Encoding independent wavefronts in a single metasurface for high-order optical vortex recognition.

The orbital angular momentum (OAM) of vortex beams has great potential in optical communications due to its communication confidentiality and low crosstalk. It is necessary to design a plausible OAM pattern recognition mechanism. Abandoning AI models that require large datasets, a single passive all-dielectric metasurface consisting of TiO nanopillars on a SiO substrate is used to recognize high-order optical vortexes. In this configuration, the proposed device is capable of simultaneously encoding the wavefront and the transmission paths in different incident OAM beams. Due to the presence of spin angular momentum (SAM), the vortex beam to be identified is spatially separated after passing through the metasurface. As a proof of concept, 14 signal channels are considered in the constructed metasurface, 12 of them can be encoded at will for the detection of any vortex beam with a predefined topological charge. These results make use of metasurfaces to enable OAM pattern recognition in an effective way, which may open avenues for the ultimate miniaturization of optical vortex communication and advanced OAM detection technologies.