Characterization of Cryogenic SiPM Down to 6.5 K

Ryoto Iwai1, Mikio Sakurai1, Aldo Antognini1,2, Ivana Belosevic1, Malte Hildebrandt2, Klaus Kirch1,2, Andreas Knecht2, Angela Papa2,3 and Alexey Stoykov2 1 Institute for Particle Physics and Astrophysics, ETH Zürich, 8093 Zürich, Switzerland 2Paul Scherrer Institute, 5232 Villigen-PSI, Switzerland 3 Dipartimento di Fisica, Università di Pisa, and INFN sez. Pisa, Largo B. Pontecorvo 3, 56127 Pisa, Italy


Introduction
At the Paul Scherrer Institute, several experiments at cryogenic temperatures are under development, such as muCool (development of ultra-cold high-brightness muon beam line) [1], HyperMu (measurement of hyperfine splittings in muonic atoms) and muonium production from super-fluid helium for testing the gravitational interaction of anti-matter [2]. These experiments make use of scintillators at cryogenic temperatures, however, their scintillating light is typically transported to an environment at warmer temperatures before being detected by photo-detectors. Operation of Silicon Photo-Multipliers (SiPM) in the cryogenic environment could reduce the complexity and allow for finer detector segmentation while improving energy and timing resolutions due to the increased light collection.
Few studies [3], [4] and [5] have already partially characterized SiPMs below liquid nitrogen temperatures. In this study, we characterized a commercially available SiPM in the temperature range from 6.5 K to 286 K.

Experimental setup
The tested SiPM is the S13370-3050CN model (also known as VUV MPPC) from Hamamatsu Photonics with a (3×3) mm 2 active area [6]. It was originally designed for detecting scintillation light of liquid xenon and liquid argon. One peculiarity of this model is the absence of the epoxy layer protecting the silicon surface. This improves not only the VUV sensitivity but also the mechanical stability at low temperatures by reducing the mechanical stress due 1 to thermal contraction. This SiPM also has metallic quenching resistors to maintain the pulse waveform also at low temperatures.
In our setup, the SiPM was thermalized to a cold stage connected to a cryo-cooler by contacting its ceramic package as shown in Fig. 1. Moreover, the two pins and the coaxial cable are thermally contacted to the cold stage. The SiPM was protected from thermal radiation by two thermal shields: one contacted to the cold stage down to 6.5 K, the other at around 100 K. The temperature was controlled by two heaters on the cold stage and maintained with a stability of ±0.1 K during the measurements.

Electrical properties
Current-voltage curves for various temperatures were measured by illuminating the SiPM with a weak LED light (Fig. 2). A baseline and Geiger mode regions of the curve were fitted with a linear and cubic functions in semilogarithmic scale and breakdown voltges were extracted as intersection points of two functions. We observed a non-linear temperature dependence of the breakdown voltage at low temperatures, which can be explained by Baraff's model [7].  Furthermore, the resistance of the quenching resistors were measured by applying forward bias voltages. We observed only an increment of 19 % from 286 K to 6.5 K. This small temperature dependence does not alter the pulse waveform at low temperatures.

Waveform analysis
Waveforms at the pre-amplifier output have been recorded for various overvoltages and temperatures as shown in Fig. 3. The SiPM was illuminated with a pulsed blue LED light and the recording was synchronized with the LED pulses. One initial problem, visible in Fig. 3, was periodic noise in phase with the LED pulse which can be subtracted in the data or suppressed in the measurement. In another previous measurement with a different setup, we could show that the prompt peak of the waveform remains basically unaffected down to temperature of 40 K. On the contrary, the amplitude of the tail decreases with decreasing the temperature, while its falling time increases. The SiPM gain was obtained by measuring the integrated charge of the single photoelectron waveform (Fig. 4 Left). The linearity with respect to bias voltages holds for all temperatures. The photo-electron peaks of the charge spectra were clearly separated even at 6.5 K. However, at low temperatures we observed a decrease of the maximal overvoltage that can be applied because of the increased after-pulse rates. For instance at 6.5 K, the maximal overvoltage was approximately 1.0 V.
Relative photon detection efficiencies (PDE) were measured by using the same pulsed LED (Fig. 4 Right). The average number of detected photons was evaluated assuming Poisson statistics. The results have relatively large non-estimated systematic uncertainties probably related to the light conditions during the data taking. However, no large decrement of relative PDEs at low temperatures were observed. Dark count rates were measured by acquiring waveform data with the LED off and a random trigger. Due to a relatively small signal-to-noise ratio for low overvoltages, we applied a digital filtering algorithm to the raw waveforms in order to improve the signal-tonoise ratio. The dark count rate was evaluated from the probability of finding a signal in a certain time window by assuming Poisson statistics. It was largely decreased with decreasing temperature: 456 kHz at 286 K and 11 kHz at 250 K with the overvoltage of 2.25 V.

Applications
We investigated the performance of a plastic scintillator-SiPM assembly at cryogenic temperatures. The scintillator was glued on the SiPM active area as shown in Fig. 5. Because of the absence of the protection layer, the scintillator size was chosen to be smaller than the active area of the SiPM to avoid a conflict with the bond wires. A stable operation of the scintillator-SiPM assembly at 44 K was confirmed by detecting β-rays of 90 Sr.

Conclusion
We characterized a cryogenic SiPM over a temperature range from 6.5 K to 286 K. The breakdown voltage depends linearly on the temperatures till around 200 K and saturates to a constant value at lower temperatures. Because of the increase of after-pulse probability, the maximal applicable overvoltage decreases at low temperatures. However, the SiPM is still operational down to 6.5 K with a similar performance as at room temperature. Furthermore, a scintillator-SiPM assembly was successfully tested at cryogenic temperatures.