ETD

• gas detectors
• hybrid detectors

This thesis is focused on the design, performance study and characterization of gaseous and hybrid X-ray detectors based on a custom, pixelated CMOS chip for applications in high-resolution astronomical polarimetry, spectroscopy and imaging. Specifically, this work combines experimental measurements on prototypes with the development of Monte Carlo simulations and dedicated event reconstruction algorithms.
The development of these new detectors is motivated by the experience acquired over the last two decades with Gas Pixel Detectors (GPDs), which enabled the launch of the Imaging X-ray Polarimetry Explorer (IXPE) mission in 2021.
A fundamental component of the GPD is its custom ASIC, named XPOL, which performs an event-driven and self-triggering readout by dynamically defining a small region of interest around the event, significantly reducing the readout time compared to a full-frame acquisition. Recently, a new generation of XPOL has been developed, achieving a readout speed about an order of magnitude higher than the original version. This performance enhancement paves the way for the next generation of GPDs, which is currently being developed to support the significantly higher event rates anticipated for future X-ray polarimetry observatories.
Motivated by these perspectives, this thesis builds upon XPOL ASICs to address two parallel goals: developing the next generation of GPDs, and extending this readout architecture to application fields different from polarimetry.
Realizing the next generation of GPDs requires solving some issues of the first generation that can be attributed to the Gas Electron Multiplier (GEM) used as the multiplication stage. Indeed, the GEM introduces systematic effects that induce variations in the detector azimuthal response (spurious modulation), leading to the detection of a false polarization even when observing unpolarized sources. Furthermore, the GEM shows rate-dependent variations in the charge multiplication factor over time caused by charge accumulation on the exposed dielectric, which leads to a degradation of the energy resolution.
A promising solution to reduce these systematic effects and enhance stability consists of replacing the GEM with micro-gap-like structures fabricated directly on top of the ASIC during the foundry manufacturing process. This innovative monolithic design defines the $\mu$GPD concept. Fabricating the multiplication stage with CMOS processes guarantees sub-micrometer tolerances, drastically improving spatial uniformity and potentially reducing spurious modulation. Moreover, minimizing the exposed dielectric surface in the amplification region limits charge accumulation during avalanches, mitigating electric field distortions and ensuring gain stability over time. To achieve this goal, it is first necessary to demonstrate the feasibility of the concept by testing and characterizing the multiplication structures under realistic operating conditions before commissioning the final detector. To this end, prototype structures were fabricated via post-processing directly onto XPOL ASICs. In this context, my work is focused on the characterization of these test structures, evaluating their breakdown limits, gain capabilities, time stability, and energy resolution.
In parallel with the gas detector developments, the fast readout capabilities and low-noise electronics of XPOL make it well suited for developing a new class of X-ray detectors for non-polarimetric applications requiring energy-resolved imaging, such as X-ray diffraction, material science, and X-ray astrophysics. This is the core of the Analog Spectral Imager for X-rays (ASIX) project, which consists of the development of hybrid pixel detectors in which a solid-state sensor is coupled to an XPOL-family ASIC, inheriting GPDs' event-driven and self-triggering readout. The event-driven approach enables taking advantage of charge sharing, typical of detectors with small pixel pitches such as the 50 $\mu$m hexagonal pixels of XPOL, to achieve sub-pixel spatial resolution through offline analysis. At the same time, optimizing sensor design parameters that influence the charge sharing geometry allows for high energy resolution by maximizing the signal-to-noise ratio, which depends on the number of pixels involved in the event. My work in the ASIX project is focused on the development of the Monte Carlo simulation and analysis software, the calibration of the detector and the development of different algorithms for offline reconstruction of incidence position and energy of the photons. These algorithms are evaluated using both dedicated simulations and experimental data collected with a prototype. Through these analyses, the spatial and energy resolution of the detector are determined. Finally, a comparison between simulations and laboratory measurements is carried out to verify the validity of the Monte Carlo simulation software.
In conclusion, the measurements on $\mu$GPD prototypes presented in this work confirm the feasibility of the concept, validating the multiplication mechanism and the chip's capability to withstand the high voltages required by the integrated amplification stage. For the ASIX project, the analysis proves that the project goals in terms of energy and spatial resolution are achievable, and that advanced reconstruction algorithms can further push the limits of its performance.