- ETH Zurich detector replaces millions of components with single scintillator block
- SPAD sensors and AI achieve 200 micrometer resolution in simulations
- Technology could sharpen PET medical scans and cut detector costs
- Prototype successfully tracked electrons from strontium-90 source
Scientists at ETH Zurich and EPFL have developed PLATON, a particle detector that uses light-field camera technology to track invisible particles in 3D. PhD student Till Dieminger, senior scientist Dr. Saúl Alonso-Monsalve, Professor Davide Sgalaberna and colleagues, together with members of the Advanced Quantum Architecture Lab at EPFL in Lausanne led by Professor Edoardo Charbon, developed and tested the first prototype, publishing results in Nature Communications in April 2026.
Instead of dividing the detector into millions of tiny units, the system uses advanced camera technology to reconstruct where the light originated. Simulations demonstrated that the detector could track these tiny particles down to a resolution of 200 micrometers.
Single scintillator block eliminates assembly bottleneck
Manufacturing such geometries with current production strategies is challenging, as they involve time-consuming and costly fabrication processes, followed by the assembly of millions of individual parts. Such systems introduce significant challenges in construction and demand a large number of readout electronics channels, leading to extremely high costs.
PLATON sidesteps this entirely. Instead of using millions of tiny components, it employs a single scintillator block with a micro-lens array and SPAD sensors. When plenoptic cameras are paired with single-photon avalanche diode (SPAD) array sensors, they can detect individual photons and potentially reconstruct particle tracks even when very little light is available.
The economics matter more than the researchers may realize. In segmented detectors, finer resolution means more readout channels, and reducing the pitch increases the number of readout channels and consequently the cost and system complexity. A monolithic approach that uses computational reconstruction instead of physical segmentation breaks this scaling problem—resolution improvements come from better optics and algorithms, not exponentially more hardware.
Strontium-90 tests validate single-photon tracking
In the lab tests, the team successfully reconstructed positions of electrons from a strontium-based source, confirming that, as suspected, this setup could viably detect particles. We built and successfully characterised the spatial resolution of the PLATON detector prototype, a plenoptic camera instrumented with a SPAD array sensor, and we developed a post-processing method suitable for photon-starved images. We also succeeded in reconstructing the position of 90Sr electrons on an event-by-event basis.
Enabling technologies are plenoptic systems and time-resolving single-photon avalanche diode array imaging sensors. Together, they enabled us, using a plenoptic camera, to reconstruct the origin of single photons in the scintillator. The micro-lens array sits between the main lens and the imaging sensor, turning each lens element into a miniature camera that captures both intensity and directional information.
PET scanners could gain resolution without detector overhaul
The technology may also lead to sharper PET medical scans. Technological improvements in detector sensitivity, image resolution, scan speed, and artificial intelligence (AI)-enabled image reconstruction are transforming PET imaging capabilities. The market is following: the global PET/CT scanner device market generated $1.87 billion in 2020 and is expected to reach $3.20 billion by 2030.
Current PET detector upgrades focus on faster scintillators and improved electronics. PLATON offers a different path—keeping the scintillator monolithic but extracting spatial information optically. These improved detector electronics will combine with computational methods including deep learning and AI to better estimate the location, time, and energy of an interaction in the detector. If the approach scales to clinical PET ring diameters, it could deliver submillimeter resolution without the manufacturing complexity of current high-resolution designs.
PLATON shifts detector resolution gains from mechanical segmentation to optical reconstruction. For neutrino experiments scaling to kiloton volumes, this matters—building and instrumenting millions of individual detector cells is a fabrication and assembly nightmare. For PET manufacturers, it offers a route to higher resolution without fragmenting the detector ring into ever-smaller crystals. The real test comes when someone attempts to scale this beyond the prototype. A case study focused on neutrino detection demonstrates full event reconstruction with a spatial resolution of two hundred micrometres, but production detectors will need to prove they can maintain that performance across cubic meters of scintillator while surviving years of operation.
How does PLATON achieve submillimeter resolution without physical segmentation?
PLATON uses a micro-lens array positioned between the main lens and a SPAD imaging sensor. Each micro-lens captures both the intensity and direction of incoming photons from the scintillator. AI algorithms then ray-trace these individual photons back to their origin point in 3D space. This optical reconstruction replaces the need to physically divide the scintillator into millions of separate detector cells, each requiring its own readout channel.
What prevents current segmented detectors from scaling to finer resolution?
Cost and complexity scale exponentially with segmentation. Finer resolution requires smaller detector segments, which multiplies the number of readout electronics channels. Each channel needs amplification, digitization, cabling, and data acquisition hardware. In large neutrino detectors, this can mean millions of individual components that must be manufactured, assembled, aligned, and maintained. Manufacturing tolerances become tighter, assembly time increases, and failure modes multiply as segment count grows.
Article Source: Scientists built a camera that can track invisible particles in 3D
