High temporal performance gamma-ray detector

Detector concept

Cherenkov emission has received attention to improve timing resolution of detectors dedicated to positron emission tomography. Since the Cherenkov photons are promptly emitted within an order of 1 ps, the timing performance of the detector would be enhanced compared to a scintillation-based detector. A monolithic Cherenkov radiator which does not emit scintillation photons was employed to suppress reflections in the radiator. However, it is challenging to reconstruct the interaction position of the gamma ray because the number of Cherenkov photons emitted when a gamma ray interacts with the radiator is only 30 at most. As a photodetector, a high-dense photodetector array with individual readout was applied to extract all temporal and spatial information of the Cherenkov photons.

■Features

Monolithic Cherenkov radiator
　→To maximize the timing performance of Cherenkov emission
Individual readout
　→To obtain all temporal- and spatial-information from a few Cherenkov photons

Required specification for photodetector

Monte Carlo simulations were performed [Publication 1]. According to the simulations, we found that requirements for the photodetector are high-dense and good single photon time resolution (SPTR).For example, the SPTR must be better than σ = 40 ps to obtain a coincidence time resolution (CTR) of 100 ps full width at half maximum (FWHM).
Moreover, we have investigated potential capability of deep learning to improve the position resolution of the proposed detector [Publication 2].

Experimental validation

An experiment measuring the CTR was performed using a high speed photodetector of MCP-PMT (R3809). The SPTR of the MCP-PMT is 25 ps FWHM.
A lead fluoride Cherenkov radiator was attached to the window face plate of the MCP-PMT using an optical glue.
As a result, the CTR better than 50 ps FWHM was successfully obtained, and this result is consistent with the simulation result [Publication 3].

Experimental validation by MCP-PMT (R3809)

Development of Cherenkov-radiator-integrated MCP-PMT

There are optical boundaries in the Cherenkov detector described above. The optical boundaries will worsen the timing performance of the detector. We have developed a Cherenkov-radiator-integrated MCP-PMT (CRI-MCP-PMT) to remove the optical boundaries from the detector. The window face plate of the ordinary MCP-PMT was replaced with the Cherenkov radiator and Al2O3 layer was inserted between the radiator and the photocathode using atomic layer deposition technique.
The CTR was measured using a pair of the CRI-MCP-PMTs, and the CTR of 41.9 ps FWHM was obtained [Publication 4]. Additionally, the CTR was improved to 30.1 ps FWHM by optimizing the analysis parameters, e.g. timing pick-off threshold level and pulse area.

Development of Cherenkov-radiator-integrated MCP-PMT

Performance improvement with lead-free materials

By reviewing all materials starting from scatch and applying an ALD technique, we succeeded in creating ALD-MCPs that ensure low noise and high signal multiplication (gain) without using materials containing lead. Eliminating lead is great step forward because its use is restricted in the RoHS directive issued by the European Union (EU) as a hazardous substance.

Performance improvement with lead-free materials

A coincidence experiment revealed that a PMT integrating the ALD-MCP can enhance the CTR from 41.9 to 35.4 ps FWHM compared to an ordinary MCP-PMT [reference 6].

See details

coincidence experiment result

Reconstruction-free imaging

We have succeeded in being the first in the world to achieve high-accuracy positron emission imaging without image reconstruction by utilizing the pair of detectors mentioned above, and unique light detection and signal processing techniques [reference 7]. Applying these successful research results are promising to achieve a completely innovative new type of radiation medical imaging system capable of making speedy diagnoses from a simple, compact setup yet with the same or higher accuracy than currently used radiation imaging systems such as positron emission tomography (PET). This will help boost inspection efficiency for detecting diseased tissues or organs, such as from cancer, and also reduce the radiation exposure dose, alleviating the load on the patient and medical staff. We will further improve the detector performance to make this new concept more practical.

See details

Reconstruction-free imaging

Reconstruction-free anatomical imaging method

PET scanners are generally combined with an X-ray computed tomography (X-CT). The data obtained by a PET scanner contains metabolic information of a patient, however, does not contain its anatomical information, which is necessary to correct the PET data and make it more quantitative. The above-mentioned proposed reconstruction-free imaging method using ultrafast timing detectors also requires the anatomical information because it is also a system measuring patient’s metabolic information. Although the anatomical information can be obtained using X-CTs, X-CTs generally are huge systems, which will not be able to maximize the potential of the compactness of the proposed reconstruction-free imaging method.

To tackle with this problem, we propose a new anatomical imaging method using the same ultrafast timing detectors to maintain the potential [reference 9]. We utilize a nature of an interaction between matter and light, so called Compton scattering. By utilizing the fact that probability of Compton scattering between matter and light is proportional to the electron density of the matter, we devised an algorithm which can localize the three-dimensional interaction position based on spatiotemporal information of the ultrafast detectors. It is demonstrated that the anatomical information can be obtained by our algorithm through a Monte Carlo simulation.

Theoretical understanding of anatomical imaging method

Although it was proved that an anatomical image can be obtained by our proposed algorithm, we also found that the spatial resolution of the anatomical image obtained by our proposed algorithm is significantly worse than that intuitively expected. Therefore, we worked on understanding a physics behind our proposed algorithm [reference 10]. Through the theoretical approach, we figure out that the spatial resolution of the anatomical image gets worse in certain conditions. These conditions are validated through Monte Carlo simulations, and we succeeded in improving the spatial resolution of the anatomical image be eliminating the conditions.