麻豆淫院

Refraction鈥攖he bending of light as it passes through different media鈥攈as long been constrained by physical laws that prevent independent control over how light waves along different directions bend. Now, UCLA researchers have developed a new class of passive materials that can be structurally engineered to "program" refraction, enabling arbitrary control over the bending of light waves.

In a study in Nature Communications, a team led by Dr. Aydogan Ozcan, the Chancellor's Professor of Electrical & Computer Engineering at UCLA, has introduced a novel device called a refractive function generator (RFG) that can independently tailor the output direction of refracted light for each input direction. This device allows light to be steered, filtered, or redirected according to custom-designed rules鈥攆ar beyond what standard materials or traditional metasurfaces can achieve.

Standard refraction, described by Snell's law, links the input and output directions of light using fixed material properties. Even advanced metasurface designs only allow limited tunability of refraction.

The RFG, however, uses a very thin stack of passive transmissive layers鈥攅ach structurally engineered through deep learning at a scale close to the diffraction limit of light鈥攖o define completely arbitrary refractive functions, effectively decoupling the input-output mappings of light refraction. The UCLA team demonstrated that these thin optical devices, spanning only a few tens of wavelengths in thickness, can perform sophisticated wave transformations such as permutation, filtered permutation and negative refraction.

To validate their approach, the researchers fabricated and experimentally tested RFGs using 3D printed materials and terahertz waves. These devices successfully bent light in precisely defined directions, successfully demonstrating arbitrary programming of refractive functions.

"This is a significant step forward in our ability to precisely control and engineer how light behaves," said Dr. Ozcan. "By programming the refraction of light using structured 3D materials, we open up new design opportunities for optical computing, communications, and imaging systems."

The study shows that these RFG devices can be designed using AI to be compact, efficient, and robust against fabrication imperfections and wavelength variations. The AI-based design framework also showcased further extensions, including wavelength and polarization multiplexing of RFGs, and unidirectional light routing using only passive, structured materials.

The authors of this work are Dr. Md Sadman Sakib Rahman, Tianyi Gan, Prof. Mona Jarrahi, and Prof. Aydogan Ozcan, all at the UCLA Samueli School of Engineering. This research was supported by the US ARO (Army Research Office).