Electron-trapping materials, due to their exceptional ability of energy storage and controllable photon release under external stimulation, have attracted considerable attention in the field of optical information storage (OIS). In this work, Gd₃Al₃Ga₂O₁₂:Ce³⁺, Yb³⁺ fluorescent ceramics, were developed using air and vacuum sintering technology. By co-doping Ce³⁺ and Yb³⁺, the trap density was significantly increased by 7.5 times compared to samples containing only Ce³⁺. Vacuum annealing further enhanced trap density by 1.6 times compared to samples sintered solely in air, while generating deep traps (1.44 eV), making Gd₃Al₃Ga₂O₁₂:Ce³⁺, Yb³⁺ an excellent OIS medium. This work is expected to facilitate the development of OIS materials.

Electron-trapping materials have attracted a lot of attention in the field of optical data storage. However, the lack of suitable trap levels has hindered its development and application in the field of optical data storage. Herein, Lu₃Al₅O₁₂:Ce³⁺ fluorescent ceramics were developed as the optical storage medium, and high-temperature vacuum sintering induced the formation of deep traps (1.36 eV). The matrix based on the garnet-structured material ensures excellent rewritability. By analyzing the thermoluminescence and photostimulable luminescence, it is found that the transition of electrons provided by Ce³⁺ between the conduction band and trap levels offers the possibility for optical data storage. As evidence of its application, the optical information encoding using 254 nm light and decoding using a light stimulus and thermal stimulus were applied. These findings are expected to provide candidate material for novel optical storage technology, and further promote the development of advanced information storage technology.

This paper proposes an AI-based end-to-end decoding method for high-density optical disks. A bidirectional RNN directly decodes raw RF signals, overcoming the jitter-noise sensitivity of traditional PRML decoders. Experimental results show significantly improved BER under jitter noise with linear computational complexity O(N).