Online program Overview Session: JS-01

Abstract

Pixel detectors for X-ray science typically require pixel pitches of less than 1 mm and operation at photon fluxes greater than 10⁹ ph s⁻¹ mm². For lead halide perovskite detectors to be viable in these applications, they must demonstrate good spatial uniformity, as well as operational stability over a period of hours or days. Here, we present results of synchrotron experiments in which the detection performance of melt-grown CsPbBr₃ devices, solution-grown FAPbBr₃ devices and CZT were each measured at 20 keV X-ray fluxes up to 10⁹ ph/s/mm², demonstrating the linear response of the material to changing X-ray fluxes. Microbeam maps of device response to single-photon fluxes of 50 keV X-rays were used to quantify spatial variations in the charge collection efficiency and rise time of events.

Abstract

Scintillator-based photon counting detectors (PCDs) couple a fast scintillator with a silicon photomultiplier (SiPM) and associated readout electronics. The characteristics of the scintillator and SiPM that are required for competitive performance are not yet clear, and it is sometimes felt that bright scintillators and high photodetection efficiency are needed. We use numerical modeling to study this question, using a cascaded system model with multiple stages to quantify the effects of light yield, SiPM sensitivity, and energy binning. Effects such as SiPM saturation, Swank noise, and K-escape are also included as optional stages. We find that using relatively dim components (5x dimmer than that assumed in previous modeling studies) introduces little harm into spectral imaging tasks, but the penalty of K-escape is large. Future prototype detectors should emphasize K-escape, stopping power, and speed rather than light yield or photodetection efficiency.

Abstract

We present the second generation of a four-side buttable module solution (AS592xM) for spectral CT detectors with a pixel matrix of 72x24 covering an area of 24.5 mm x 8.57 mm with a pitch of 365 µm x 340 µm. In combination with a suitable direct detection material like CdTe or CZT it detects individual X-ray photons with five programmable thresholds. Improvements have been made to reduce power consumption down to 0.35mW/pixel while maintaining low noise performance (380 e rms). Count rate stability is based on two different methods: a static leakage compensation (SLC) as well as a base line restorer circuit (BLR). The new SLC in combination with an auto-zero scheme in the frontend suppresses the effect of potential 1/f and Random Telegraph Noise (RTN) noise from the amplifiers. The adjustable BLR compensates for static and dynamic baseline shifts up to 700nA of leakage current per pixel. Measurements with and without CZT will be shown to quantify the energy resolution and count rate stability over input flux for different operation modes to quantify S/N for the different energy bins. Additionally, the MTF is characterized to project spatial resolution on system level.

Abstract

In this presentation, we will report the design and imaging performance of our third-generation Alpha-SPECT-mini system (the Alpha-SPECT-mini-ZM, where ZM stands for Zoom-in Microscope), which combines ultra-narrow micro-ring collimators and a vibrational aperture-scanning scheme to provide ultrahigh spatial resolution for in vivo 3-D autoradiography of α-RPT in small lab animals. The system uses a cylindrical ring of 10 detector panels, each comprising 2×6 HEXITEC/CZT detector modules. Each CZT module has a 2×2 cm² active area with 80×80 pixel array at 250 μm pitch and offers 1.5 keV FWHM energy resolution at 122 keV. The Alpha-SPECT-mini-ZM system uses a two-section collimator ring that combines pinhole apertures for whole-body imaging and ultra-narrow micro-ring apertures for zoomed-in ultrahigh-resolution imaging of a focal area with a 5-mm-diameter FOV, switched via a motorized stage. To offer an ultrahigh imaging resolution that overcomes the limited detector pixel size, the system employs a vibrational scanning strategy with MUCRB-guided acquisition optimization. Preliminary simulations with micro-hot-rod phantoms show the micro-ring collimator achieving better than 160 μm resolution at practical sensitivity. Further phantom studies will be carried out to characterize the performance of the system.

Abstract

Although large language models (LLMs) perform well on general-purpose programming tasks, their accuracy drops sharply when applied to domain-specific simulation codes such as Monte Carlo N-Particle (MCNP). Within the nuclear sector, security policies and export-control rules typically forbid uploading inputs to cloud-hosted LLM services, so review tools must instead run on-premise atop openly released models. This study introduces a coordinated multi-agent framework that automatically inspects and repairs MCNP input decks. Three role-specific agents—a Supervisor, a Fixer, and a Validator—drive the workflow, supported by 16 auxiliary tools and an optional retrieval-augmented generation (RAG) module that consults the MCNP user manual. To measure performance, we assembled a 360-case benchmark covering six representative nuclear-engineering scenarios. Our pipeline reaches a 75.8% success rate, exceeding a single-agent reference by 15.5 percentage points and confirming that role-based orchestration is an effective strategy for trustworthy MCNP code repair in fully local deployments.

Abstract

Particle physics experiments rely on complex detectors, where trigger and data acquisition (TDAQ) systems play a key role. These systems are typically expensive, large-scale, and accessible only in specialized laboratories, limiting students’ early exposure to the fundamentals of TDAQ. To address this gap, we developed a compact, low-cost educational platform that reproduces the main elements of a particle-detector readout chain in a form suitable for university laboratories. The platform is based on an Arduino UNO R4 Wi-Fi, a custom piggy-back board, and a detector module consisting of a CsI(Tl) scintillator coupled to a silicon photomultiplier. The system integrates bias-voltage generation and temperature monitoring, programmable pulse amplitude discrimination, peak detection, digitization, and USB or Wi-Fi data transfer to a personal computer. 
The platform enables students to explore the main TDAQ concepts, including trigger formation, latency, dead time, buffering, and threshold setting. By progressively implementing different firmware strategies, from simple trigger polling to interrupt-driven and hardware-assisted acquisition, students gain direct insight into the relation between system design and detector performance. Educational activities on slow control, such as bias-voltage calibration, detector monitoring, and gain stabilization, are also possible. After firmware development, students can perform gamma-spectroscopy measurements using low-activity radioactive materials. Examples include potassium chloride salt substitutes, where the 1460.8 keV photopeak from ⁴⁰K can be observed, and thoriated tungsten welding electrodes, which exhibit several gamma lines from thorium decay products. In this contribution, we discuss the platform, the TDAQ topics addressed, and student experiments that demonstrate its educational potential.