KAIST "Enhancing Neuromorphic Computing Reliability, Early Commercialization Expected"

A technology to enhance the reliability of neuromorphic computing has been developed in Korea. Neuromorphic computing is a technology that mimics the human brain to implement hardware-based artificial intelligence computation and has recently gained attention. However, the conductive change device (memristor) used as a unit device has shown limitations in implementing large-scale neuromorphic computing systems due to irregular device characteristics that reduce reliability. The developed technology is expected to overcome these limitations and accelerate the commercialization of neuromorphic computing.

KAIST announced on the 21st that Professor Choi Seon-hyun's research team from the Department of Electrical Engineering and Computer Science, in collaboration with Hanyang University’s research team, has developed an ‘Aliovalent ion’ doping method that can improve the reliability and performance of next-generation memory devices.

Aliovalent ions refer to ions that have a different valence (a measure of covalent bonding) from the originally existing atoms.

(From left) Professor Choi Seon-hyun, Master’s student Bae Jong-min from the Department of Electrical Engineering at KAIST, Postdoctoral researcher Kwon Cho-a, and Professor Kim Sang-tae from the Department of Nuclear Engineering at Hanyang University. Provided by KAIST

The joint research team improved the uniformity and performance of the devices by doping aliovalent ions to address the biggest problem of irregular device characteristic changes in existing next-generation memory devices. They elucidated the improvement mechanism through experiments and atomic-level simulations.

They also reported that injecting aliovalent halide ions appropriately into the oxide layer can enhance the reliability and performance of the devices, solving the problem of irregular device reliability.

The joint research team experimentally confirmed that this method can improve device operation uniformity, operating speed, and performance.

In particular, atomic-scale simulation analysis revealed that the device performance improvement effect, consistent with experimental results, appears in both crystalline and amorphous environments.

According to the joint research team, the doped aliovalent ions attract nearby oxygen vacancies, enabling stable device operation, and expand the space around the ions, allowing for faster device operation.

Professor Choi Seon-hyun said, "The aliovalent ion doping method developed this time is a revolutionary way to significantly improve the reliability and performance of neuromorphic devices," adding, "The research results can contribute to the commercialization of next-generation memristor-based neuromorphic computing, and the performance improvement principles confirmed in experiments can be applied to various semiconductor devices."

Meanwhile, this research was conducted with support from the National Research Foundation of Korea’s New Device Core Technology Development Project, New Material PIM Device Project, Excellent Young Researcher Project, Nano Convergence Technology Institute’s Semiconductor Process-Based Nano Medical Device Development Project, and the National Supercomputing Center Innovation Support Program.

The first authors are Master’s student Bae Jong-min from KAIST’s Department of Electrical Engineering and Computer Science and postdoctoral researcher Kwon Cho-ah from Hanyang University. The research results were published in the June issue of the international journal Science Advances.

This content was produced with the assistance of AI translation services.