Engineering Integral and Differential Dispersion With Subwavelength‐Scaled Planar Metastructures for Image Encryption

ABSTRACT Optical analog computing offers inherent advantages such as ultrahigh computational density and minimal power consumption, yet practical deployment remains limited by bulky and fabrication‐intensive architectures. Inspired by the ability of epsilon‐near‐zero (ENZ) materials to manipulate dispersion within a subwavelength scale, we propose a planar pixel metastructure strategy that emulates and tailors such compact dispersion responses within a fully planar, fabrication‐compatible architecture. Instead of employing lossy plasmonic materials or complex three‐dimensional (3D) waveguide effective ENZ media, the reconfigurable pixelated architecture allows deterministic dispersion engineering at the subwavelength scale, enabling compact and integrable pixel metastructure processing units (pixel‐MPUs). As a proof of concept, we design and experimentally verify pixel‐MPUs that realize integration and differentiation functions, and further demonstrate real‐time image encryption and decryption using cascaded operators with negligible latency and energy consumption. The proposed pixel‐MPUs combine functionality, scalability, and broad planar‐process compatibility, providing a practical foundation for highly integrated and manufacturable optical analog computing systems.