SiPM time resolution: From single photon to saturation S. Gundacker a,n , E. Auffray a , N. Di Vara a , B. Frisch a , H. Hillemanns a , P. Jarron a , B. Lang b , T. Meyer a , S. Mosquera-Vazquez b , E. Vauthey b , P. Lecoq a a European Organization for Nuclear Research (CERN), 1211 Geneva 23, Switzerland b Physical Chemistry Department - Sciences II - University of Geneva 30, Quai Ernest Ansermet, 1211 Geneva 4, Switzerland article info Available online 8 February 2013 Keywords: SiPM TOF-PET Single photon time resolution Multi pixel photon counter (MPPC) NINO Femtosecond laser abstract The time resolution of photon detection systems is important for a wide range of applications in physics and chemistry. It impacts the quality of time-resolved spectroscopy of ultrafast processes and has a direct inﬂuence on the best achievable time resolution of time-of-ﬂight detectors in high-energy and medical physics. For the characterization of photon detectors, it is important to measure their exact timing properties in dependence of the photon ﬂux and the operational parameters of the photo- detector and its accompanying electronics. We report on the timing of silicon photomultipliers (SiPM) as a function of their bias voltage, electronics threshold settings and the number of impinging photons. We used ultrashort laser pulses at 400 nm wavelength with pulse duration below 200 fs. We focus our studies on different types of SiPMs (Hamamatsu MPPC S10931-025P, S10931-050P and S10931-100P) with different SPAD sizes (25 mm, 50 mm and 100 mm) coupled to the ultrafast discriminator ampliﬁer NINO. For the SiPMs, an optimum in the time resolution regarding bias and threshold settings can be reached. For the 50 mm type, we achieve a single photon time resolution of 80 ps sigma, and for saturating photon ﬂuxes better than 10 ps sigma. & 2013 Elsevier B.V. All rights reserved. 1. Setup The tests comprised a series of systematic studies of SiPMs in terms of: SiPM ﬁll factor or SPAD size, i.e. 25, 50 and 100 mm; SiPM bias voltage; Discriminator (NINO [1]) threshold; optical density (OD) of the light attenuators. We characterized the SiPMs regarding their break-down voltage and dark count. An overview of their properties can be seen in Table 1. As can be seen in Fig. 1, the 50 mm type has the lowest dark count rate, even for high bias overvoltages. In our investigations we always used SiPMs (MPPCs) with the same active area of 3  3 mm 2 produced by Hamamatsu Photo- nics. The setup can be seen in Fig. 2. The data was acquired with a fast LeCroy Oscilloscope DDA 735Zi 40 Gs/s that achieves 1 ps time resolution by interpolation. The femtosecond laser operated at 400 nm wavelength with a pulse width of 200 fs. For each attenuation factor, measured in optical density (OD) of the laser beam we scanned the SiPM bias and NINO threshold to ﬁnd the optimum values. 2. Time reference To establish a precise trigger we split the beam into two, each illuminating a 25 mm – SiPM with approx. 8000 photons. The performance of the trigger was then tested by measuring the time delay between the two 25 mm type SiPMs. The signal of the SiPMs was directly fed into the oscilloscope without using any further electronics. We obtained a coincidence time resolution of s ¼ 4:1 ps. Thus the trigger jitter is about s ¼ 4:1 ﬃﬃ 2 p ps ¼ 2:9 ps. 3. Data analysis NINO uses the time-over-threshold method and produces a square pulse. The leading edge gives the time information, and the pulse width is proportional to the input charge [3]. A small area around the peak of the pulse width histogram was selected and the corresponding delay time spectrum was plotted, Fig. 3. With this selection we reduce the inﬂuence of time walk and the Poissonian photon ﬂux jitter on the time distribution. Contents lists available at SciVerse ScienceDirect journal homepage: www.elsevier.com/locate/nima Nuclear Instruments and Methods in Physics Research A Table 1 Properties of three different photodetectors from HAMAMATSU, with 3  3 mm 2 active area. Type: S10931 SPAD size (mm 2 ) Number of cells Fill factor (%) Break down (V) Opt. bias for PET [2] (V) 100P 100  100 900 78.5 69.3 70.3 050P 50  50 3600 61.5 70.5 72.4 025P 25  25 14,400 30.8 69.2 73 0168-9002/$ - see front matter & 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.nima.2013.01.047 n Corresponding author. Tel.: þ41 22 767 4623. E-mail address: stefan.gundacker@cern.ch (S. Gundacker). Nuclear Instruments and Methods in Physics Research A 718 (2013) 569–572