Impact of front-end parameters of the ARCADIA MD3 on charged particle detection

This paper presents the first in-beam characterization of the ARCADIA MD3 sensor, a 200 μm thick FD-MAPS developed using a customized LFoundry 110 nm process, to investigate how front-end parameters affect tracking performance when tested with a 120 GeV proton beam at Fermilab.

The Gist

Imagine you are trying to take a crystal-clear photo of a speeding bullet passing through a room. To do this, you need a camera that is incredibly fast, incredibly sensitive, and doesn't get overwhelmed by the sheer speed of the event.

This paper is about a team of scientists building a new, super-advanced "camera" for particle physics called the ARCADIA MD3, and they are testing how to tweak its settings to get the sharpest possible picture of subatomic particles.

Here is the story of their experiment, broken down into simple concepts:

1. The Camera: A "Thick" Sensor

Most standard particle sensors are like thin sheets of paper. The ARCADIA team built a sensor that is "thick" (about 200 micrometers, which is roughly the width of two human hairs).

• The Analogy: Think of a thin sensor as a single sheet of paper. If a bullet (a particle) hits it, it might just graze the surface. But the ARCADIA sensor is like a thick block of wood. When the bullet hits it, it travels deep inside, leaving a long, clear trail of "dust" (electric charge) behind it.
• Why it matters: Because the sensor is thick, it captures more of the particle's energy, making it easier to see and track, even if the particle is moving very fast or if the sensor gets damaged by radiation later on.

2. The Test: The "Bullet" Range

The team took their new sensor to Fermilab (a giant particle accelerator in the US) to test it.

• The Setup: They built a "telescope" out of three of these sensors. Two were used as reference cameras to know exactly where the particle should be, and the third one (the "Device Under Test") was the one they were trying to tune.
• The Beam: They fired a beam of protons (tiny particles) at the sensors. It's like shooting a stream of tiny bullets through a series of three windows to see how well the middle window records the path.

3. The Problem: Tuning the "Sensitivity Knobs"

The sensor has little electronic circuits inside every single pixel (the tiny squares that make up the image). These circuits have "knobs" (called bias currents) that control how sensitive they are.

• The Analogy: Imagine you are adjusting the volume and bass on a stereo system.
  • If the volume is too low, you miss the music (the particle).
  • If the volume is too high, the speakers distort and you hear noise instead of music.
  • The scientists needed to find the "Goldilocks" setting: not too loud, not too quiet, but just right.

They tested three specific knobs:

1. The Threshold Knob ($I_D$): How loud does the signal have to be before the sensor says, "I saw something!"?
2. The Feedback Knobs ($I_{BIAS}$ and $I_{FB}$): These control how the sensor resets itself after seeing a particle, like how quickly a microphone stops picking up sound after a loud noise.

4. The Results: Finding the Sweet Spot

The scientists ran the beam through the sensor while turning these knobs to different settings. Here is what they found:

• The "Cluster" Effect: When a particle hits the sensor, it doesn't just light up one tiny square; it usually lights up a small group of squares (a "cluster").
  • If the knobs were set too high (too sensitive), the sensor got confused and lit up too many squares, making the image "fuzzy."
  • If they turned the Feedback Knobs down to a specific setting, the "fuzziness" disappeared. The cluster of light became tighter and more precise.
• The Precision: By finding the perfect setting, they could pinpoint the particle's location with an accuracy of about 4.6 micrometers.
  • The Analogy: The sensor has a grid of squares (pixels). If you just count which square was hit, your accuracy is limited to the size of that square (like guessing a location to the nearest city block). But because the sensor is thick and the settings were tuned perfectly, they could tell you the location within a single street inside that city block. They achieved "sub-pixel" precision.

5. The Conclusion

The paper concludes that the ARCADIA sensor is a huge success.

• The "Thick" Advantage: The extra thickness helps spread the charge out, which actually helps the computer calculate the exact center of the hit more accurately.
• The Tuning Matters: It's not enough to just build a good sensor; you have to tune the electronic "knobs" correctly. The team found that the feedback knobs had a bigger impact on the picture quality than the threshold knob.

In Summary:
The scientists built a super-thick, high-tech camera for catching invisible particles. They went to a giant particle accelerator to test it, fiddled with the electronic volume and reset knobs, and discovered the perfect settings. With these settings, the camera can track particles with incredible precision—much better than the size of its own pixels would suggest. This technology could be used in future giant particle colliders or even in space telescopes to see the universe in sharper detail than ever before.

Technical Summary

1. Problem Statement

The ARCADIA INFN R&D project aims to develop a Fully Depleted Monolithic Active Pixel Sensor (FD-MAPS) using a customized LFoundry 110 nm CIS process. This technology is designed for future high-luminosity collider experiments (e.g., FCC-ee) and space/medical applications, requiring high radiation hardness, low power consumption, and excellent spatial resolution.

