Researchers has developed a novel ferroelectric optoelectronic memory device that integrates sensing, memory, and logic operations, offering potential for artificial intelligence (AI) and Internet of Things (IoT) applications. Published in ACS Applied Materials & Interfaces, the study highlights the device's ability to perform high-speed ultraviolet (UV) photodetection and long-term data storage. The research team, led by SHEN Dezhen from the Changchun Institute of Optics, Fine Mechanics and Physics, designed a p-GaN/ZnGa₂O₄/BaTiO₃/n-ITO heterojunction device. This structure combines wide-bandgap semiconductors with ferroelectric materials to achieve superior optoelectronic performance. The device demonstrated a peak responsivity of 7 mA/W at 0 V bias, with fast response times of 6 ms (rise) and 12 ms (fall). Additionally, the device exhibited excellent long-term storage characteristics, retaining its photocurrent for over five months.
Traditional computing systems based on the Von Neumann architecture face challenges due to the physical separation of memory and computational units, leading to high power consumption and speed losses. The new ferroelectric optoelectronic memory addresses these issues by integrating sensing, memory, and computing into a single device, significantly improving efficiency and reducing energy consumption.
The device's fabrication involved growing a p-GaN thin film on a GaN/sapphire substrate, followed by the deposition of ZnGa₂O₄ and BaTiO₃ layers using metal-organic chemical vapor deposition and magnetron sputtering, respectively. An indium tin oxide (ITO) film was then added, and the device was annealed to improve crystal quality and conductivity. The resulting heterojunction demonstrated high rectification ratios and excellent UV photodetection performance.
The researchers also demonstrated a 5 × 5 ferroelectric optoelectronic memory array capable of imaging, storing, and reading out information. The device could perform "AND" and "OR" logic gate operations based on its initial polarization state and input signals, showcasing its potential for in-memory computing and AI applications.
This study opens new avenues for the application of ferroelectric materials in AI and IoT, offering a promising solution for integrated sensing, memory, and computing systems.
中文
