A New Concept of Liquid Xenon Time Projection Chamber for Medical Imaging

Liquid xenon time projection chambers offer a homogeneous detection medium with excellent intrinsic energy resolution, fast scintillation, and true three-dimensional position sensitivity, making them an attractive alternative to crystal-based detectors for positron emission tomography (PET). In this work, we present a new single-phase liquid xenon time projection chamber (TPC) concept optimized for medical imaging, employing combined scintillation and electroluminescence-based ionization readout to enable low-noise signal amplification and intrinsic depth-of-interaction measurement. We evaluate the system-level performance of this detector concept using Monte Carlo simulations based on OpenGATE and Geant4, with direct comparison to conventional LYSO-based PET systems. The study focuses on detection sensitivity, energy-based event selection efficiency, and reconstructed spatial resolution. While LYSO detectors provide higher absolute stopping efficiency due to their higher density, liquid xenon detectors exhibit improved photopeak purity as a result of superior intrinsic energy resolution, leading to enhanced rejection of scattered events. Point-source reconstruction studies demonstrate that the intrinsic three-dimensional position sensitivity of the liquid xenon TPC translates into a reconstructed spatial resolution of approximately 1mm full width at half maximum (FWHM) at the system level, compared to approximately 4mm for LYSO-based systems under comparable conditions. These results indicate that liquid-xenon-based PET detectors can achieve competitive or superior imaging performance, particularly for applications requiring high spatial resolution, large axial acceptance, and scalable detector geometries.

💡 Research Summary

This paper presents a comprehensive study on a novel detector concept for Positron Emission Tomography (PET) based on a single-phase liquid xenon time projection chamber (TPC). It addresses the limitations of conventional segmented inorganic scintillator crystals (like LYSO), such as high cost, fixed geometries, and lack of intrinsic depth-of-interaction (DOI) information, by proposing a homogeneous liquid xenon detection medium.

The core innovation lies in the detector’s operational principle. It utilizes both the prompt scintillation light and ionization electrons produced by 511 keV gamma-ray interactions in liquid xenon. The scintillation provides fast timing for coincidence detection. The ionization electrons are drifted under an electric field to a designated region where they undergo electroluminescence (EL), a process that amplifies the signal with low noise while preserving the excellent intrinsic energy resolution of liquid xenon. The time difference between the scintillation and EL signals provides precise depth information, and the spatial pattern of the EL light gives transverse coordinates, enabling true, continuous 3D position reconstruction without discrete crystal segmentation.

The system-level performance was rigorously evaluated using Monte Carlo simulations (OpenGATE/Geant4) and directly compared to a standard LYSO-based PET system across a range of detector thicknesses. The analysis yielded two key, nuanced findings regarding efficiency and resolution.

Firstly, on detection efficiency: due to its higher density, LYSO demonstrated a higher absolute probability of interacting with and stopping 511 keV photons (higher stopping power). However, liquid xenon, benefiting from its superior intrinsic energy resolution (2.1% FWHM simulated vs. 11.2% for LYSO), exhibited a higher “photopeak purity.” This means a larger fraction of true, full-energy absorption events were correctly identified within the selected energy window, leading to more effective rejection of scattered events. Thus, while the raw interaction count may be lower, the quality of the retained coincidence data is higher for liquid xenon.

Secondly, and most strikingly, on spatial resolution: The intrinsic 3D position sensitivity of the liquid xenon TPC translated into a vastly superior reconstructed image resolution at the system level. Assuming position resolutions of 1 mm in depth and 2 mm transversely for liquid xenon (based on electron diffusion and prior studies), the simulated point-source reconstruction achieved approximately 1 mm Full Width at Half Maximum (FWHM). In contrast, the LYSO-based system, limited by its 4x4x19 mm³ crystal size, yielded about 4 mm FWHM. This nearly fourfold improvement highlights the fundamental advantage of continuous position sensing over discrete segmentation.

The paper concludes that liquid-xenon-based PET detectors offer a competitive and potentially superior alternative, particularly for applications demanding very high spatial resolution, large axial coverage, and scalable detector geometries. It acknowledges that the lower stopping power compared to LYSO is a trade-off but suggests it can be mitigated by the improved scatter rejection and future integration of Time-of-Flight (TOF) information, for which liquid xenon’s fast scintillation is also highly suitable. Future work is directed towards experimental validation of module and system performance, optimization of timing, and studies on practical implementation in medical imaging environments.