Test-Beam Performance of the AstroPix Silicon Sensor for Imaging Calorimetry

This paper presents the first 2025 test-beam results of the AstroPix-v3 HVCMOS sensor, demonstrating its stable performance, high hit efficiency, and capability to effectively discriminate between electromagnetic and hadronic showers when integrated with a Pb/SciFi calorimeter, thereby validating its suitability for future space-based gamma-ray missions and the ePIC experiment at the Electron-Ion Collider.

The Gist

Imagine you are trying to take a high-speed photograph of a firework exploding in the dark. You need a camera that is incredibly fast, very sensitive, and can see exactly where every spark lands. This is essentially what the scientists in this paper are trying to do, but instead of fireworks, they are studying tiny particles of energy (like electrons and pions) crashing into a detector.

Here is a breakdown of their work using simple analogies:

The Camera: "AstroPix"

The main character in this story is a new type of digital sensor called AstroPix. Think of it as a super-advanced, high-resolution digital camera chip.

• What it is: It's a "High-Voltage CMOS" sensor. In plain English, it's a silicon chip that can handle a strong electrical "push" (voltage) to make its internal layers deeper. This helps it catch particles better and faster.
• The Goal: The scientists built this chip for two main jobs:
  1. Space Missions: To act as the "eye" of future telescopes looking for gamma rays from space.
  2. Particle Colliders: To be the "imaging layer" inside a massive machine called the Electron-Ion Collider (EIC), helping to see how particles break apart.

The Experiment: The "Test Drive"

Before putting this new camera on a real spaceship or a giant collider, the team needed to test it. They took the third version of their chip (called AstroPix-v3) to two different "test tracks" (particle beam facilities) in Japan (KEK) and Switzerland (CERN).

They set up two different scenarios to see how the camera performed:

Scenario A: The Solo Run (Standalone Mode)
They let the camera sit alone in the path of a beam of particles.

• The Result: They found that the camera works best when you give it a strong electrical "push" (a bias voltage of -400 Volts). At this setting, it successfully caught about 68% of the particles hitting it.
• The Catch: It didn't catch 100% of them because the "active" part of the chip wasn't fully deep yet. The scientists say future versions will be deeper and catch even more.

Scenario B: The Sandwich Run (Interleaved Mode)
This is the more complex and exciting part. They built a "sandwich."

• The Layers: They placed layers of the AstroPix camera between blocks of lead and special plastic fibers (called Pb/SciFi).
• The Analogy: Imagine throwing a ball into a stack of thick blankets.
  • If you throw a light, bouncy ball (an electron), it bounces around wildly, creating a wide, messy cloud of sparks as it hits the blankets.
  • If you throw a heavy, dense rock (a pion/hadron), it punches straight through with very little bouncing or spreading.
• The Test: The scientists shot both types of particles at their sandwich.
  • The Camera's Job: The AstroPix layers acted like a high-speed security camera, taking pictures of the "sparks" (hits) as they traveled through the sandwich.
  • The Synchronization: Since the camera takes pictures continuously (like a video stream) but the other detectors only take pictures when triggered, the team had to use a "master clock" to sync everything up perfectly. They succeeded, ensuring every picture was time-stamped correctly.

The Big Discovery: Telling the Difference

The most important result was that the AstroPix camera could clearly tell the difference between the "bouncy ball" (electron) and the "heavy rock" (pion).

• Electrons (The Fireworks): When an electron hit the sandwich, the camera saw a wide spread of hits. The sparks flew far apart, creating a large, fuzzy cloud. The number of sparks also increased as the particle went deeper.
• Pions (The Rocks): When a pion hit the sandwich, the camera saw a tight, narrow line of hits. The particle didn't spread out much.

By looking at how "spread out" the hits were and how many hits there were, the camera could instantly identify what kind of particle it was looking at.

The Conclusion

The paper concludes that this new "AstroPix" camera works exactly as hoped.

1. It is stable and reliable.
2. It can take clear, high-detail pictures of how particles spread out (shower development).
3. It is excellent at distinguishing between different types of particles based on how they scatter.

Because it works so well in these tests, the scientists are confident it is ready to be used in future space telescopes and inside the massive particle colliders to help us understand the universe better.

