Geometric Correction Method for Large-Area Pixel-Detectors for Improved Resolution in X-ray Imaging

Abstract

Large-area hybrid pixel detectors are widely used in synchrotron facilities in small-angle X-ray scattering and crystallography beamlines but also for imaging techniques, due to their high dynamic range and single-photon sensitivity. However, their modular construction introduces gaps that limit the effective use of detector pixels and degrade data quality and reconstruction fidelity for imaging techniques. Here we present a method for geometric calibration and correction framework for the PIMEGA 540D detector installed at the Cateretê beamline of Sirius. This detector consists of modular constructions with large active area (∼ 170 × 170 mm2), with 300 μm thick silicon sensors, and pixel sizes of 55 × 55 μm2. It has a tiled architecture with reduced gaps minimising data losses; however it introduces geometric complexities requiring accurate calibration. Existing calibration approaches typically assume planar detector geometry and rely on continuous transformations, which may introduce artifacts in the scattering signal. The presented method is based on the alignment of detector substructures through discrete rotations and integer in-plane translations within a common coordinate system, while explicitly accounting for gaps between elements, calibrated using the scattering of a grid pattern, without the need of a priori knowledge of the detector mounting. By applying these corrections, the full active detector area can be exploited in diffraction and scattering experiments. We demonstrate the impact of this approach using ptychographic measurements of a Siemens star test pattern and a nano porous glass sample, both achieving spatial resolution of 8 nm in two and three-dimensions, respectively. The results highlight the importance of accurate detector geometry for high-resolution imaging and provide a practical method for large-area detectors geometrical corrections, of broad interest to the photon science community.