Development of a comprehensive PMT optical model for the JUNO experiment

This study establishes a comprehensive, individualized optical model for the 17,612 PMTs in the JUNO experiment by integrating mass testing and reflectance data to map photocathode and anti-reflective coating thicknesses, thereby refining detector simulations and energy response accuracy beyond previous uniform assumptions.

The Gist

Imagine the JUNO experiment as a giant, underwater camera trying to take a picture of invisible particles called neutrinos. To do this, it uses a massive sphere filled with a special glowing liquid. Surrounding this sphere are over 17,000 giant "eyes" called Photomultiplier Tubes (PMTs). These eyes are designed to catch the faint flashes of light produced when a neutrino interacts with the liquid.

For the camera to take a perfect picture, the scientists need to know exactly how each of these 17,000 eyes sees the world. However, not all eyes are identical, and even a single eye doesn't see light the same way across its entire surface.

This paper is about building a much better "instruction manual" for how these eyes work. Here is the breakdown in simple terms:

1. The Problem: The "One-Size-Fits-All" Mistake

Previously, scientists treated all the giant eyes of the same brand as if they were clones. They assumed that the light-sensitive coating on the front of every eye was perfectly smooth and uniform, like a factory-made sheet of glass.

But in reality, these coatings are more like hand-painted canvases. The thickness of the paint (the light-sensitive layer) varies slightly from one eye to another, and even across the surface of a single eye. Some spots are thicker, some are thinner. This means some parts of an eye catch light better than others, and some eyes reflect light differently than their neighbors. The old "uniform" model was like assuming every person in a crowd has the exact same height and weight—it's a useful average, but it's not accurate enough for high-precision science.

2. The Solution: A "Fingerprint" for Every Eye

The team in this paper created a comprehensive optical model. Think of this as giving every single one of the 17,612 eyes its own unique fingerprint.

To do this, they didn't just guess; they measured.

• The Reflectance Test: They shined a light on 669 of these giant eyes and measured how much light bounced off them (like checking how shiny a mirror is). They found that the "shininess" varied wildly between different brands and even different spots on the same eye.
• The Efficiency Test: They used data from previous tests to see how many photons (light particles) each eye actually caught.

By combining these two sets of data, they could work backward to figure out the thickness map of the coatings on every single eye. It's like looking at a shadow and deducing the exact 3D shape of the object casting it.

3. The Analogy: The Sunglasses and the Lens

Imagine the PMT is a pair of sunglasses.

• The ARC (Anti-Reflective Coating): This is like a special anti-glare spray on the lens. If the spray is too thick in one spot and too thin in another, some light bounces off (wasted) while some gets through. The paper mapped out exactly how thick this spray is on every part of every lens.
• The PC (Photocathode): This is the film inside the glasses that turns light into an electrical signal. If the film is uneven, some areas are super-sensitive, and others are dull. The paper mapped out this unevenness too.

4. The Results: A New Reality

When they compared their new, detailed model against the old, simple model, they found some surprising differences:

• For the "HPK" brand eyes: The new model says they reflect more light than we thought.
• For the "NNVT" brand eyes: The new model says they reflect significantly less light (up to 40% less in some cases) than the old model predicted.
• The Catch: While the amount of light caught (the efficiency) only changed by a small amount (a few percent), the amount of light bouncing around inside the detector changed a lot.

Why This Matters

In the JUNO experiment, light doesn't just travel in a straight line; it bounces off the walls and the eyes before being caught. If you get the "bouncing" (reflectance) wrong, your calculation of the energy of the neutrino is wrong.

By creating this detailed, eye-by-eye map, the scientists can now simulate the detector's behavior with much higher precision. It's the difference between using a blurry, low-resolution map to navigate a city versus using a high-definition GPS that knows exactly where every pothole and traffic light is. This ensures that when JUNO finally detects a neutrino, the scientists can trust the data they get.

In short: They stopped treating 17,000 complex cameras as identical clones and started treating them as the unique, slightly imperfect, hand-crafted instruments they actually are. This makes the entire experiment more accurate.

