🔬 physics

CMOS Image Sensors for CEvNS Detection at Nuclear Reactors

This paper proposes and analyzes a CMOS image sensor-based detector with active shielding and optimized readout parameters, demonstrating its feasibility for detecting coherent elastic neutrino-nucleus scattering (CEvNS) from nuclear reactors within days to weeks, thereby establishing CIS technology as a promising platform for reactor monitoring and precision neutrino measurements.

Original authors: Santiago Ezequiel Perez, Dario Rodrigues, Miguel Sofo-Haro

Published 2026-08-11

📖 4 min read☕ Coffee break read

CC BY 4.0

Original authors: Santiago Ezequiel Perez, Dario Rodrigues, Miguel Sofo-Haro

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine the universe is filled with ghostly particles called neutrinos. They are the ultimate hiders: they zip through planets, stars, and even your own body without ever saying "hello" or leaving a scratch. Because they are so shy and interact so weakly with matter, catching one is like trying to hear a whisper in a hurricane. For decades, scientists had to build massive detectors deep underground to catch just a few of these elusive visitors.

However, there is a special trick nature plays called "Coherent Elastic Neutrino-Nucleus Scattering" (CEvNS). Think of it like this: usually, a neutrino bumps into a single atom and bounces off. But in this rare event, the neutrino hits the entire nucleus of an atom all at once, like a bowling ball hitting a bowling pin. Because the whole nucleus moves together, the "kick" is much stronger and easier to feel. This discovery is a big deal because it means we might be able to build tiny, compact detectors to catch these ghosts, rather than needing giant underground caverns. This could let us monitor nuclear reactors for safety or even look for new, hidden laws of physics that the Standard Model doesn't explain.

The Paper's Mission: Catching Ghosts with a Digital Camera

In this paper, the authors propose a new way to catch these neutrino ghosts using a technology you likely have in your pocket right now: a CMOS image sensor (CIS). You know these as the tiny cameras in your smartphone. The team suggests that by tweaking these sensors, they can become super-sensitive neutrino detectors.

The Problem with the Old Cameras
The researchers looked at a previous type of sensor called a "Skipper-CCD." These are amazing because they can count individual electrons, making them incredibly sensitive. However, they have a major flaw: they are slow. Reading a single image from a Skipper-CCD takes hours. Imagine trying to catch a speeding car by taking a photo, but the camera shutter takes three hours to open and close. By the time you get the picture, the car is long gone. This slowness also means they can't use "active shielding"—a safety system that acts like a bouncer to kick out background noise (like cosmic rays) the moment they arrive.

The New Solution: The High-Speed CMOS
The authors argue that CMOS sensors are the perfect upgrade. Unlike the slow, serial-readout of the old cameras, CMOS sensors read all their pixels at the same time (parallel readout). This allows them to snap pictures at lightning speed—up to 1,000 frames per second (fps).

To make this work, the team developed a clever strategy to handle the "bouncers" (background noise). Since they can't stop every single cosmic ray muon from hitting the sensor, they use the camera's speed to their advantage. They set up a system where if a muon detector sees a particle coming, the camera instantly rejects that specific frame. Because the camera is so fast, it only loses a tiny fraction of its time (less than 10%), keeping the "live time" high. This is like a security guard who can instantly spot a troublemaker and only discard the one second of video they caused, rather than shutting down the whole camera for hours.

The Simulation Results
The authors didn't build the final detector yet; instead, they ran detailed computer simulations to see how well this idea would work. They modeled a detector with about 100 grams of sensitive silicon, a readout noise of just 1 electron (meaning it's very quiet), and a speed of 200 frames per second.

Their simulations suggest that with these specs, the detector could spot neutrinos from a nuclear reactor located 16 meters away in less than 50 days. If they can improve the sensor even further—getting the noise down to 0.2 electrons and the speed up to 1,000 fps—the detection time could shrink dramatically to just 7 days.

What This Means
The paper concludes that this technology is a very promising path forward. It suggests that we don't need massive, expensive underground labs to study these particles anymore. Instead, we could use compact, fast, and relatively cheap camera chips to monitor nuclear reactors for safety or to hunt for new physics. While the results are currently based on models and simulations, the authors are confident that the technology is mature enough to make this a reality, potentially turning the next generation of neutrino experiments into something as small and portable as a high-tech camera.

Technical Summary: CMOS Image Sensors for CEvNS Detection at Nuclear Reactors

Problem Statement
Coherent Elastic Neutrino-Nucleus Scattering (CEvNS) offers a promising channel for reactor neutrino experiments, enabling compact detectors for applications such as reactor safeguarding and monitoring, as well as searches for physics beyond the Standard Model (BSM). However, detecting CEvNS is challenging due to the low nuclear recoil energies involved and the necessity of suppressing background events. While Skipper-CCD technology has demonstrated ultra-low ionization thresholds ( $\sim$ 5 eV) and single-electron sensitivity, its serial readout architecture results in long acquisition times (hours per image), precluding the use of active shielding systems based on time-coincident vetoes against cosmic muons. Conversely, standard Complementary Metal-Oxide-Semiconductor (CMOS) Image Sensors (CIS) offer high frame rates but have historically faced challenges regarding readout noise and target mass. This work addresses the need for a detector platform that integrates low ionization thresholds, high temporal resolution, and substantial silicon mass to enable active background suppression in reactor environments.

