학술논문

A Silicon Photomultiplier Readout ASIC for Time-of-Flight Applications Using a New Time-of-Recovery Method
Document Type
Periodical
Source
IEEE Transactions on Nuclear Science IEEE Trans. Nucl. Sci. Nuclear Science, IEEE Transactions on. 65(5):1196-1202 May, 2018
Subject
Nuclear Engineering
Bioengineering
Impedance
Jitter
Timing
Energy resolution
Semiconductor device measurement
Positron emission tomography
Detectors
Application specific integrated circuits
charge measurement
CMOS
PET
photodetectors
silicon photomultipliers
time of recovery
time over threshold
time-of-flight PET
timing electronics
Language
ISSN
0018-9499
1558-1578
Abstract
Silicon photomultiplier timing chip (STiC) is a 64-channel mixed-mode application-specific integrated circuit in the $0.18-\mu \text{m}$ CMOS technology from Univited Microelectronics Corporation (UMC) for silicon photomultiplier (SiPM) readout with very high timing resolution. It is designed for time-of-flight measurements in positron emission tomography and high-energy physics experiments. In order to achieve the best timing performance without compromising the charge/energy measurement, a novel time-based signal processing technique called “time-of-recovery (ToR)” method has been developed and implemented in the analog front end of the chip. This technique converts the incoming charge into a digital pulse with a linearized time-over-threshold width. Measurements have shown a time jitter smaller than 20 ps for the analog front end and smaller than 40 ps for the time-to-digital converter and the digital part. A coincidence time resolution of 214-ps full-width at half-maximum (FWHM) has been obtained with STiC using $3.1{\times }3.1{\times }15$ mm 3 LYSO:Ce crystals and Hamamatsu multi pixel photon counters (S12643-050CN(X)). The measured energy resolution for a 511-keV photon is 11.2% FWHM after correcting for SiPM saturation effects. In this paper, we report on the details of the ToR method and how it is embedded within the STiC design; results of performance measurements as well as the 128-channel front-end module used for the EndoTOFPET-US project are also presented.