Technical Summary

Technical Summary: Test-Beam Performance of the AstroPix Silicon Sensor for Imaging Calorimetry

Problem Statement
High-Voltage CMOS (HVCMOS) Monolithic Active Pixel Sensors (MAPS) are evolving as a critical technology for next-generation gamma-ray telescopes and particle physics experiments. These sensors offer high spatial resolution and low material budgets, essential for precise event reconstruction and background suppression in the medium- to soft-gamma-ray energy range. While the AstroPix sensor, derived from the ATLASPix chip, has been developed for space-based missions (e.g., AMEGO-X) and proposed as an imaging layer for the Barrel Imaging Calorimeter (BIC) at the future Electron–Ion Collider (EIC), its performance in a controlled beam environment had not been previously reported. Specifically, there was a lack of data regarding hit efficiency, position and energy resolution, timing performance, and the sensor's ability to distinguish between electromagnetic (EM) and hadronic showers when interleaved with calorimeter modules.

Methodology
To address these gaps, the Korean Barrel Imaging Calorimeter (KoBIC) group conducted test-beam campaigns in 2025 at two facilities: the KEK Photon Factory Advanced Ring (PF-AR) and the CERN Proton Synchrotron (PS) T10 beam line. The study utilized the third prototype, AstroPix-v3, a full-reticle sensor with a $2 \times 2 \text{ cm}^2$ active area, a $35 \times 35$ pixel matrix, and a $500 \, \mu\text{m}$ pixel pitch, fabricated with a deep n-well (DNW) structure in a $720 \, \mu\text{m}$ bulk.

The experimental setup involved two primary configurations:

1. Standalone Mode: Three AstroPix-v3 chips were spaced 6 cm apart to function as tracking layers.
2. Interleaved Mode: AstroPix-v3 layers were placed between prototype lead/scintillating-fiber (Pb/SciFi) calorimeter modules to simulate the BIC design.

Data acquisition (DAQ) presented a synchronization challenge, as AstroPix-v3 operated in continuous readout mode while the Pb/SciFi calorimeter and trigger detectors used a trigger-based scheme. The team implemented a common global timestamp system using a 125 MHz Trigger Control Board (TCB). AstroPix timestamps were corrected by subtracting the minimum Time-over-Threshold (ToT) value to account for intrinsic delays, and events were matched if the time difference ($\Delta t$) fell within a 20 $\mu$s window.

Beam conditions included a high-purity 4.5 GeV/c electron beam at KEK and a mixed electron/pion beam (0.5–11.5 GeV/c) at CERN. Particle identification (PID) at CERN was achieved using threshold Cherenkov counters, allowing the separation of electron-induced (EM) and pion-induced (hadronic) events.

Key Contributions and Results

• Synchronization and Timing: The study successfully demonstrated a robust synchronization scheme between continuous-readout MAPS and trigger-based calorimeters. Timing offsets were confined within 5 $\mu$s, and spatial correlations between layers confirmed reliable event matching.
• Hit Efficiency: In standalone mode using an 8 GeV/c pion-dominated beam, the hit efficiency of AstroPix-v3 was measured as a function of bias voltage. Efficiency increased monotonically with voltage, reaching a maximum of approximately 68% at −400 V. The authors attribute this to the partial depletion depth (estimated at ~100 $\mu\text{m}$) of the current sensor design, noting that future iterations with deeper depletion are expected to yield higher efficiencies.
• Electromagnetic Shower Imaging: When interleaved with Pb/SciFi modules, AstroPix-v3 successfully captured the lateral development of EM showers. Hit maps and position residual distributions showed a progressive broadening of the spatial hit pattern from upstream to downstream layers, consistent with the generation of secondary electrons and photons within the calorimeter.
• EM vs. Hadronic Shower Separation: Using Cherenkov-based PID, the system demonstrated clear discrimination between electron and pion events.
  • Hit Multiplicity: Electron-induced events exhibited significantly higher hit multiplicities in downstream layers compared to pion-induced events, which remained dominated by single-hit events. At 4 GeV/c, the fraction of multi-hit events for electrons exceeded 60%, whereas pions showed no significant momentum dependence.
  • Spatial Distribution: Electron-induced events displayed broader radial position residuals ($dR$) relative to the upstream reference layer, reflecting lateral EM shower development. In contrast, pion-induced events remained narrowly distributed, reflecting the delayed onset and limited development of hadronic showers in the tested volume.

Significance and Claims
The paper claims that these results constitute the first test-beam validation of the AstroPix-v3 sensor in a combined tracking and calorimetry environment. The findings demonstrate that AstroPix-v3 provides effective, high-granularity imaging of shower development. Specifically, the sensor's ability to distinguish between EM and hadronic showers using hit multiplicity and spatial correlation observables supports its viability as an imaging layer for the BIC in the ePIC experiment at the future EIC. Furthermore, the results reinforce AstroPix's potential for future space-based medium-energy gamma-ray missions, where precise two-dimensional hit-position information is essential for event reconstruction and background suppression. The authors conclude that while the current efficiency is limited by the sensor's depletion depth, the technology is well-suited for the intended applications, with future iterations expected to improve performance.