Plutonium dioxide particle imaging using a high-resolution alpha imager for radiation protection

The internal exposure of workers who inhale plutonium dioxide particles in nuclear facilities is a crucial matter for human protection from radiation. To determine the activity median aerodynamic diameter values at the working sites of nuclear facilities in real time, we developed a high-resolution alpha imager using a ZnS(Ag) scintillator sheet, an optical microscope, and an electron-multiplying charge-coupled device camera. Then, we designed and applied a setup to measure a plutonium dioxide particle and identify the locations of the individual alpha particles in real time. Employing a Gaussian fitting, we evaluated the average spatial resolution of the multiple alpha particles was evaluated to be 16.2 ± 2.2 μmFWHM with a zoom range of 5 ×. Also, the spatial resolution for the plutonium dioxide particle was 302.7 ± 4.6 µmFWHM due to the distance between the plutonium dioxide particle and the ZnS(Ag) scintillator. The influence of beta particles was negligible, and alpha particles were discernible in the alpha–beta particle contamination. The equivalent volume diameter of the plutonium dioxide particle was calculated from the measured count rate. These results indicate that the developed alpha imager is effective in the plutonium dioxide particle measurements at the working sites of nuclear facilities for internal exposure dose evaluation.


Scientific Reports
| (2021) 11:5948 | https://doi.org/10.1038/s41598-021-84515-z www.nature.com/scientificreports/ Alpha particle imaging detectors using position sensitive photomultiplier tubes (PSPMTs) and silicon photomultipliers (SiPMs) have also been developed in the past [12][13][14][15] . They can obtain alpha particle counts in real-time and are used to estimate PuO 2 particle diameters. However, it is desirable to acquire optical images and alpha particle imaging simultaneously. The optical images are useful information for the subsequent analysis using different methods if the exact location of the PuO 2 particles is to be determined.
On the other hand, high-resolution alpha imagers have been considered for clinical applications, such as the targeted alpha therapy (TAT), mainly described with 211 At being injected into a patient. Some high-resolution alpha imagers have recently been developed for the TAT. Bäck et al. have developed an alpha imager that combines a high-sensitivity charged-coupled device (CCD) camera with a ZnS(Ag) sheet. This alpha imager can be applied to identify the alpha particle distribution of 211 At in the organs of mice 16 . The spatial resolution of this imager was 35 ± 11 μm. Miller et al. developed an alpha camera named iQID camera, and the spatial resolution of ∼ 20 µm FWHM was achieved 17 .
Yamamoto et al. have developed a high-resolution alpha imager using a CCD camera and an optical fiber structure scintillator plate, which can visualize alpha particle trajectories with an ultrahigh spatial resolution (~ 11 µm) 18 .
Pratx et al. developed radioluminescence microscopy using a CdWO 4 scintillator plate and a microscope to capture the individual beta particle track 19 . Radioluminescence microscopy is a new technique for visualizing radionuclides in live cells and acquires a sequence of frames with short exposure to count the ionization tracks.
Kurosawa et al. have developed a high-resolution alpha imager by combining a complementary metal-oxide-semiconductor (CMOS) camera with a microscope 20 . This alpha imager can choose its zoom range and is capable of an image resolution of 2.1 µm for optical imaging. As the spatial resolution for alpha particles is better than 100 µm, which is a better resolution than those of the conventional alpha imagers used for AMAD value measurements at the workplaces of nuclear facilities, such as the IP detectors, an alpha imager using a camera and a microscope can be used to determine in real time the diameters of PuO 2 particles.
Alpha and beta contamination existed in the FDNPS site. For use at the FDNPS site, the alpha imager must be sensitive to alpha particles only. In this study, we develop a high-resolution alpha imager using a CCD camera and an optical microscope based on previous studies and applies the technique to measure an actual PuO 2 particle.

