Wavelength-Multiplexed Diffractive Optical Storage Enables Massively Parallel Image Retrieval

The explosive growth of data generated by artificial intelligence, cloud computing, and modern digital infrastructures is placing increasing pressure on existing information storage technologies. Although magnetic storage systems such as hard disk drives remain the dominant platform for digital storage, they face challenges including rising costs, limited lifespan, and relatively slow information retrieval. To address these challenges, researchers at the University of California, Los Angeles (UCLA) have developed a new optical information storage platform that uses engineered diffractive structures to store and rapidly retrieve thousands of images.

The UCLA team introduced a wavelength-multiplexed diffractive optical storage system composed of multiple passive dielectric layers that are spatially engineered using deep learning. Instead of storing information electronically, the system stores thousands of distinct images in a compact stack of diffractive layers, densely packed within a transparent material. Each image is assigned a unique illumination wavelength, enabling information to be selectively retrieved simply by changing the wavelength of the incident light.

Similar to tuning a radio to different stations, illuminating the diffractive structure with different wavelengths reveals different stored images, all projected onto the same output field of view. Wavelength-multiplexed diffractive optical storage offers a robust, high-density solution that is particularly vital for the long-term, stable archiving of massive datasets. By eliminating mechanical components and the degradation risks associated with traditional magnetic media, this technology ensures the reliable preservation and rapid access to information over extended periods.

Through numerical simulations in the visible spectrum, UCLA researchers demonstrated that their diffractive storage platform could store and reconstruct more than 4,000 independent image patterns while maintaining high image fidelity and minimal interference between wavelength channels. The reconstructed images exhibited an average peak signal-to-noise ratio exceeding 48 dB, highlighting the high quality of the retrieved information.

To experimentally validate the concept, the team fabricated a proof-of-concept two-layer diffractive device and demonstrated the storage and retrieval of six different image patterns. By sequentially illuminating the device with six wavelengths ranging from 500 nm to 740 nm, the researchers successfully reconstructed six distinct images within the same output field of view, confirming the feasibility of wavelength-encoded optical storage in a physical system.

Importantly, the proposed architecture is highly scalable and can operate across different regions of the electromagnetic spectrum without relying on complex material dispersion engineering. Furthermore, the wavelength-multiplexing strategy can be combined with other optical multiplexing approaches, such as polarization, illumination angle, or spatial shifting, to increase storage capacity further.

“Our work establishes wavelength-multiplexed diffractive optics as a scalable platform for all-optical information storage,” said Professor Aydogan Ozcan of UCLA, who is the corresponding author of the work. “By encoding information across wavelength channels, a single passive optical structure can support thousands of independent images, opening new opportunities for long-term storage of information, information processing, and display technologies.”

The demonstrated combination of large storage capacity, compact form factor, and rapid wavelength-encoded readout could benefit a wide range of applications, including optical data storage, information security, advanced display systems, and future photonic computing hardware. As demand for efficient, long-term data storage continues to grow, diffractive optical storage technologies may offer an attractive alternative to conventional electronic memory systems.

The study was supervised by Professors Aydogan Ozcan and Mona Jarrahi of UCLA. The other authors of this work include Che-Yung Shen, Yuhang Li, Cagatay Isil, Jingxi Li, Leon Lenk, Tianyi Gan, Guangdong Ma, and Fazil Onuralp Ardic of UCLA.