Flexible conformal polarization-sensitive structural color optical switch

Area of Science:

• Photonics and Materials Science
• Artificial Intelligence in Optics

Background:

• Traditional metasurface structural color design is computationally intensive.
• Existing optical switching mechanisms lack efficiency and speed.

Purpose of the Study:

• To develop a bidirectional deep neural network (Bi-DNN) platform for the inverse design of flexible, polarization-sensitive color switches.
• To significantly accelerate the design process compared to conventional simulations.

Main Methods:

• Implemented a bidirectional deep neural network (Bi-DNN) for simultaneous prediction of structural parameters and output colors.
• Utilized mechanical strain tuning (-10% to 50%) to control the metasurface's optical properties.
• Validated performance using finite-difference time-domain (FDTD) simulations for comparison.

Main Results:

• Achieved prediction speeds 3 orders of magnitude faster than FDTD (64 ms per sample).
• Demonstrated a wide color gamut (92% sRGB) with high accuracy (ΔE < 2) under strain.
• Enabled polarization-switchable color display/hiding with a 17.1 dB extinction ratio.
• Confirmed angular stability (0°-20° incidence).

Conclusions:

• The Bi-DNN platform offers a new paradigm for high-performance optical switches.
• Advances inverse-design foundations for directional color metasurfaces.
• Presents applications in next-generation displays, smart wearables, and optical encryption.