Assessment of the Imaging Performance of the CITIUS High-Resolution Detector for Heavy Charged Particles and Neutrons

This paper demonstrates that the CITIUS high-resolution X-ray detector, through its gain-selecting architecture and substantial charge sharing enabled by long carrier drift distances, achieves significantly improved spatial resolution for imaging heavy charged particles and neutrons, as validated by experimental irradiation with alpha particles and Geant4 simulations.

The Gist

Imagine you have a super-sensitive camera, but instead of taking pictures of your cat or a sunset, it's designed to take "photos" of invisible, high-speed particles like alpha particles (tiny helium nuclei) and neutrons. This camera is called CITIUS.

Originally, scientists built CITIUS to take incredibly sharp pictures of X-rays for a giant machine called SPring-8. But the researchers in this paper asked a fun question: "Can this X-ray camera also take great pictures of heavier, tougher particles?"

Here is the story of how they found out, explained with some everyday analogies.

1. The Problem: The "Fuzzy" Photo

When a heavy particle hits a thick piece of silicon (the camera's sensor), it doesn't just stop at one tiny spot. Think of it like dropping a heavy stone into a deep, thick pool of honey.

• The stone creates a splash (the particle hits).
• But the ripples spread out as they travel through the honey (the electrical charge spreads out as it moves through the silicon).
• By the time the ripples reach the surface (the camera's pixels), they have spread over a wide area.

If the camera tries to guess where the stone hit based on this spread-out ripple, the photo looks blurry. The researchers wanted to know: How blurry is it, and can we fix it?

2. The Experiment: Testing the Camera

To test this, they used a source of alpha particles (from a radioactive element called Americium-241) and shot them at the CITIUS camera. They changed the "tension" of the camera sensor (called back-bias voltage) to see how the charge spread.

They built a virtual twin of the experiment on a computer (using a tool called Geant4). It was like creating a video game simulation of the experiment. They tweaked four "knobs" in the simulation until the virtual results matched the real-world photos perfectly:

1. Source fuzziness: How much the energy of the alpha particles varied naturally.
2. Coating mix: How much gold was in the protective layer of the source.
3. Charge spread: How much the "ripples" spread out in the silicon.
4. Camera noise: The static or "grain" in the camera's electronic signal.

The Result: They found that even when the sensor was fully "stretched out" (high voltage), the charge still spread out quite a bit (about 26.5 micrometers). This is actually a good thing! It means the charge is shared between many pixels, giving the camera lots of data to work with.

3. The Magic Trick: The "Smart" Gain Selector

Here is the most exciting part. The CITIUS camera has a special feature called a Gain-Selecting Architecture.

Imagine you are listening to a band.

• The "Medium" setting is like a standard volume knob. It's good for loud instruments but gets distorted if the music is too loud.
• The "High" setting is like a super-sensitive microphone. It hears the quietest whispers perfectly but gets blown out by loud drums.

Usually, you have to pick one setting. If the particle is strong, you use the "Medium" setting, and the image is a bit blurry. If the particle is weak, you use "High," but you might miss the loud parts.

CITIUS is different. It's like a smart sound system that listens to each pixel individually.

• If a pixel sees a huge splash of charge (a loud drum), it automatically switches to the "Medium" setting to handle it without distortion.
• If a pixel sees a tiny ripple (a whisper), it switches to the "High" setting to hear it clearly.

By mixing these settings, the camera can calculate the exact center of the splash with incredible precision.

4. The Results: From Blurry to Crystal Clear

The researchers ran simulations to see how sharp the pictures would be for alpha particles and neutrons (using a special boron layer to catch neutrons).

• Without the smart trick (Single Gain): The picture of an alpha particle was about 9.1 micrometers wide (a bit fuzzy).
• With the smart trick (Multi Gain): The picture sharpened down to 1.2 micrometers!

That is like going from a blurry, pixelated photo to a 4K HD image. For neutrons, the improvement was even more dramatic, going from a blurry 26 micrometers down to a crisp 1.9 micrometers.

The Bottom Line

This paper proves that CITIUS, originally built for X-rays, is actually a superstar for heavy particles and neutrons too.

Because the silicon is thick (allowing the charge to spread out like a wide ripple) and the camera is "smart" enough to adjust its sensitivity pixel-by-pixel, it can pinpoint exactly where a particle hit with amazing accuracy. It's like having a detective that can look at a messy crime scene (the spread-out charge) and figure out exactly where the bullet landed, even if the bullet was moving incredibly fast.

The scientists are now preparing to build real experiments to show off this new superpower!

Technical Summary

1. Problem Statement

While the CITIUS detector was originally developed as a high-speed, high-resolution X-ray imaging detector for the SPring-8-II synchrotron facility, its potential application for heavy charged particles (e.g., alpha particles) and neutrons remained unquantified.

• The Challenge: Imaging these particles requires thick silicon sensors to ensure sufficient interaction depth. However, thick sensors introduce significant charge diffusion as carriers drift to the readout nodes.
• The Gap: Existing mobility and recombination models often lack accuracy under high-injection conditions (where transient charge density exceeds the silicon impurity concentration), making it difficult to predict spatial resolution and optimize detector performance without empirical characterization.
• Objective: To experimentally characterize the CITIUS sensor's response to alpha particles, determine key physical parameters (diffusion, noise), and validate a simulation model to predict spatial resolution for both alpha particles and neutrons.

