A metalens is an advanced optical element that uses a metasurface—an array of nanoscale resonators (NRs)—to manipulate light at the subwavelength scale. In contrast to conventional lenses, which focus light through curved glass or plastic surfaces, metalenses control optical wavefronts by precisely engineering the geometry, dimensions, and spatial arrangement of NRs. This approach enables effective aberration correction and high focusing efficiency and resolution, while offering a significantly thinner and lighter form factor compared to traditional lenses.

The widely used approach to design a metalens starts from the generation of a Look-Up Table (LUT), i.e., a pre-calculated library containing dissimilar NRs that can impart a phase in the 0-2π range by simply tuning a few geometrical parameters. Once the LUT is generated, the NRs are carefully placed in the metalens to assign, at every spatial point, the required phase to achieve the desired performance. However, to compute the LUT, the NRs are usually placed in a perfectly periodic array, an assumption that fails when the NRs are then displaced into the real metalens. Thus, the LUT method implies an intrinsic error limiting its performance. Some recent studies have raised concerns regarding the reported performance of metalenses in the existing literature. For this reason, the modelling of a full metalens to evaluate its actual focusing efficiency, and methods beyond the standard LUT approach for the metalens design, are necessary.

In this work, we consider a 2D cylindrical metalens focusing light in the NIR range. After calculating the LUT and engineering the metalens, we first evaluate its actual performance and then we envision how to improve them.