Scintillation light calibrations, systematic uncertainties, and triggering efficiency in the MicroBooNE detector

This paper presents a comprehensive five-year analysis of scintillation light performance in the MicroBooNE detector, detailing its simulation, calibration, and triggering efficiency while reporting novel observations of a significant long-term light yield decline and an unexpectedly high single photoelectron noise rate.

The Gist

Imagine the MicroBooNE detector as a giant, high-tech underwater camera, but instead of water, it's filled with 170 tons of super-cold liquid argon. Its job is to take "photos" of ghostly particles called neutrinos as they zip through the tank.

However, neutrinos are shy; they rarely interact with anything. When they do, they leave two types of clues:

1. Electric Charge: Like a trail of footprints in the sand (which the camera's wires catch).
2. Scintillation Light: A tiny, instant flash of blue light (which the camera's "eyes" catch).

This paper is essentially the maintenance log and performance review of the "eyes" (the light detectors) over five years of operation. Here is the story of what they found, explained simply.

1. The Eyes and the Flashlight

The detector has 32 giant "eyes" called Photomultiplier Tubes (PMTs). Think of these as incredibly sensitive night-vision goggles.

• How they work: When a neutrino hits the argon, it creates a flash of invisible ultraviolet light. Before the goggles can see it, a special coating (TPB) on the lens acts like a translator, turning that invisible UV light into visible blue light.
• The Trigger: The camera is too big to record every single second of every day. It needs a "trigger" to start recording. The system waits for a flash of light that happens at the exact same time the neutrino beam arrives. If the flash is too dim, the camera ignores it.

2. The "Flickering" Problem (Calibration)

Just like old lightbulbs or camera sensors, these electronic eyes don't stay perfect forever.

• The Issue: Over time, the sensitivity of the eyes drifted. Sometimes they got too sensitive; sometimes they got sluggish.
• The Fix: The team used a clever trick. Even when no neutrinos were hitting the tank, there was a constant, tiny "static" noise in the system—random single flashes of light (called Single Photoelectrons or SPEs). It was like hearing a constant, faint buzzing in a quiet room.
• The Analogy: Imagine you are trying to calibrate a microphone. Instead of playing a test tone, you listen to the room's background hum. By measuring that hum, you can figure out exactly how loud the microphone is set. The MicroBooNE team used this constant "buzz" to constantly recalibrate their eyes, ensuring they didn't miss any real neutrino flashes.

3. The Dimming Light (The Big Discovery)

The most surprising finding in this paper is that the detector got dimmer over time.

• The Observation: In the first two years, the amount of light the detector saw dropped by about 50%. It was like the sun slowly setting inside the tank.
• The Mystery: The team investigated everything. Did the argon get dirty? Did the coating on the lenses peel off? Did the electronics age?
  • They tested the argon for contaminants (like checking if the water in a pool is dirty).
  • They checked the lenses.
  • Result: They couldn't find a definitive "smoking gun." It's a mystery, but they know it happened.
• The Impact: Even though the light got half as bright, the camera was still sensitive enough to catch the neutrinos they cared about. They built a "dimming map" into their software to correct for this, ensuring their scientific results remained accurate.

4. The "Too Noisy" Eyes

Another surprise was that the "eyes" were much noisier than expected.

• The Expectation: They thought the background "buzz" (the random single flashes) would be quiet, like a library.
• The Reality: It was more like a busy coffee shop. The rate of these random flashes was about 200,000 times per second per eye.
• The Silver Lining: While this sounds like a problem, it actually helped them. Because the noise was so constant and predictable, it gave them a steady stream of data to calibrate their detectors without needing to stop the experiment to run special tests.

5. Catching the Shy Neutrinos (Efficiency)

The biggest question for the scientists was: "If the light got so dim, did we miss the faint, low-energy neutrinos?"

• The Test: They looked at cosmic rays (particles from space) that passed through the very back of the tank, where the light is weakest.
• The Result: Even in the darkest corner of the tank, with the light dimmed by half, the camera still caught 80-100% of the events it needed to. The "trigger" system was robust. It was like having a security camera that could still spot a thief even if the hallway lights were turned down to 50%.

Summary: What Does This Mean?

This paper is a testament to the reliability of the MicroBooNE experiment.

1. We can trust the data: Even though the detector changed over time (getting dimmer and noisier), the team figured out how to correct for it.
2. We learned a lot: They discovered that liquid argon detectors can have mysterious, long-term changes in how they see light.
3. Future-proofing: These lessons are crucial for the next generation of even bigger detectors (like DUNE). If we know how to handle a "dimming" detector, we can build better ones for the future.

In short, the MicroBooNE team spent five years watching a giant tank of liquid argon, learned that its "eyes" were slowly going blind and getting noisy, figured out how to fix the glasses, and confirmed that they still saw exactly what they needed to see.

Technical Summary

1. Problem Statement

The MicroBooNE experiment, a Liquid Argon Time Projection Chamber (LArTPC) located at Fermilab, relies heavily on scintillation light detection for two primary functions:

1. Triggering: Identifying beam-induced neutrino interactions amidst a high background of cosmic rays.
2. Timing ($t_0$): Determining the interaction time to reconstruct the drift position of ionization charge.

However, the long-term operation of the detector (2015–2021) revealed significant, unexplained time-dependent variations in the light detection system:

• A ~50% decline in light yield over the first two years of operation.
• A Single Photoelectron (SPE) noise rate of $\sim$200 kHz per PMT, which is significantly higher than the expected $\sim$50 kHz based on standard radio-impurities and intrinsic PMT noise.
• The need to quantify systematic uncertainties related to light modeling (attenuation, scattering, yield) to ensure the validity of physics analyses, particularly for low-energy neutrino interactions.

