Imaging performance evaluation of large photon sieves with high-efficiency simulation algorithms

Photon sieve (PS) is a kind of diffractive lens that has great application potential in ultraviolet and X-ray imaging. It is very difficult to simulate the imaging performance of large PSs with numerous pinholes because of the high computing cost. No study has systematically analyzed a large PS through simulation. Based on individual far-field models with slight modifications, we introduce a symmetry assumption and a dimension-expansion algorithm with GPU. This approach makes the simulation approximately 1000 times faster and can be widely used to simulate point spread functions (PSFs) under both ideal and non-ideal conditions, considering factors such as chromatic aberration, depth of focus (DOF), wrinkles, and manufacturing uncertainty. For a smaller field of view (FOV), we derive Fresnel Zone Plate (FZP) diffraction formulas to quickly compute the FZP’s diffracted field and find a strong linear relationship between the PSF of a PS and that of its homologous FZP. By utilizing the PSF of the FZP and the linear factor, we achieve simulations in ideal cases within a few seconds. We systematically evaluate the imaging performance of an Hα PS with 1.7×10⁸ pinholes and predict nonideal impacts on the PS. We believe this is the first systematic simulation-based analysis of the comprehensive performance including manufacturing and installation uncertainties for such a large PS. These simulations offer effective methods for analyzing the systematic performance of large PSs and can be widely applied in their design.