Customizing a vectorial point spread function for high-dimensional imaging via metalens

Abstract

Metalenses have recently been utilized to unlock new paradigms in miniaturized, integrated imagers for multidimensional imaging by leveraging advanced multiplexing of point spread functions (PSFs). However, further progress remains largely confined to a scalar PSF framework, where spatial and polarization multiplexing faces inherent trade-offs, including limited compactness, severe aberrations, and inter-channel crosstalk. Here, we propose and experimentally demonstrate a metalens-based vectorial PSF architecture that enables high-dimensional imaging through a tailored polarization response along the optical axis. The transition from scalar to vectorial PSF facilitates the encoding of depth, wavelength, and polarization information within two complementary geometric characteristics of the PSF, enabling decoupled retrieval of different combinations of high-dimensional light field information from a single-shot image. As a concrete demonstration of this capability, we showcase two distinct four-dimensional (4D) imaging modalities: one captures 2D intensity, polarization, and wavelength, while the other resolves 2D intensity, depth, and polarization, validating high compatibility without any modification to the optical configuration. Furthermore, we demonstrate its application in characterizing and localizing anisotropic samples. This vectorial PSF framework may open an avenue for advanced PSF engineering in applications ranging from microscopy to machine vision.