Optical metasurfaces are subwavelength-thick arrays of meta-atoms that have attracted significant attention due to their superior capabilities compared with conventional optical devices. Designing metasurfaces for practical applications requires system-level models that accurately predict their responses. The conventional approach for modeling metasurfaces is to ignore the coupling among the meta-atoms and to model metasurfaces as phase, amplitude, or polarization masks that are independent of the incident light’s wavefront, which is an inaccurate assumption for large incident angles. In this talk, we will introduce a novel technique for the modeling and design of metasurfaces based on the discrete-space impulse response (DSIR) concept. Because the waves propagating in free space are spatially band-limited, the incident, the transmitted, and the reflected waves can be represented using discrete-space signals that are obtained by sampling these waves at the Nyquist rate (at half a wavelength intervals). As a result, we can define discrete-space impulse responses for metasurfaces that relate the transmitted/reflected waves to the incident waves. We show that such impulse responses are local, accurately model the interactions among neighboring meta-atoms, and completely characterize the metasurfaces’ response to any incident waves. We also introduce a new approach for designing metasurfaces using the DSIR concept. As a proof-of-concept, we present the characterization results of a high numerical aperture meta-lens that is designed using the DSIR technique and show that its focusing efficiency is higher than that of a similar meta-lens designed using the conventional technique.
|