Technical Summary

Technical Summary: Development of a Comprehensive PMT Optical Model for the JUNO Experiment

Problem Statement
The Jiangmen Underground Neutrino Observatory (JUNO) utilizes 17,612 20-inch photomultiplier tubes (PMTs) to detect neutrinos. Precise detector simulations are critical for determining the neutrino mass ordering and measuring oscillation parameters. However, existing optical models relied on simplified assumptions, specifically that PMTs of the same type possess uniform photocathode (PC) and anti-reflective coating (ARC) properties. This study addresses the limitations of such models by accounting for the non-uniformity of photon detection efficiency (PDE) across the PMT surface and the significant variations in PDE and reflectance among individual PMTs. These factors are crucial because the operating environment (water immersion) and photon incidence angles in JUNO differ significantly from mass testing conditions in air, leading to event-dependent PDE variations that simplified models cannot accurately capture.

Methodology
The authors developed a comprehensive PMT optical model by extending a previous framework [6] to incorporate empirical data from mass production testing and new reflectance measurements. The methodology involves the following steps:

1. Data Integration: The model utilizes PDE data from two sources: the "container system" (measuring average PDE for all 17,612 PMTs) and the "scanning station system" (measuring PDE uniformity for a subset of ~3,200 PMTs at seven specific zenith angles). Additionally, new reflectance data were collected for 669 PMTs (276 HPK, 115 NNVT normal-QE, and 174 NNVT high-QE) at four zenith angles using an integrating sphere setup.
2. Model Extension: The model assumes that while the spectral responses of the refractive index ($n$) and extinction coefficient ($k$) are identical for PMTs of the same type, the thicknesses of the PC and ARC vary, causing non-uniformity. The thickness distribution is modeled as a function of the zenith angle ($\theta$) using an empirical quadratic formula.
3. Parameter Extraction: A chi-squared minimization procedure simultaneously fits the measured reflectance and PDE data to constrain the thickness maps of the PC and ARC for each PMT. This process decouples the quantum efficiency (QE) and collection efficiency (CE) from the measured PDE.
4. Simulation and Correction: A Geant4-based optical simulation was employed to distinguish between direct photon absorption and absorption of transmitted photons within the PMT structure. This allowed for the correction of PDE data to isolate the true absorptance at the incident position.
5. Spectral Extension: Using dispersion relations and literature data, the optical properties ($n$, $k$, and escape factor) were extended from the original 390–500 nm range to cover the full JUNO spectrum of 300–700 nm.

Key Contributions

• Thickness Mapping: The study successfully derived thickness maps for the PC and ARC for all 17,612 20-inch PMTs, revealing non-uniformities that were previously unaccounted for.
• Collection Efficiency (CE) Determination: By decomposing the PDE data and accounting for internal optical processes, the authors determined the CE as a function of the zenith angle for different PMT types.
• Comprehensive Optical Parameters: The refractive index, extinction coefficient, and escape factor were evaluated across the 300–700 nm range for three distinct PMT types: HPK dynode PMTs, NNVT normal-QE PMTs, and NNVT high-QE PMTs.
• Reflectance Characterization: New, high-precision reflectance measurements were established, correcting for surface curvature effects and providing a dataset for PMTs not covered by the scanning station.

Results

• Reflectance Variations: The comprehensive model predicts significantly different reflectance values compared to the simplified uniform model. For HPK PMTs, the reflectance is predicted to be higher, whereas for NNVT PMTs (both normal and high-QE), the reflectance is lower by approximately 30% to 40%.
• PDE Impact: The changes in the average PDE are more modest, remaining at the level of a few percent for small angles of incidence (AOI) but reaching 5% to 10% at large AOIs.
• Thickness Distributions: The fitted thickness maps show distinct dependencies on the zenith angle for different PMT types. For instance, the ARC thickness for NNVT normal-QE PMTs was found to have a negligible effect on reflectance and absorptance, leading to a fixed value in the model.
• CE Behavior: The calculated CE generally decreases as the zenith angle increases. Some values exceeding unity were observed, attributed to the assumption of a constant escape factor spectrum across PMTs of the same type, which prevents full decoupling of CE and escape factor variations with the available data.

Significance
The paper asserts that this comprehensive PMT optical model is essential for enhancing the accuracy of JUNO detector simulations and refining the energy response model. By moving beyond the assumption of uniform optical properties, the model provides a more realistic representation of how light interacts with the detector components. The authors state that this model will be integrated into the JUNO simulation program to evaluate its impact on energy resolution, non-uniformity, and non-linearity. Furthermore, the comprehensive approach offers a valuable reference for other PMT-based applications seeking to improve detector performance through precise optical modeling. The study concludes that while the PDE changes are moderate, the significant shifts in reflectance predictions necessitate this refined model for high-precision physics analysis.