Development of a Cherenkov-Based Time-of-Flight Detector Using Silicon Photomultipliers

This paper presents the development and validation of high-precision Time-of-Flight detectors using high-refractive-index Cherenkov radiators coupled with Silicon Photomultipliers, achieving a system-level time resolution better than 33.2 ps and 100% detection efficiency through rigorous optimization and CERN beam tests.

The Gist

Imagine you are trying to catch a speeding bullet. To do this, you need a camera that can take a picture so fast that the bullet looks frozen in mid-air. In the world of particle physics, scientists are trying to do exactly this, but instead of bullets, they are tracking subatomic particles like protons and pions moving at nearly the speed of light.

This paper describes a new, ultra-fast "camera" system designed to catch these particles and measure exactly when they arrive. The goal is to reach a timing precision of about 33 picoseconds. To put that in perspective: a picosecond is to a second what a second is to about 32 years. It's incredibly fast.

Here is how they did it, explained through simple analogies:

1. The "Flashbulb" Trick (Cherenkov Radiation)

Usually, to time something, you wait for it to hit a sensor. But particles move so fast that waiting for them to stop or interact can be too slow.

Instead, the scientists used a clever trick called Cherenkov radiation. Think of a supersonic jet breaking the sound barrier; it creates a loud "sonic boom." Similarly, when a charged particle moves through a transparent material (like glass) faster than light can travel through that material, it creates a "light boom."

• The Analogy: Imagine running through a pool of water faster than the ripples can spread. You create a cone-shaped wave behind you. That wave is the Cherenkov light.
• The Benefit: This light is emitted instantly (in picoseconds). It acts like a perfect, instantaneous flashbulb that goes off the moment the particle passes through.

2. The "Sensor Array" (SiPMs)

To catch this flash, they used Silicon Photomultipliers (SiPMs).

• The Analogy: Think of a standard camera sensor as a single large window. If a tiny speck of dust hits it, you might miss it. But imagine a wall made of thousands of tiny, individual windows (pixels). If a flash of light hits the wall, even if it's small, it will likely hit several of these tiny windows at once.
• The Innovation: They didn't just use one sensor; they used a grid (array) of these sensors. When a particle zips through a thin piece of glass (the radiator) placed in front of the sensors, it creates a "splash" of light that hits a small cluster of these tiny windows simultaneously.

3. The "Swarm" Strategy

The key to their success wasn't just catching the light, but how they processed it.

• The Problem: If you only look at the single brightest pixel, you might get a good time, but it's not perfect.
• The Solution: They looked at the whole "swarm" of pixels that lit up. By combining the timing signals from all the pixels in that cluster, they could average out the errors.
• The Analogy: Imagine trying to guess the exact time a race started by listening to one person shouting "Go!" You might be off by a fraction of a second. But if you have 100 people shouting "Go!" at the same time, and you average their voices, you get a much more precise moment of the start.

4. The "Glass Window" Optimization

They tested different types of glass (radiators) and how to glue them to the sensors.

• The Challenge: Light bounces around. If the glass isn't glued perfectly to the sensor, light bounces off and is lost, or it bounces around too much and confuses the timing.
• The Fix: They found that using a specific type of glass (Fused Silica) and a special "optical glue" (silicone resin) acted like a perfect highway for the light, guiding it straight into the sensors without losing any energy. They also figured out that a glass thickness of about 1 millimeter was the "Goldilocks" zone—not too thin (not enough light) and not too thick (too much blurring).

5. The "Super-Brain" Electronics

Even with perfect sensors, if the computer reading the data is slow, the whole system fails.

• The Hardware: They tested two different electronic "brains" (chips) to read the sensors. One was good, but the other (called RadioPico) was like a Formula 1 car compared to a sedan. It had incredibly low "jitter" (electronic noise).
• The Result: When they combined the perfect glass setup with the super-fast RadioPico electronics, they achieved the world-class timing of 33.2 picoseconds.

Why Does This Matter?

In the future, particle accelerators (like the Large Hadron Collider) will be running much faster, creating millions of collisions every second. It's like trying to hear a single conversation in a stadium full of screaming fans.

• The "4D" Solution: Current detectors can tell you where a particle is (3D). This new technology adds the 4th dimension: Time.
• The Benefit: By knowing the exact time a particle arrived (down to 33 picoseconds), scientists can separate the "real" signal from the "noise" of other collisions happening at the same time. It's like having a super-fast shutter speed that freezes the action, allowing them to see individual particles clearly even in a chaotic crowd.

Summary

The team built a high-speed particle detector that uses a "light boom" (Cherenkov radiation) to create an instant flash. They caught this flash with a grid of tiny sensors, averaged the signals from the whole group to cancel out errors, and used ultra-fast electronics to read the result. The result is a detector that is 100% efficient at catching particles and can time them with a precision that was previously thought impossible for this type of device. This paves the way for the next generation of experiments that will explore the deepest secrets of the universe.

Technical Summary

1. Problem Statement

The primary challenge in modern high-energy physics (HEP) is the need for high-precision Time-of-Flight (TOF) detectors capable of identifying charged particles over broad momentum ranges, particularly in high-luminosity environments where pile-up suppression and 4D tracking are critical.

• Limitations of Current Tech: Traditional Ring-Imaging Cherenkov (RICH) detectors are effective for particle identification (PID) but often lack the sub-30 ps timing resolution required for next-generation TOF systems. Microchannel Plate Photomultipliers (MCP-PMTs) offer excellent timing but suffer from high cost, sensitivity to magnetic fields, and large material budgets.
• The SiPM Gap: Silicon Photomultipliers (SiPMs) offer magnetic field insensitivity, low material budget, and high Photon Detection Efficiency (PDE). However, using them for direct Cherenkov detection is inefficient because Cherenkov photons generated within the SiPM's protective resin are confined to a single pixel, leading to poor detection efficiency and limited timing precision due to the small number of photoelectrons (PEs).

