Search

The intermediate phases between the initial interaction of ionizing radiation in scintillators and the production of scintillation light have been extensively studied over the years. Observable features of scintillators such as energy and time resolution have been in part explained through pre-scintillation mechanisms. Whereas the early energy transfer processes following the photoelectric (or Compton) interaction of an energetic photon with an atomic electron in scintillator materials were fully described, their spatio-temporal characteristics have had limited quantification. For 511 keV photon interaction in (TOF-)PET detectors, the production of a primary electron carrying most of the energy leads to non-localized energy deposit distribution having some temporal extent. Here, we characterize the spatial, temporal and energy evolution of the primary electron generated by 511 keV photons in different scintillators using Monte Carlo simulations. We aim to provide a general quantitative description detailing the dynamics of energy transfer, such as the electron ejection angle, track length and tortuosity, time course and remaining energy as a function of the initial interaction point. Simulation results show that the electron track is tortuous with a short extent and a rapid energy transfer within 100 μ m in high-Z (e.g., BGO) or high density (e.g., LuAP:Ce) materials while the extent nearly doubles in lighter materials like Ba F 2 in which the track tortuosity is 10%–50% lower. The time fluctuations of the photoelectron full energy deposit remain below 1 ps FWHM, therefore adding negligible jitter to the time resolution of coincident detectors. These simulations provide a better understanding of the main carrier dynamics from simple characteristics as some detector concepts require low-range, contained energy deposit (e.g., pixelated crystal arrays) whereas others benefit from a greater extent of energy deposition (e.g., structures designed for energy sharing). [ABSTRACT FROM AUTHOR]

The 3M-ESR multilayer polymer film is a widely used reflector in scintillation detector arrays. As specified in the datasheet and confirmed experimentally by measurements in air, it is highly reflective ( > 98 % ) over the entire visible spectrum (400–1000 nm) for all angles of incidence. Despite these outstanding characteristics, it was previously found that light crosstalk between pixels in a bonded LYSO scintillator array with ESR reflector can be as high as ∼30–35%. This unexplained light crosstalk motivated further investigation of ESR optical performance. Analytical simulation of a multilayer structure emulating the ESR reflector showed that the film becomes highly transparent to incident light at large angles when surrounded on both sides by materials of refractive index higher than air. Monte Carlo simulations indicate that a considerable fraction (∼25–35%) of scintillation photons are incident at these leaking angles in high aspect ratio LYSO scintillation crystals. The film transparency was investigated experimentally by measuring the scintillation light transmission through the ESR film sandwiched between a scintillation crystal and a photodetector with or without layers of silicone grease. Strong light leakage, up to nearly 30%, was measured through the reflector when coated on both sides with silicone, thus elucidating the major cause of light crosstalk in bonded arrays. The reflector transparency was confirmed experimentally for angles of incidence larger than ~ 60 ° using a custom designed setup allowing illumination of the bonded ESR film at selected grazing angles. The unsuspected ESR film transparency can be beneficial for detector arrays exploiting light sharing schemes, but it is highly detrimental for scintillator arrays designed for individual pixel readout. [ABSTRACT FROM AUTHOR]

Abstract Single photon avalanche diodes (SPAD) are detectors used for applications requiring high timing resolution and single photon sensitivity. Medical imaging modalities such as positron emission tomography benefit from higher timing resolution enabling time-of-flight (ToF) measurements. One of the current challenge in the development of SiPM detectors is combining both high photodetection efficiency (PDE) and high single photon timing resolution (SPTR). Although SiPM have a high fill factor, they are limited for sub-10 ps FWHM SPTR for two reasons: the high output capacitance directly affects the rise time and a timing skew is associated with the SPAD position in the array. The proposed solution to obtain ToF capabilities with both high SPTR and PDE is to make a 3D digital SiPM where every SPAD is individually connected to a quenching circuit (QC) and a time-to-digital converter (TDC). While the 3D integration process is underway, we designed and tested a 2D version of a single channel of the detector composed of a SPAD, a QC and a TDC in TSMC CMOS 65 nm. This paper presents a SPAD and QC with 9 ps FWHM SPTR and a chain composed of a SPAD, a QC and a TDC with 17.5 ps FWHM SPTR. [ABSTRACT FROM AUTHOR]

Coincidence timing resolution (CTR) of 10 ps FWHM would drastically increase the contrast of images in pre-clinical and clinical positron emission tomography (PET) due to the capability to localize the annihilation site along the line of response with unprecedented accuracy. Currently, the scintillators and the photodetectors with their electronic readout are the two limiting factors to reach such a timing resolution. On the photodetector side, the single photon timing resolution (SPTR) must be improved below 4 ps RMS to reach a CTR of 10 ps FWHM. To this end, we propose to use a 3D digital silicon photomultiplier (3DdSiPM) where each single photon avalanche diode (SPAD) has its own quenching circuit (QC) and time-to-digital converter (TDC). To reach an SPTR of 4 ps RMS in an array of SPAD requires to correct each readout pixel individually to minimize the impact of TDC least significant bit (LSB) non-uniformities and SPAD-to-SPAD skew. For this purpose, we developed an array of 256 pixels of SPAD readout circuits optimized for high timing precision with embedded digital signal processing. The latter is a code-to-time conversion where each pixel is individually corrected for TDC LSB variation and skew. Measurement results on single pixels (QC and TDC) show timing jitter down to 8 ps RMS. The uncorrected array timing jitter is about 87 ps RMS and is reduced to 18 ps RMS after skew correction and TDC LSB correction. [ABSTRACT FROM AUTHOR]