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

This paper presents a compact, reproducible laboratory setup using picosecond laser pulses and waveform digitization to characterize the gain, transit time spread, and noise properties of photomultiplier tubes across a range of supply voltages and temperatures, providing a standardized reference for single photoelectron measurements in low-light physics applications.

The Gist

Imagine you are trying to listen to a single, tiny whisper in a room full of people shouting. That is essentially what scientists do when they study dark matter or neutrinos (ghostly particles that rarely interact with anything). They use giant tanks of liquid and sensitive "ears" called photomultiplier tubes (PMTs) to catch the faintest flashes of light these particles might make.

But here's the problem: If your "ear" isn't perfectly calibrated, you might mistake a random crackle of static for a whisper, or miss a real whisper entirely. To fix this, the authors of this paper built a super-precise testing lab to check how well these "ears" work.

Here is a breakdown of their work using simple analogies:

1. The Setup: A "Whispering Booth"

Think of the experimental setup as a soundproof recording booth for light.

• The Laser: Instead of a human whispering, they use a picosecond laser. This is a light source so fast it fires a "blip" of light in a trillionth of a second. It's like a camera flash that happens faster than a hummingbird's wingbeat.
• The PMT: This is the microphone. It sits inside a black, light-tight box (so no outside light sneaks in) and waits to catch that single blip.
• The Climate Chamber: They can turn this booth into a freezer or a warm room. They tested the microphones at temperatures ranging from a chilly -50°C (like a deep freezer) to a cozy +20°C (room temperature).

2. The Mission: Counting the "Single Drops"

The goal is to see how the PMT reacts to Single Photoelectrons (SPE).

• The Analogy: Imagine a single raindrop hitting a tin roof. You want to know: Does the roof make a loud ping? Is the ping always the same volume? Does it take the same amount of time for the sound to reach your ear every time?
• The Test: They fire the laser so weakly that, on average, only one "raindrop" (photon) hits the PMT at a time. They record thousands of these events to build a profile of how the PMT behaves.

3. The Tools: "Digital Magnifying Glasses"

They didn't just listen; they recorded the sound waves in extreme detail using a high-speed digital recorder (an oscilloscope).

• Charge Spectrum (The Volume): They measured how "loud" the signal was. They developed a clever math trick called self-convolution.
  • The Metaphor: Imagine you have a bag of marbles. You don't know how big the "single marble" is, but you see some clumps of two marbles stuck together. By looking at the shape of the "two-marble clumps," you can mathematically work backward to figure out exactly what a "single marble" looks like. This helps them separate real signals from noise without guessing.
• Transit Time Spread (The Speed): They measured how long it took for the signal to travel from the light hitting the tube to the electrical signal leaving it.
  • The Metaphor: If you drop a ball down a slide, does it always take exactly 2 seconds? Or does it sometimes take 1.9 and sometimes 2.1? This "jitter" is called Transit Time Spread (TTS). The faster and more consistent the slide, the better the detector.

4. What They Discovered

After testing many tubes, they found some interesting rules of thumb:

• Voltage is the Volume Knob: Turning up the electricity (voltage) makes the signal louder (higher gain) and the timing more precise (lower TTS). It's like turning up the gain on a microphone; the sound gets clearer, but you have to be careful not to distort it.
• Cold is Good for Volume: When they cooled the tubes down to -50°C, the signal got slightly louder (about 0.1% louder for every degree dropped). It's like how a cold battery might hold a charge differently, or how cold metal conducts electricity better.
• Timing Doesn't Care About Temperature: Interestingly, while the volume changed with temperature, the speed (timing) stayed the same. The "slide" took the same amount of time whether it was hot or cold.
• The "Ghost" Signals: Sometimes, the PMT fires a tiny signal before the real one (a "prepulse") or long after (a "late pulse"). They found these "ghosts" happen very rarely (less than 2% of the time), which is good news for scientists trying to find real signals.
• Cables Matter: The length of the wire connecting the tube to the computer actually changed the results! Shorter wires gave a slightly louder and faster signal. This is a crucial lesson: if you compare two different labs, you must make sure their "wires" are the same length, or you aren't comparing apples to apples.

5. Why Does This Matter?

This paper provides a recipe book for scientists.

• Before this, every lab might have had its own weird way of testing these tubes.
• Now, they have a standardized, compact, and reliable method (a "table-top" setup) that anyone can use.
• This ensures that when a giant experiment (like one looking for dark matter) says, "We found a signal!" we can be 100% sure it wasn't just a glitch in the equipment.

In a nutshell: The authors built a tiny, precise "light-testing lab" to make sure the giant ears used to hunt for the universe's biggest mysteries are listening clearly, consistently, and without any static.

Technical Summary

1. Problem Statement

Modern low-background neutrino and dark matter experiments rely on large arrays of photomultiplier tubes (PMTs) to detect faint scintillation and Cherenkov light. To control systematic uncertainties in reconstructed energy, position, and event topology, precise characterization of the Single Photoelectron (SPE) response, gain, and timing is essential.