While the sensor architecture (thick, fully depleted substrate) is established, optimizing the front-end electronics is critical for performance. The specific problem addressed in this study is understanding how adjustable front-end bias currents and voltages affect the tracking performance (specifically spatial resolution and cluster formation) of the ARCADIA Main Demonstrator 3 (MD3) sensor. Without this optimization, the sensor may not meet the stringent requirements of future particle physics experiments.

2. Methodology

The study utilized an experimental approach involving beam tests and systematic parameter scanning:

• Device Under Test (DUT): The ARCADIA MD3 sensor with an active thickness of 200 µm.
• Experimental Setup:
  • Location: Fermilab Test Beam Facility (FTBF).
  • Beam: 120 GeV proton beam (Summer 2024).
  • Configuration: A trigger-less telescope consisting of two reference tracking planes (Planes 0 and 2) and the DUT (Plane 1) placed in between. All planes had a 200 µm active thickness and were biased to achieve full depletion (backside voltages: -80V to -90V).
  • Geometry: The DUT was placed perpendicular to the beam to evaluate orthogonal charged particle tracks.
• Data Acquisition: A trigger-less system with timestamp synchronization was used. Data was processed spill-by-spill using Corryvreckan software.
• Analysis Strategy:
  • Tracks were reconstructed using the two external reference planes.
  • Clusters on the DUT were associated with tracks based on spatial and temporal windows.
  • Parameter Scans: The study systematically varied specific front-end parameters controlled by Digital-to-Analog Converters (DACs):
    • $V_{CASN}$: Fixed at a reference value (DAC 5) to set the amplifier baseline.
    • $I_D$ (Discriminator Current): Controlled via a 2-bit DAC; directly proportional to the threshold.
    • $I_{FB}$ (Feedback Branch Current): Controlled via a 2-bit DAC; affects baseline level and return speed.
    • $I_{BIAS}$ (Main Branch Bias Current): Ideally kept equal to $I_{FB}$ to optimize the amplifier operating point.

3. Key Contributions

• First In-Beam Characterization: This paper presents the first in-beam performance data for the ARCADIA MD3 sensor with a 200 µm active thickness.
• Front-End Optimization Study: It provides a detailed analysis of how specific analog front-end currents ($I_D$, $I_{BIAS}$, $I_{FB}$) influence the sensor's physical performance metrics, specifically cluster size and spatial resolution.
• Demonstration of FD-MAPS Advantages: The study validates that the thick, fully depleted architecture enhances charge sharing, which is crucial for achieving spatial resolution superior to the binary limit defined by the pixel pitch.

4. Key Results

The analysis focused on cluster width (maximum elongation) and residual width (spatial resolution, $\sigma$ of the Gaussian fit of residuals).

• Impact of Discriminator Current ($I_D$):

  • Increasing $I_D$ led to a slight increase in single and double-pixel clusters.
  • The average cluster width decreased slightly.
  • The residual width increased slightly (variation < 5% between $I_D=0$ and $I_D=3$).
  • Conclusion: $I_D$ has a relatively minor impact on spatial resolution compared to other parameters.

• Impact of Bias Currents ($I_{BIAS}$ and $I_{FB}$):

  • These parameters had a significant impact on spatial resolution.
  • Optimal Setting: The minimum residual width was achieved when $I_{BIAS} = I_{FB} = 2$.
  • Degradation: A significant widening of the residual width occurred when $I_{BIAS} = I_{FB} = 3$.
  • Cluster Width: Increasing these currents significantly decreased the average cluster width.

• Performance Metrics:

  • Spatial Resolution: The minimum residual width achieved was 4.6 – 4.7 µm.
  • Cluster Size: This resolution was achieved with an average cluster width of 1.65 pixels.
  • Comparison to Binary Resolution: The achieved resolution (4.6–4.7 µm) is significantly better than the binary resolution limit of 7.2 µm (determined by the 25 µm pixel pitch), proving the effectiveness of charge sharing in the thick substrate.
  • Asymmetry: A slight asymmetry was observed where the column axis showed slightly larger widths than the row axis, attributed to minor mechanical misalignment of the DUT.

5. Significance

• Validation of Technology: The results confirm that the ARCADIA MD3, with its 200 µm fully depleted substrate, is a viable candidate for future high-energy physics trackers. The ability to achieve ~4.7 µm resolution with a 25 µm pitch demonstrates the power of the FD-MAPS architecture.
• Design Guidance: The study establishes that $I_{BIAS}$ and $I_{FB}$ are critical tuning parameters. Maintaining $I_{BIAS} \approx I_{FB}$ is essential for optimizing the amplifier operating point and maximizing spatial resolution.
• Future Outlook: This work lays the groundwork for further optimization. The authors note that a complete characterization of efficiency and resolution is ongoing, and these findings will guide the final design of sensors for the FCC-ee and other next-generation experiments.