Integration of Non‐Volatile Multi‐Bit Storage and Logic Computing in GaN Enhancement‐Mode Devices for In‐Memory Computing
Silicon‐based in‐memory computing, through seamless integration of computation and storage capabilities, offers a promising solution to the data transfer bottlenecks inherent in traditional von Neumann architectures, addressing the growing demands of big data and artificial intelligence applications...
Saved in:
| Published in: | Advanced functional materials |
|---|---|
| Main Authors: | , , , , , , , |
| Format: | Journal Article |
| Language: | English |
| Published: |
28.08.2025
|
| ISSN: | 1616-301X, 1616-3028 |
| Online Access: | Get full text |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Silicon‐based in‐memory computing, through seamless integration of computation and storage capabilities, offers a promising solution to the data transfer bottlenecks inherent in traditional von Neumann architectures, addressing the growing demands of big data and artificial intelligence applications. Despite its potential, research on non‐volatile memory and in‐memory computing based on third‐generation semiconductor materials like gallium nitride (GaN) remains relatively underexplored. Herein, this study successfully fabricates GaN enhancement‐mode devices by optimizing Complementary Metal Oxide Semiconductor (CMOS)‐compatible processes and demonstrates their multi‐level non‐volatile memory functionality alongside in‐memory computing capabilities. By leveraging intrinsic defects at the interfaces, charge‐trapping mechanism is utilized to achieve exceptional non‐volatile memory characteristics, including a stable memory window of up to 5.01 V, an on/off current ratio exceeding 10 4 , remarkable reliability (retention time >10 6 s, endurance > 10 4 cycles), and robust resistance to read disturbances (>10 4 cycles), even under high temperature (125 °C). Adjusting the programming gate voltage further enables high‐density multi‐bit data storage, supporting complex data patterns and significantly enhancing storage density. Moreover, logic operations such as implication (IMP) and False are directly implemented within GaN memory, embedding computational functionality into the storage process. These findings underscore the potential of GaN‐based in‐memory computing as a transformative approach for next‐generation computing systems. |
|---|---|
| ISSN: | 1616-301X 1616-3028 |
| DOI: | 10.1002/adfm.202513256 |