Methodology
The authors propose a CIS-based detector architecture and develop an analytical framework to evaluate its feasibility for CEvNS detection. The methodology includes:

Active Shielding Strategy: Unlike Skipper-CCDs, CIS architectures support parallel pixel readout with frame rates up to 1000 frames per second (fps). The authors propose exploiting the Poisson statistics of cosmic muon arrivals. By rejecting frames coincident with muon detector triggers, the system can achieve high live times (>90%) while suppressing muon-induced neutron backgrounds. The analysis assumes a muon rate of $\sim$ 200 Hz for a detector volume comparable to previous experiments (e.g., CONUS+).
Intrinsic Detector Background (IDB) Modeling: An analytical model is developed to quantify intrinsic backgrounds arising from dark current (thermal generation of electron-hole pairs) and readout noise. The model treats dark current as a Poisson process and readout noise as a Gaussian distribution. It extends beyond single-pixel analysis to account for "polykings"—geometrical shapes of connected pixels formed by random noise fluctuations—calculating the probability of noise events mimicking signal events (defined as groups of up to four pixels).
Signal and Significance Calculation: The CEvNS signal rate is calculated for silicon targets using reactor antineutrino fluxes and Standard Model cross-sections. Two experimental scenarios are evaluated: a high-flux configuration ( $H_1 = 0.014$ GW $_{th}$ /m $^2$ ) and a reduced-flux configuration ( $H_2 = 0.004$ GW $_{th}$ /m $^2$ ). The discovery significance is determined using the median discovery significance formula for Poisson counting experiments with known background, rather than a Gaussian approximation, to ensure accuracy at low event counts.

Key Contributions

Feasibility of High-Speed CIS for CEvNS: The paper demonstrates that modern CIS technology, including "Skipper-CMOS" architectures and high-voltage depleted monolithic active pixel sensors (DMAPS), can achieve the necessary combination of sub-electron readout noise, high frame rates, and thick silicon substrates (up to 725 $\mu$ m).
Analytical Background Model: The introduction of a probabilistic model for IDB that accounts for the spatial correlation of noise pixels (polykings) allows for a more accurate estimation of background rates at the event level, rather than just the pixel level.
Active Shielding via High Frame Rates: The work establishes that high frame rates (hundreds to thousands of fps) enable effective active shielding against cosmic muons without the dead-time penalties associated with serial readout technologies, making surface-level or low-overburden reactor experiments viable.

Results
The authors evaluate a feasible detector design with the following parameters: $\sim$ 100 g of sensitive silicon mass (256 chips of 1 Mpix each), 1 $e^-$ readout noise, 1.5 $e^-$ /pix/s dark current, and 1000 fps operation.

Background Suppression: With an energy threshold set above 0.064 keV to suppress IDB, and assuming a conservative flat physical background of 100 DRU (Differential Rate Units), the detector can effectively isolate the CEvNS signal.
Detection Time:
- For the high-flux scenario ( $H_1$ ), a 3 $\sigma$ significance is achievable in approximately 17 days.
- For the reduced-flux scenario ( $H_2$ ), the detection time extends to approximately 150 days.
Performance Improvements: The study projects that improving detector performance to 0.2 $e^-$ readout noise and 1000 fps (with optimized dark current) could reduce the detection time for the high-flux scenario to roughly 7 days. Even with a more conservative 1 $e^-$ noise and 200 fps, detection within 50 days is established as feasible.

Significance and Claims
The paper claims that CIS technology represents a promising platform for reactor monitoring and precision neutrino measurements. The authors assert that the integration of parallel readout and low-noise capabilities allows for the implementation of active shielding, a critical component for background reduction that was previously incompatible with ultra-low threshold detectors like Skipper-CCDs.

The work positions CIS sensors as a viable path toward next-generation CEvNS experiments. Specifically, it suggests that:

Reactor Monitoring: Current CIS capabilities (1 $e^-$ noise, 200 fps) are sufficient for reactor monitoring applications and detector characterization.
BSM Physics: Further advancements toward sub-electron noise levels would significantly enhance the signal-to-background ratio, enabling sensitive searches for non-standard neutrino interactions, light mediators, and millicharged particles.
Scalability: The use of commercial CMOS foundries and hybrid integration techniques (e.g., SPROCKET readout circuits) offers a scalable and cost-effective route to producing large-mass, high-performance neutrino detectors.

The authors conclude that thick CMOS sensors featuring parallel and non-destructive readout are particularly attractive candidates for future CEvNS experiments, potentially reducing the required exposure time from months to weeks or days.

More like this