No abstract

Three types of SiPMs (Silicon Photomultiplier) with an active area of 3 × 3 mm 2 manufactured by KETEK with cell sizes of 50 µm (PM3350), 60 µm (PM3360) and 75 µm (PM3375) have been investigated. All devices have optical trenches in between the cells to suppress direct crosstalk. Their breakdown voltage at room temperature is about 23 V and the gain at an overvoltage U over = 3.4 V is > 6 • 10 6 . The temperature variation of the breakdown voltage is < 16 mV/K and the gain coefficient with temperature is < 1% for overvoltages U over > 1.7 V. The photodetection efficiency (PDE) at 420 nm and U over = 3.4 V is 51% for PM3350, 55% for PM3360 and 58% for PM3375. At U over = 3.4 V, the dark count rates are < 470 kHz/mm 2 at 20 • C and the afterpulse probability is < 9% at −20 • C. Single photon timing of 230 ps FWHM for PM3350, 320 ps for PM3360 and 375 ps for PM3375 have been achieved. To test their performance in PET (Positron Emission Tomography), energy spectra of 22 Na with LYSO (Lutetium Yttrium Oxyorthosilicate, Lu 1.8 Y .2 SiO 5 :Ce) and GAGG (Gadolinium Aluminum Gallium Garnet, Gd 3 (Ga,Al) 5 O 15 :Ce) scintillators with a size of 2×2×6 mm 3 have been acquired. The saturation corrected energy resolution (FHWM) at 511 keV was with LYSO 12.3% for PM3350, 13.4% for PM3360, 12.4% for PM3375 and with GAGG 10.8% for PM3350. Coincidence timing (FWHM) at U over = 3.4 V was with LYSO 174 ps for PM3350, 178 ps for PM3360, 157 ps for PM3375 and with GAGG 430 ps for PM3350.

The Silicon Photomultiplier (SiPM) is a promising photo-detector for a variety of applications. However, the high dark count rate (DCR) of the SiPM is still a contemporary problem. Decreasing the DCR would significantly broaden the range of possible applications. In this work we present a novel method for the spatially resolved characterization of crystal defects in SiPMs. The contribution of crystal defects to the DCR is evaluated by exploiting the effect of "hot carrier luminescence" (HCL), which is light that is emitted during the Geiger mode operation of avalanche photodiodes (SiPM micro-cells). Spatially confined regions with an enhanced light emission intensity (hotspots) are identified within the active areas of SiPM micro-cells. By correlating the detected light intensity and the DCR, a significant contribution of up to 56 % of the DCR can be attributed to less than 5 % of the micro-cells. The analysis of the temperature dependence of the emitted light identifies the Shockley-Read-Hall-Generation to be the dominant mechanism responsible for the occurrence of hotspots. The motivation of this work is to generate a deeper understanding of the origin of hotspots in order to suppress their contribution to the DCR of SiPMs.

Abstract. Despite several advantages of Silicon Photomultipliers (SiPM) over Photomultiplier Tubes (PMT) like the increased photon detection efficiency (PDE), the compact design and the insensitivity to magnetic fields, the dark count rate (DCR) of SiPM is still a large drawback. Decreasing of the SiPM dark count rate has become a modern task, which could lead to an enormous enhancement of the application range of this promising photo-detector. The main goal of this work is to gain initial information on the dark generation and identify the dominating contributions to dark currents. The chosen approach to fulfill this task is to extract characteristic activation energies of the contributing mechanisms from temperature dependent investigations of dark currents and DCR. Since conventional methods are not suited for a precise analysis of activation energies, a new method has to be developed. In this paper, first steps towards the development of a reliable method for the analysis of dark currents and dark events are presented. IntroductionThe dark count rate (DCR) of the Silicon Photomultiplier (SiPM) is still a limiting factor for the extension of its application range and requires a significant reduction. In order to achieve this requirement, a distinction of contributions determining the DCR is needed to be able to identify the dominating mechanisms. As reported in [1,3,4,5] the extraction of characteristic activation energies from the temperature dependence of the dark current is able to provide information on the dominating mechanisms. Unfortunately these methods are not suited for a precise determination of the needed energy levels, since the investigated dark currents are determined by a mixture of effects depending on voltage on the one hand (e.g. generation current, electric field effects) and overvoltage on the other hand (e.g. gain, crosstalk, afterpulsing). In this paper, we propose a new method to separate the mentioned effects using the responsivity of the detector as an appropriate reference for the measured dark current. This approach allows us

No abstract