Results
Optical images. Figure 1 presents the optical images of the calibration sample with the zoom ranges 5 ×, 10 ×, and 20 ×. The slit width of the calibration sample was 25 µm. The relation between the image resolution and FOV with the same zoom ranges is presented in Table 1. A trade-off could be observed between the image resolution and FOV. For instance, at the zoom range of 5 ×, the resolution and FOV were 3.28 µm/pixel and 1679.4 µm × 1679.4 µm, respectively, whereas when 20 × was used, the corresponding values were 0.81 µm/pixel and 412.9 µm × 412.9 µm. Radiation imaging. Alpha source imaging. Figure 2 presents the image of an alpha particle measured using the high-resolution alpha imager with a zoom range of 5 × and along with the particle's intensity profile. The full width at half maximum (FWHM) of the alpha particle intensity profile shown in Fig. 2(b) was evaluated us-  www.nature.com/scientificreports/ ing the Gaussian fitting, at 15.9 ± 0.7 μm with a zoom range of 5 × . The average spatial resolution of the multiple alpha particles was evaluated to be 16.2 ± 2.2 μm with a zoom range of 5 ×. Figure 3a presents the image of an alpha particle from the active area of the 241 Am alpha source obtained using the "Sequence mode" with 50-ms intervals. The entire area of the image was the active area of the 241 Am alpha source. The alpha particles were confirmed at different locations within 50 ms. Similarly, Fig. 3b presents the image of alpha particles obtained using the "Acquire mode" for 10 min. Here the alpha particles were distributed all over the image.
Beta source imaging. Figure 4a presents the image from the active area of the 90 Sr/ 90 Y beta source obtained using the sequence mode with 50-ms intervals. The entire area of the image was the active area of the 90 Sr/ 90 Y beta source. No particle was confirmed within 50 ms. Figure 4b presents the image obtained using the acquire mode for 10 min. Similarly, no particle could be observed on the image. These results indicate that the influence of beta particles is negligible and that alpha particles can be distinguished against them.
PuO 2 particle. Figure 5a presents the alpha particles emitted from the PuO 2 particle with the 50-ms intervals of the sequence mode. The locations of the individual alpha spots can be identified in real time. Figure 5b presents the alpha particles emitted from the PuO 2 particle using the acquire mode for 30 min. The number of alpha particles corresponded with that measured using the commercial ZnS(Ag) scintillation counter, at a detection efficiency (D eff ) of 99.3% based on Eq. (1). Figure 5c shows a superimposition of the optical image and alpha particle image, and Fig. 5d shows the optical image with an arrow showing the position of the PuO 2 particle. Figure 5e shows an intensity profile of the alpha particle emitted from a plutonium particle. The spatial resolution for the  Radon-222 progeny. Figure 6 shows the alpha particles emitted from the 222 Rn progeny using the acquired mode for 30 min. Alpha particles are distributed throughout the image. The distribution of alpha particles of 222 Rn was different from the PuO 2 particles, which correlated well with the results obtained from other alpha imaging detectors 21 .
Evaluation of the equivalent volume diameter (d e ) of a PuO 2 particle. Figure 7 presents the conversion factor (CF) curve used for evaluating the diameter of a PuO 2 particle as a function of the count rate calculated using the Monte Carlo simulation. The equation of the curve can be approximated as y = 3.84 × (d e
Where the alpha counts detected in the experiment were 214.3 ± 5.2 cpm, the d e of the PuO 2 particle was evaluated to be 6.9 μm. Accordingly, the equivalent aerodynamic diameter (d ae ) was determined to be 17.4 μm using Eq. (6).

Discussion
We have developed the high-resolution alpha imager, measured the actual PuO 2 particle, and calculated its d e and d ae . At a zoom range of 5 ×, the resolution was 3.28 µm/pixel, but with a larger FOV of 1679.4 × 1679.4 µm. Thus, lower zoom ranges like 5 × appropriate the identification of the locations of PuO 2 particles on a sample.
Using the sequence mode allowed the measurement of individual alpha particle spots with a high spatial resolution (~ 17 µm FWHM). This feature is useful in evaluating the activity and diameter of PuO 2 particles. Moreover, the electron-multiplying charged-coupled device (EMCCD) camera is capable of changing the exposure time intervals from 13.9 ms to 2 h. The short time interval measurement helps avoid the overlapping of multiple alpha particles, as well as accurately measures the number of alpha particle spots. Conventional detectors, such as the CR-39 and IP detectors, are passive detectors that need long exposure times; if the exposure time is not appropriate, then multiple alpha particles will overlap during the measurement. The "no-fading effect, " as in the case of IP detectors, was another advantage of our developed alpha imager.
