Phys.org May 19, 2023
Despite recent progress in optical imaging and metrology there remains a substantial resolution gap between atomic-scale transmission electron microscopy and optical techniques. An international team of researchers (UK, Singapore) demonstrated atomic scale metrology by collecting single-shot images of the diffraction pattern of topologically structured light scattered on a suspended nanowire to determine its position relative to the fixed edges of the sample. They trained a deep learning algorithm that could predict the positions of a given nanowire based on the scattered light pattern recorded by the team’s sensor. If a sub-wavelength object moves in such a field, the scattering pattern of light on the object is very sensitive to the shape and position of the object. They employed artificial intelligence, a deep learning analysis of the scattered light intensity profile to reconstruct the object’s position. Their proof-of-principle experiments performed remarkably well. According to the researchers their work opens the field of picophotonics, the science of light-matter interactions on the picometer scale… read more. TECHNICAL ARTICLE