A research team from the University of California, Los Angeles (UCLA) has developed a novel refractive function generator (RFG) capable of precisely controlling light refraction through artificially designed material structures. This study, published in Nature Communications, breaks the limitations of traditional Snell's law, bringing new possibilities to optical technologies.

Led by Professor Aydogan Ozcan from UCLA's Department of Electrical and Computer Engineering, the team used deep learning algorithms to design special three-dimensional transmissive layer structures. Professor Ozcan stated: "Our RFG devices can arbitrarily program the refraction behavior of light, opening new pathways for optical system design." The researchers validated this technology in the terahertz band using 3D-printed materials, successfully achieving various complex light wave transformations, including negative refraction.

The key innovation in this programmable optical material lies in its passive structural design. Unlike traditional metamaterials, RFG requires no external energy input and achieves precise light path control solely through carefully designed microstructures. The research team demonstrated that devices with thicknesses of only tens of wavelengths can effectively resist manufacturing defects and wavelength variations.

This technology holds potential applications in optical computing, communications, and imaging. Co-author Professor Mona Jarrahi noted: "This AI-based design method can be further extended to achieve more complex functions such as wavelength and polarization multiplexing." The research team plans to continue optimizing material performance and exploring more practical application scenarios.