• The Gap: While large-scale experiments (e.g., Borexino, JUNO) utilize dedicated, high-throughput test stands for qualification, there is a lack of compact, flexible laboratory setups for detailed R&D studies.
• Specific Needs: Researchers require a reproducible method to characterize PMTs under controlled low-light conditions, including variable temperatures, to understand device-to-device variations, gain stability, and timing properties (Transit Time Spread - TTS) without relying on complex parametric models for every analysis step.

2. Methodology and Experimental Setup

The authors developed a compact, table-top characterization unit designed for reproducible SPE-level measurements.

• Light Source: A PILAS DX diode laser (407 nm) with picosecond pulse widths (21–34 ps FWHM) and timing jitter <3 ps. The beam is attenuated via a neutral density filter (OD 5) to ensure 5–10% of triggers yield a detected PMT response, maintaining SPE-level conditions.
• Environment: The PMT is housed in a light-tight aluminum box or a Binder MKT 115 climate chamber, allowing temperature scans from −50°C to +20°C. The setup acts as a Faraday cage to minimize electromagnetic interference.
• Synchronization: A synchronized Data Acquisition (DAQ) chain uses a Rohde & Schwarz oscilloscope to generate TTL (laser trigger) and NIM (digitizer trigger) signals. The laser repetition rate is kept low (≤20 kHz) to ensure no pulse overlap within the 100 ns acquisition window.
• Data Acquisition: Waveforms are digitized using an Acqiris DC282 digitizer (up to 8 GS/s, 10-bit resolution) without thresholding.
• Analysis Framework:
  • Waveform Processing: Baseline noise, pulse height, start time (constant fraction), and integrals (trace and pulse) are extracted.
  • Trace Charge Spectrum Modeling: A model-independent approach extends the standard Poissonian model. It uses an Exponentially Modified Gaussian (EMG) distribution to account for baseline fluctuations and dark pulses, and a Heaviside-modified Gaussian for signal components to handle truncation at the pedestal.
  • DPE Quantification: A novel self-convolution method is introduced. By convolving the measured SPE pulse charge distribution with itself, the authors derive the Double Photoelectron (DPE) distribution shape without assuming a specific parametric model, allowing for robust DPE contamination estimation.
  • Timing Analysis: Transit time spectra are analyzed using Kernel Density Estimation (KDE) to separate Prepulses (PP), Main Pulses, and Late Pulses (LP) based on time differences relative to the trigger.

3. Key Contributions

• Compact, Versatile Test Stand: A validated, table-top setup capable of characterizing PMTs at SPE levels across a wide temperature range (−50°C to +20°C).
• Model-Independent DPE Estimation: Introduction of a data-driven self-convolution technique to quantify DPE contributions directly from charge spectra, reducing reliance on theoretical assumptions about the underlying SPE shape.
• Comprehensive Parameter Extraction: A unified workflow to extract Gain, Peak-to-Valley (P/V) ratio, Transit Time Spread (TTS), and populations of prepulses/late pulses from a single dataset.
• Temperature and Cabling Analysis: Systematic study of how temperature and cable length affect gain and TTS, providing critical correction factors for experimental setups.

4. Key Results

The system was tested on eight ET Enterprises 9821B, two 9821QB, and one Hamamatsu R9980 PMT.

• Gain Behavior:
  • Gain follows an exponential dependence on supply voltage ($G \propto e^{U/U_0}$).
  • Temperature Effect: Cooling increases gain by approximately 0.1% to 0.29% per °C (depending on the temperature range), attributed to changes in dynode secondary emission factors and reduced circuit resistance.
  • Cable Effect: Shorter cables yield higher apparent gain due to reduced attenuation.
• Timing (TTS):
  • TTS decreases as supply voltage increases.
  • Temperature Independence: No significant temperature dependence was observed for TTS in the tested range.
  • Cable Effect: Longer cables increase the apparent TTS due to signal dispersion.
  • Values clustered around 2.0–2.2 ns (FWHM) for 9821B tubes, consistent with datasheets.
• Pulse Populations:
  • Prepulses: ~0.02% of events, occurring ~12.6 ns before the main peak with reduced charge (bypassing the first dynode).
  • Late Pulses: ~1.4% of events, occurring ~25 ns after the main peak, caused by backscattering at the first dynode.
• Peak-to-Valley (P/V) Ratio:
  • Increases with voltage.
  • Ranges from ~1.0 to 2.3 for 9821B/9821QB and up to 4.5 for R9980.
  • No clear temperature dependence was found.
• Device Variation: Significant tube-to-tube variations were observed in gain scaling parameters ($U_0$) and absolute TTS values, highlighting the necessity of individual unit characterization.

5. Significance

• Reference Standard: This work provides a reproducible, practice-oriented reference for SPE-level PMT characterization, essential for the calibration of large-scale liquid scintillator and Cherenkov detectors (e.g., for the proposed LENA observatory or current LS R&D programs).
• Systematic Uncertainty Control: By quantifying the effects of temperature and cabling on gain and timing, the study enables more accurate corrections in low-light physics experiments, directly impacting the precision of energy and position reconstruction.
• Methodological Advancement: The model-independent DPE extraction method offers a robust tool for analyzing charge spectra in scenarios where parametric models may fail or be too complex to fit.
• Scalability: The methodology is transferable to other photosensor types (e.g., SiPMs) and can be extended to measure additional observables like afterpulsing and photon detection efficiency by incorporating reference sensors.