2. Methodology

The authors propose a novel approach: coupling a thin, high-refractive-index transparent window (acting as a Cherenkov radiator) directly to a SiPM array. This generates a cluster of fired pixels rather than a single hit, improving efficiency and timing statistics.

Key Methodological Steps:

• Radiator Optimization: The study focused on optimizing the radiator material (Fused Silica/SiO₂, MgF₂, NaF) and thickness. High refractive index ($n$) materials lower the momentum threshold for Cherenkov emission and increase photon yield. SiO₂ was selected for its radiation hardness and precise thickness control.
• Monte Carlo Simulations: A comprehensive Geant4-based simulation framework was developed to model:
  • Cherenkov photon generation (Frank-Tamm formula).
  • Optical transport, including reflections at interfaces (window, resin, SiPM).
  • SiPM response (PDE, SPAD pitch, fill factor).
  • Electronic noise and jitter.
  • The simulation identified a trade-off: thinner radiators reduce geometric time spread but lower photon yield, degrading statistical timing resolution.
• Reflectance Characterization: The team measured total, diffuse, and specular reflectance of various SiPM arrays (Hamamatsu S13361 series) with different protective resins (epoxy vs. silicone) and window materials. They developed a model to extrapolate the reflectance at the SiPM internal interfaces (Resin-ARC-Passivation-Silicon) to optimize Anti-Reflective Coatings (ARC).
• Experimental Setup (Beam Tests):
  • Location: CERN-PS T10 beamline (10 GeV/c negative beam: ~95% pions).
  • Prototype: A telescope of two SiPM arrays (A0 and A1) separated by 28 cm, flanked by fiber trackers (T0, T1) for triggering and tracking.
  • Configurations: Tested various SiPM sizes (1.3 mm, 2.0 mm, 3.0 mm), SPAD pitches (50 µm, 75 µm), and window materials (SiO₂, MgF₂) with thicknesses of 1 mm and 2 mm.
  • Electronics: Two readout architectures were compared:
    1. Petiroc 2A: 32-channel ASIC with 40 ps LSB TDC.
    2. Radioroc 2 + picoTDC: 64-channel ASIC coupled to a picoTDC with 3.05 ps LSB, offering superior timing resolution.
  • Cooling: Arrays were cooled to 0°C to minimize Dark Count Rate (DCR).

3. Key Contributions

• Cluster-Based Timing Strategy: Demonstrated that coupling a radiator window to a SiPM array creates a "cluster" of fired pixels. By averaging timestamps from multiple pixels (especially those with high charge), the system overcomes the "dead area" limitations of single-pixel detection and achieves 100% charged-particle detection efficiency.
• Optical Coupling Optimization: Identified that silicone resin coupled with SiO₂ windows provides the lowest reflectance and best UV transmission compared to epoxy resin, significantly reducing photon loss.
• ARC Modeling: Developed a model to characterize internal SiPM reflectance and proposed custom multi-layer ARCs (e.g., CeF₃/ZnS or HfO₂/TiO₂) that could theoretically reduce reflectance by a factor of two compared to commercial coatings.
• Electronics Benchmarking: Quantified the impact of front-end electronics on timing resolution, showing that low-jitter ASICs (Radioroc 2) are essential to unlock the intrinsic potential of the sensor.

4. Results

• Detection Efficiency: Achieved >99.9% detection efficiency for charged particles with thresholds up to 14 PEs (for 3x3 mm² SiPMs) and 7 PEs (for 1.3x1.3 mm² SiPMs). This confirms the approach is robust even in high-radiation environments where DCR suppression is necessary.
• Time Resolution:
  • Single Array (Max Charge): Using the Petiroc board, the single-array time resolution was approximately 53.0 ps (for 3x3 mm² SiPMs with >30 PEs).
  • System Level (Best Performance): Using the Radioroc 2 + picoTDC electronics, the system achieved a time resolution of 33.2 ps for the S13361-3075 array (3x3 mm², 75 µm pitch) coupled to a 1 mm SiO₂ window.
  • Cluster Averaging: For smaller SiPMs (1.3 mm) or tracks hitting pixel boundaries, averaging the timestamps of multiple fired pixels improved resolution, approaching the ideal $1/\sqrt{N_{SiPM}}$ scaling.
• Comparison with Bare SiPMs: A control test with bare SiPMs (no window) showed a time resolution of 186 ps and significantly lower efficiency (30% non-response), highlighting the critical role of the radiator window in signal generation.
• Material Performance: SiO₂ and MgF₂ windows showed similar performance due to close refractive indices, but SiO₂ with silicone resin offered the best overall optical coupling.

5. Significance

This work establishes a viable path toward sub-30 ps TOF detectors for future HEP experiments (e.g., at the High-Luminosity LHC).

• Scalability: The use of SiPMs allows for large-area, cost-effective detectors that are insensitive to magnetic fields, unlike MCP-PMTs.
• Radiation Hardness: The use of fused silica and SiPMs ensures stability in high-luminosity environments.
• Future Impact: The demonstrated 33.2 ps resolution, combined with 100% efficiency, meets the stringent requirements for pile-up suppression and 4D tracking in next-generation colliders. The paper also provides a roadmap for further improvements via custom ARCs and optimized readout electronics, pushing the limits of Cherenkov-based timing technology.