Time-based signal sampling using sawtooth-shaped threshold.

Energy measurement of scintillation pulses by using time-over-threshold (TOT) has advantages of low readout cost, low power consumption, and less complexity compared with conventional analog-to-digital converter (ADC) approaches. Therefore, TOT is attractive in positron emission tomography (PET) systems based on silicon photomultiplier (SiPM) arrays requiring many readout channels. However, poor energy resolution and linearity of TOT leads to degradation of the overall PET detector and system performance, which is unsuitable for high-performance PET systems. To overcome these limitations, we propose a novel time-based signal-sampling method, sawtooth threshold sampling (STS), for PET scintillation detectors. This method uses a sawtooth-shaped threshold signal generated adaptively to the input scintillation pulse. Based on the time difference between the rising and falling edges of the digital pulse train generated by comparing the input scintillation pulse to the sawtooth signal, we can estimate the input scintillation pulse amplitude at the several time points. We compared several curve-fitting and numerical integration methods for energy estimation from the STS samples. Coincidence data between two identical scintillation detectors composed of one-to-one coupled SiPM and LGSO crystal (3  ×  3  ×  20 mm) was measured using the proposed STS circuit. For timing- and energy-resolution measurement, STS (10.5%  ±  0.21% and 200  ±  5.8 ps) was superior than simple TOT (16.4%  ±  0.52% and 224  ±  8.2 ps) and was similar to high-speed ADC (9.84%  ±  0.11% and 193  ±  4.3 ps). In conclusion, the proposed method can be a cost-effective solution for data collection in future SiPM-based PET systems.