Further, the proposed alpha imager demonstrates no beta sensitivity, even with the measurement of the 1-MBq 90 Sr-90 Y beta source. As the ZnS(Ag) scintillator was only 3.25-mg/cm 2 (approximately 8-μm) thick, there were a few possibilities that beta particles absorb their energy in the scintillator. For example, the range of a 5.5-meV alpha particle is approximately 10 mg/cm 2 in the ZnS(Ag) scintillator. In the FDNPS site, where beta contamination activities were much higher than those of alpha, our alpha imager could prove useful in measuring the alpha-emitting particles.
Moreover, the measured d e and d ae values of the PuO 2 particle were 6.9 and 17.4 μm, respectively, the latter being in the 1.0-28.0-μm range of the previous IP detector measurements 22 . The calculated CF values for PuO 2 particles were calculated in the range d e of 1-10 μm (d ae of 2.7-25.4 μm). The AMAD value of 5 μm is representative of workplace aerosol value 23 . The value of 5 μm is within this range.
In a facility like the one from which we obtained our samples, the PuO 2 particles generated by the dismantling of process equipment were measured using a cascade impactor 24 . The results were distributed in the range of 1-11 μm measured using a cascade impactor. Our measured results (d e , equivalent particle size = 6.9 μm) are well within that range.
The abovementioned methods are useful in determining the AMAD values in real time at the working site of nuclear facilities, which will help evaluate the internal exposure dose of workers instantaneously in case of internal exposure accidents. and (e) intensity profile of the alpha particle emitted from a plutonium particle. The arrows in (a) show the individual alpha particle locations. The arrows in (d) show the PuO 2 particle location. The images of using the "Sequence mode" with 50-ms intervals can be found as Supplementary Fig. S3 online.

Conclusions
The proposed high-resolution alpha imager is a combination of a ZnS(Ag) scintillator sheet, an optical microscope, and an EMCCD camera. The alpha imager is capable of visualizing individual alpha particle spots in real time with a high spatial resolution (~ 16.2 ± 2.2 μm with a zoom range of 5 ×). As the influence of beta particles was found to be negligible, alpha-emitting particles can be distinguished in a mixed alpha and beta contamination environment, as in any FDNPS sites. The values of the d e and d ae of the PuO 2 particle were calculated from the measured count rate. Based on these results, the developed alpha imager will be useful for PuO 2 particle measurements at the working sites of nuclear facilities for instantaneous internal exposure dose evaluation.

Methods
High-resolution alpha imager. Figure 8 presents a schematic drawing of the high-resolution alpha imager developed using a ZnS(Ag) scintillator sheet (EJ-440, Eljen Technology, TX, USA), an optical microscope (BX53MRF-S, Olympus Corp., Tokyo, Japan), and an EMCCD camera (ImagEM X2, Hamamatsu Photonics k.k., Shizuoka, Japan) (Fig. 9). The thickness of the ZnS(Ag) scintillator was 3.25 mg/cm 2 (approximately 8 μm). The EMCCD camera was mounted on the optical microscope to capture the images of scintillation light or optical light. Output signals from the EMCCD camera were transferred to a PC using an IEEE1394 cable. A data acquisition software with "Acquire mode, " to identify a time-integrated image, and "Sequence mode, " to identify a series of single images, was used. The CCD camera was automatically cooled at − 65 °C after it was switched on to reduce thermal noise. The optical microscope was covered with a black curtain to block unwanted light from the outside.  Radiation imaging. Alpha source imaging. We used the alpha imager to measure 5.5 meV alpha particles emitted from the 241 Am alpha source. A schematic drawing of the alpha source imaging is presented in Fig. 10. The source has a 5-mm-diameter active area and a 3.37-kBq activity. Alpha particles deposited their energies in the scintillator, and the light emitted from the scintillator that passed through the lens was captured by the CCD camera.
Beta source imaging. Because alpha contamination is sometimes accompanied by beta contamination and needs to be distinguished, it is important to evaluate the beta sensitivity to the high-resolution alpha imager. Figure 11 presents the schematic drawing of the beta source imaging. Using the alpha imager, we measured a  www.nature.com/scientificreports/ 1-MBq 90 Sr − 90 Y beta source with a 10-mm-diameter active area. Also, the ZnS(Ag) scintillator was placed onto the beta source.
