Abstract

The principles of information multiplexing in ptychography have been sufficiently demonstrated with few channels [1]. In contrast to a single source approach—where the sample is scanned in a set of overlapping positions with respect to a confined illuminating beam and a detector downstream the object records the intensity of the diffracted field—a ptychography experiment can employ a set of incoherent sources that illuminates the same region of the object and pixelated detector records the incoherent sum of the diffracted fields. A multi-wavelength [2] or mixed-states [3] forward model can be used to deconvolve the object and probe. This approach has been used successfully in VIS [4], XUV [5], and X-Rays [6] not only to obtain the object spectral response, but also to scan simultaneously independent regions of the sample. In the latter, a performance increase in terms of the scanned area per time can be achieved. The question of how many channels can be effectively coded and decoded and what are the sampling requirements is still open. We conducted a numerical analysis with synthetic data to explore the constraints and performance increase when using 1, 4, and 16 incoherent beams with unique scanned regions. The simulations were done at a wavelength of 13.5 nm (92 eV) and each beam was modulated by an ampilutde mask with a 8 μm diameter placed 100 μm upstream the sample. A single scanned area covered a diameter of 24 μm, and the detector was placed in the far-field 10 cm downstream the object. The average linear overlap for each probe was set to four different scenarios [80%, 85%, 90%, 94%], that corresponded to an increasing number of scanning points [200, 400, 800, 2000].

© 2023 IEEE