of the multi-radiation imaging module, the magnitudes of the detection signal for each radiation source (X-ray and neutrons) and sample were compared via the irradiation test after placing lead (20 cm × 25 cm × 4 cm) and polyethylene blocks (20 cm × 25 cm × 15 cm) in front of the imaging module. The thicknesses of the phantoms of both materials were determined using a Geant4 tool, and a 10-% transmittance was realized for a 6-MeV X-ray and 14-MeV neutron [19]. Figure 5 shows the transmission ratio of the calculated (red square) and measured (black dot) values of the 6-MeV X-rays that penetrate the 20-cm-thick lead and polyethylene inside the air cargo. The transmittance of an X-ray is calculated using the following equation related to attenuation: ${I_{trans}} = {I_0}{e^{ - \mu t}}$, where ${I_{trans}}$ is the number of transmitted intensities, ${I_0}$ is the number of incident photon intensities, $\mu $ is the linear attenuation coefficient $({\textrm{c}{\textrm{m}^{ - 1}}} )$, and t is the thickness of the attenuator $({\textrm{cm}} )$. The calculations show that when no sample is placed, that is, in air, the transmission rate of an X-ray is 100%. However, after penetration through lead and polyethylene, the values are 11.81% and 58.94%, respectively. The measured transmittance rates of lead and polyethylene were 8.88% and 53.34%, respectively. They were similar to the linear attenuation coefficients based on the calculated value. Notably, the experimental results show a similar tendency to the calculated value, indicating that the developed module is suitable for X-ray imaging applications. The lower transmission rate of the measured value than the calculated value is because of the loss of light detection in the photodiode. According to the aforementioned results, the developed detection module confirmed the feasibility of X-ray imaging. Unlike X-ray signal processing, which integrates the charges generated by the photodiode for a certain period, the neutron signal processing module is configured to count only the peak energy that exceeds the discrimination level. Therefore, instead of a transmission-rate test with different samples, a neutron-spectrum-acquisition test was conducted. Figure 6 shows the 14-MeV neutron spectrum acquired by the developed detecting module. Notably, a neutron peak is noticeably observed. Further experimentation is required to distinguish the material information with sufficient neutron flux.

 

Fig. 5. Transmission ratio of the calculated (red square) and measured (black dot) value of the 6-MeV X-rays that penetrate the 20-cm thick lead and polyethylene inside the air cargo.

Download Full Size | PDF

 

Fig. 6. 14-MeV neutron spectrum acquired using the developed detecting module.

Download Full Size | PDF

4. Conclusion

In this study, a double-sided silicon-photodetector-based X-ray and neutron simultaneous X-ray and neutron detecting radiation sensor module was fabricated, and the feasibility was tested for multi-radiation imaging. Firstly, a DSSP was fabricated via a dual (top-bottom) align photolithography process along with the conventional SSSP fabrication process. At a sensor operation voltage of $- $5 V, the reverse leakage current was approximately $- $10 nA. The primary reason for the increased leakage current compared with that of the SSSP is attributed to the contamination and photoresist torn-off during the double-sided fabrication process. Therefore, process equipment modification and process optimization were required to reduce the leakage current. The spectral responsivities were 0.24 and 0.30 A/W for the scintillator emission at wavelengths of 475 and 580 nm, respectively. Specially designed PCB-mounted DSSPs were equipped with readout systems and tested under external light via a radiation imaging feasibility test. A line pattern image was successfully acquired using a CWO-coupled Si PIN sensor with self-developed readout and data acquisition systems. Finally, an X-ray and neutron irradiation test was conducted to verify the feasibility of the developed multi-radiation imaging module. The X-ray transmission test results with different materials were comparable to the calculated values, indicating the suitability of the developed module for X-ray imaging applications. In the neutron irradiation test, a 14-MeV neutron spectrum was acquired by the developed detecting module, and a neutron peak was noticeably observed. This proves that the fabricated sensor operated appropriately as an image sensor for multi-radiation imaging systems. In the future, we will use a material discrimination test with neutron irradiation. Our results confirm the successful combination of two radiation generators and two imaging modules into one generator and one imaging system, which can help in implementing a cost-efficient and lightweight portable multi-radiation security inspection.