Picosecond laser test unit for photosensor characterization at ambient and low temperatures

Abstract

Accurate single photoelectron (SPE) characterization of photosensors is essential for controlling systematic uncertainties in low-light neutrino and dark matter detectors. We present a compact laboratory setup for the characterization of photosensors under controlled, low-light conditions. Specifically, we demonstrate its use with photomultiplier tubes (PMTs) operated at the SPE-level, using picosecond laser pulses and waveform digitization to determine key PMT properties. Measurements as a function of supply voltage and temperature ($-50^\circ$C to $+20^\circ$C) are performed on ET Enterprises 9821(Q)B tubes and a Hamamatsu R9980 assembly, which show exponential gain-voltage behavior and device-to-device variation. Cooling increases the gain by $\sim 0.1\,\%/^\circ$C, while the transit time spread (TTS) and peak-to-valley ratio (P/V) exhibit no clear temperature dependence. TTS decreases with voltage. Late pulses remain at the percent level and prepulses at the sub-percent level. Cable length affects both apparent gain and TTS. A model-independent, data-driven self-convolution method is introduced to quantify double photoelectron contributions from pulse charge spectra. The procedures provide a reproducible, practice-oriented reference for SPE-level PMT characterization and can be extended to other photosensor types.

Figures (10)

• Figure 1: Schematic of the experimental setup for photosensor characterization. A TTL signal triggers the laser, while a synchronized NIM pulse triggers the digitizer. The attenuated laser beam illuminates the PMT inside a light-tight enclosure.
• Figure 2: Photograph of the PMT inside the light-tight enclosure. The fiber output is positioned about 35 cm from the photocathode, ensuring full illumination.
• Figure 3: Example waveform from a 3-inch 9821B PMT at the SPE-level. The baseline mean (green) and its standard deviation (light blue band) quantify the noise. The dashed blue line marks the pulse height, the dashed black line the FWHM, the purple dot the start time from constant fraction timing (20 %), and the shaded red and rose areas the trace and pulse integrals, respectively.
• Figure 4: Example trace charge distribution at SPE intensity from a 9821B PMT, operated at 1800 V supply voltage. The blue curve is the best-fit model. Colored components indicate noise (purple dashed), SPE (green dash-dotted), DPE (cyan dotted), and TPE (gold dashed) contributions. Vertical red dashed lines mark the SPE selection window. The green marker denotes the SPE mean $\mu_{\text{SPE}}$.
• Figure 5: Example pulse charge spectrum at SPE intensity for a 9821B PMT operated at 1800 V. The gold curve is a Gaussian KDE and the green marker indicates the SPE mean charge. Vertical red dashed lines denote the SPE selection window. The cyan dash-dotted curve is a DPE model fitted to the high-charge tail to estimate DPE contamination within the SPE window.