2. Methodology

The study employed a combination of experimental irradiation and Monte Carlo simulation:

A. Experimental Setup

• Detector: CITIUS sensor with $728 \times 384$ square pixels ($72.6 \, \mu\text{m}$ pitch) and a $650 \, \mu\text{m}$ thick depleted silicon substrate.
• Source: An $^{241}\text{Am}$ source emitting alpha particles ($\approx 5.5 \, \text{MeV}$), placed $9.4 \, \text{mm}$ from the sensor backside.
• Conditions: Measurements were taken at four back-bias voltages ($V_b$): 400 V, 300 V, 200 V, and 170 V (spanning above and below the full depletion voltage of 200 V).
• Readout Mode: Initially operated in a single-gain (medium-gain) configuration to characterize charge diffusion with minimal ambiguity.

B. Simulation and Modeling (Geant4)
A Geant4 model was constructed to replicate the experiment, featuring four tunable parameters determined via template fitting against measured signal cluster shapes ($m_0$ and $m_2$):

1. $\sigma_s$ (Source Energy Spread): Intrinsic Gaussian broadening of the alpha source energy.
2. $r$ (Gold Fraction): Composition ratio of the unknown Au–Pd coating on the source.
3. $\sigma_d$ (Lateral Charge Diffusion): The spread of charge after drifting $650 \, \mu\text{m}$.
4. $\sigma_n$ (Readout Noise): Per-pixel noise in the medium-gain channel.

C. Resolution Evaluation
Using the calibrated model, simulations were run for:

• 4 MeV Alpha Particles: Varying incident angles and pixel sizes.
• Cold Neutrons: Simulating the $^{10}\text{B}(n, \alpha)^7\text{Li}$ reaction with a $200 \, \text{nm}$ boron converter layer.
• Gain Architectures: Comparing Single-Gain (medium only) vs. Multi-Gain (dynamic switching between high and medium gain based on charge magnitude).

3. Key Contributions

• Quantitative Characterization: Successfully extracted physical parameters of the CITIUS sensor under high-injection conditions, a regime where standard models are often inaccurate.
• Template Fitting Framework: Developed a robust method to simultaneously determine source properties (energy spread, coating composition) and sensor properties (diffusion, noise) by fitting simulated cluster shape distributions to experimental data.
• Validation of Gain-Selecting Architecture: Demonstrated that CITIUS's ability to switch gain modes dynamically significantly enhances spatial resolution for heavy particles, a feature previously optimized for X-rays but now proven effective for neutrons and alphas.

4. Key Results

A. Calibrated Parameters (Best-Fit Values)

• Source Energy Spread ($\sigma_s$): $5\%$ (consistent across all voltages).
• Au–Pd Coating Gold Fraction ($r$): $0.4$.
• Readout Noise ($\sigma_n$): $10,000 \, e^-$ (medium-gain channel), showing no significant variation with back-bias voltage.
• Charge Diffusion ($\sigma_d$) at 400 V: $26.5 \, \mu\text{m}$ over a $650 \, \mu\text{m}$ drift distance.
  • Note: Diffusion increased as voltage decreased ($40.5 \, \mu\text{m}$ at 170 V), but sufficient charge sharing was observed even at the highest bias.

B. Spatial Resolution Improvements
The study evaluated resolution using the Line Spread Function (LSF) standard deviation. The Multi-Gain configuration (switching to high gain for low charge pixels) yielded dramatic improvements over Single-Gain:

Particle Type	Pixel Size	Single-Gain Resolution	Multi-Gain Resolution	Improvement
Alpha Particles (4 MeV)	$70 \, \mu\text{m}$	$9.1 \, \mu\text{m}$	$1.2 \, \mu\text{m}$	~7.6x
Cold Neutrons	$70 \, \mu\text{m}$	$26 \, \mu\text{m}$	$1.9 \, \mu\text{m}$	~13.7x

• Neutron Specifics: The multi-gain mode ensures that pixels with low charge (due to the $^7\text{Li}$ recoil or edge effects) are read with high sensitivity, preventing the loss of spatial information that occurs in single-gain modes.
• Pixel Size Dependence: For smaller pixel sizes ($<30 \, \mu\text{m}$), charge sharing ensures most pixels fall within the high-gain dynamic range, further optimizing resolution.

5. Significance

• Expanded Application Scope: The results confirm that CITIUS is not limited to X-ray imaging but is a versatile detector capable of high-resolution imaging for heavy charged particles and neutrons.
• Superior Performance: The combination of a thick silicon sensor (enabling substantial charge sharing) and a gain-selecting architecture allows the detector to overcome the traditional trade-off between dynamic range and spatial resolution.
• Future Impact: These findings pave the way for demonstration experiments in neutron and alpha-particle imaging, potentially revolutionizing applications in materials science, nuclear physics, and neutron radiography where sub-micron resolution is required. The authors note that demonstration experiments are currently in preparation.