PuO 2 particle imaging. Using the same alpha imager, we measured a PuO 2 particle obtained from a MOX fuel facility. Figure 12 presents a schematic drawing of the PuO 2 particle imaging. A single PuO 2 particle was attached on a flat sheet and covered with a thin-film sheet. The PuO 2 particle was then covered with a thin polyethylene film. The alpha source and the plutonium particle were set on the X-Y-Z stage. Subsequently, the ZnS(Ag) scintillator sheet was placed on top of the thin polyethylene film. Moreover, we measured the same PuO 2 particle using a commercial ZnS(Ag) scintillation counter (JDC-817, Hitachi Aloka Medical, Tokyo, Japan), having a 100% detection efficiency for alpha particles with a 2 π angle. The detection efficiency of the alpha imager was evaluated by comparing with the measured count rate using the commercial ZnS(Ag) scintillation counter according to the following equation: (1) D eff = C alpha imager C ZnS(Ag) counter × 100   Radon-222 imaging. 222 Rn progeny was also measured using the alpha imager. Figure 13 shows a schematic drawing of 222 Rn progeny imaging. The radon progeny were collected using an air sampler (DSM-361, Hitachi Aloka Medical, Ltd. Mitaka, Tokyo, Japan) through a 47-mm diameter membrane filter (Toyo Roshi Kaisha, Ltd. Tokyo, Japan). Air sampling was performed in an environment with a 222 Rn concentration of 200 Bq/m 3 for ~ 4 h, and filter paper was collected after sampling. The membrane filter was fixed on a glass plate with a diameter of 50 mm × 50 mm, and the ZnS(Ag) scintillator was placed on top of the membrane filter.
Method for auto-identification of alpha particle spots. As presented in Fig. 14, the alpha particle was confirmed as a spot when measuring with the alpha imager. The alpha particle spots were auto-identified by processing to acquire images as follows. First, we applied a 2D Gaussian filter (smoothing filter) to remove the noise distributed in the image caused by the image sensors. We used the OpenCV library to convolve with an image the 5 × 5 Gaussian filter kernel: After applying the filter to the image, we employed a binary processing onto the image to detect the edge of the alpha particle spot. Finally, we obtained the number of contours (alpha particles).  for evaluating the AMAD value in the internal exposure dose calculation. d e is defined as the diameter of the spherical particle with the same volume as the particle considered 25 . First, d e was calculated from the number of detected counts of the alpha particles measured by the alpha imager. We obtained d e by calculating a CF to convert the number of detected counts by our imager into the d e of a PuO 2 particle using a Monte Carlo particle and heavy ion transport code system (PHITS) ver. 2.8 simulation 26 . CF was used to convert the measured count rate to the equivalent volume diameter of PuO 2 particles. We reproduced a setup in a Monte Carlo simulation code which is similar to the actual PuO 2 particle imaging. Figure 15 presents the geometry of the simulation. The alpha particles were emitted to isotropic angles from the PuO 2 particle. The Pu isotopic composition of the PuO 2 particle was based on a MOX fuel that we measured using our imager. The alpha activities A i of the Pu isotope were determined by multiplying the specific activity SA i of each isotope by the isotopic composition IC i as follows: Accordingly, the diameters of the Pu particle varied from 1.0 to 10.0 μm in the simulation. The density of the PuO 2 particle was 11.46 g/cm 3 . CF was calculated using where N alpha denotes the number of detected alpha counts measured using our imager.
Particles suspended in the atmosphere move differently at different densities. Therefore, to compare the properties of particles of different densities, the d ae value should be calculated. d ae can be numerically solved from d e using the equation 22 , where λ = 0.0712 μm; χ and ρ are the particle shape factor (1.8, cited from the experimental data 27 ) and mass density, respectively; ρ 0 is the unit density (1 g/cm 3 ); and C(d) is the slip correction (often called the Cunningham factor) 22 . The value of d ae is defined by ICRP as the value at which 50% of the airborne activity in a specified aerosol is associated with particles greater than the AMAD 28 .
Received: 21 July 2020; Accepted: 3 February 2021  Figure 15. Geometry of the Monte Carlo simulation setup, which is the same as that of the measurement. The equivalent volume of the PuO 2 particle is varied from 1 to 10 µm.