Optical analog computing of spatial differentiation and edge detection with dielectric metasurfaces

Area of Science:

• Photonics and Optical Computing
• Nanotechnology and Metamaterials

Background:

• Optical analog computing offers a pathway to overcome the limitations of digital processing for specific tasks.
• Metasurfaces, engineered nanostructured surfaces, provide a versatile platform for manipulating light.
• Spatial differentiation and edge detection are fundamental operations in image processing and computational tasks.

Purpose of the Study:

• To propose and demonstrate a novel method for optical analog computing of spatial differentiation and edge detection.
• To achieve these computations using a single layer of dielectric metasurface.
• To explore the capabilities of silicon nanodisks for optical transfer function engineering.

Main Methods:

• Engineering the spatial dispersion of electric dipole resonance in silicon nanodisks within a metasurface.
• Designing the metasurface to achieve the optical transfer function for second-order derivation.
• Numerical validation using one-dimensional spatial functions and image processing.

Main Results:

• Successful realization of optical analog computing for spatial differentiation and edge detection.
• Demonstration of two-dimensional differentiation and arbitrary polarization capabilities.
• Achieved high efficiency and large spatial bandwidth at visible wavelengths.

Conclusions:

• A single-layer dielectric metasurface can effectively perform optical analog computing for spatial differentiation and edge detection.
• The electric dipole resonance mechanism in silicon nanodisks is key to achieving these optical functions.
• This work facilitates advancements in optical computing using artificial nanostructures.