The paper addresses the challenge of calibrating these time-varying effects, measuring the efficiency of the light-based trigger in the lowest-light regimes, and characterizing the anomalous SPE rates to ensure robust physics results.

---

2. Methodology

The authors employed a combination of data-driven calibration, simulation, and dedicated R&D runs:

• PMT Gain Calibration:

  • Instead of using external LED/laser sources (which interrupt data taking), the team utilized the continuous stream of random Single Photoelectrons (SPEs) naturally occurring in the PMT waveforms.
  • A Constant Fraction Discriminator (CFD) algorithm isolated SPE pulses.
  • A multi-photoelectron fitting model (Poisson distribution for photon arrival convolved with Gaussian gain response) was used to extract the mean gain ($\mu_{SPE}$) for each PMT weekly.
  • High Voltage (HV) adjustments were made to maintain a target gain of $\sim$20 ADC/PE.

• Light Yield Stability Measurement:

  • Used Cosmic-Ray Muon Tracks (Anode-Piercing and Cathode-Piercing) as a stable, high-statistics calibration source. These tracks are reconstructed using charge information alone, minimizing bias from light variations.
  • Calculated light yield (PE/cm) over time by comparing the number of detected photons to the track length.
  • Applied a time-dependent calibration to Monte Carlo (MC) simulations to correct for the observed decline.

• Triggering Efficiency Measurement:

  • Focused on the "lowest-light" regime (near the cathode, farthest from PMTs) where trigger efficiency is lowest.
  • Used CRT-tagged cosmic muons (Coincidence with the Cosmic Ray Tagger) to select events without imposing a light-based trigger threshold.
  • Measured the fraction of events passing the standard 20 PE software trigger threshold as a function of reconstructed visible energy (30–400 MeV).

• Systematic Uncertainty Evaluation:

  • Generated alternative photon lookup libraries to test variations in:
    1. Uniform light yield decrease (25% reduction).
    2. Rayleigh scattering length (increasing from 66 cm to 120 cm).
    3. Attenuation length (reducing from 20 m to 8 m).

• SPE Rate Characterization:

  • Analyzed SPE rates over time and their dependence on the TPC electric field polarity and magnitude using dedicated R&D runs where the HV polarity was reversed (positive vs. negative).

---

3. Key Contributions

• Data-Driven Gain Calibration: Demonstrated a robust method for continuous PMT gain calibration using random background SPEs, eliminating the need for dedicated calibration runs that would interrupt physics data taking.
• First-Time Observations:
  • Reported a ~50% decline in light yield concentrated in the first two years of operation, which stabilized around 2018.
  • Reported an unexpectedly high SPE rate ($\sim$200 kHz) and its correlation with the electric field polarity and argon purity.
• Trigger Efficiency in Low-Light Regime: Provided the first direct measurement of scintillation light triggering efficiency for events near the cathode (the most difficult region to trigger), confirming high efficiency even after the light yield decline.
• Systematic Framework: Established a comprehensive set of light-response systematic uncertainties (yield, scattering, attenuation) to be applied in future MicroBooNE analyses.

---

4. Key Results

• PMT Gain Stability: The gain calibration successfully maintained PMT gains within $\pm$10% of the target (20 ADC/PE) throughout the 5-year run. The gain was found to be stable after HV adjustments in 2017 and 2018.
• Light Yield Decline:
  • A 35% decline was observed for Anode-Piercing Tracks (APT) and a 55% decline for Cathode-Piercing Tracks (CPT) between late 2015 and early 2018.
  • The decline stabilized after early 2018. A slight recovery was observed in August 2019, coinciding with a filter regeneration.
  • The cause remains unknown; impurity analysis (ICP-MS and GC-MS) did not find contaminants sufficient to explain the decline, and the electron lifetime (charge collection) did not correlate with the light yield drop.
• Triggering Efficiency:
  • For events near the cathode with 150 MeV visible energy, the trigger efficiency is >80%.
  • At 200 MeV, efficiency reaches ~100%.
  • The efficiency remains high even with the 50% light yield reduction, confirming the trigger's robustness for the experiment's flagship low-energy excess searches.
• SPE Rate Anomalies:
  • The average SPE rate is $\sim$200 kHz per PMT (integrated $\sim$8 MHz), roughly 4x higher than expected.
  • The rate declined by ~37% between 2016 and 2019, correlating with the light yield decline and argon purity changes.
  • Electric Field Dependence: Dedicated R&D runs showed that the SPE rate decreases as the electric field strength increases. The rate is lower in reverse polarity (positive HV) compared to normal polarity, suggesting a dependence on the microphysics of scintillation or charge neutralization processes.

---

5. Significance

• Physics Validation: The results confirm that the MicroBooNE physics program, particularly the search for the low-energy excess (a key goal regarding the LSND/MiniBooNE anomalies), was not compromised by the light yield decline. The trigger remained efficient, and systematic uncertainties were quantified.
• Benchmark for Future Detectors: As one of the longest-running surface LArTPCs, MicroBooNE provides critical "lessons learned" for next-generation detectors like DUNE and SBND.
  • The high SPE rate and its dependence on electric field polarity offer new insights into LArTPC noise mechanisms that must be modeled for future large-scale experiments.
  • The successful use of random SPEs for gain calibration offers a blueprint for maintaining detector performance without interrupting data taking.
• Systematic Uncertainty Management: The paper establishes a rigorous framework for handling time-dependent light response variations, ensuring that future analyses can correctly account for detector aging and environmental changes.

In summary, this paper provides a comprehensive characterization of the MicroBooNE light detection system, turning potential anomalies (light yield decline, high SPE noise) into well-understood systematic effects, thereby securing the integrity of the experiment's physics results and laying the groundwork for future LArTPC operations.