Team led by Professors Chul-Ho Lee and Gunuk Wang of KU-KIST Graduate School of Converging Science and Technology develops a 2D-material-based low-power brain-mimicking electronic device
Published and featured on the cover of Advanced Materials
A research team led by Professor Chul-Ho Lee and Gunuk Wang from the KU-KIST Graduate School of Converging Science and Technology at Korea University has succeeded in using phase-engineered 2D heterostructures to develop a new electronic device capable of mimicking the human brain while offering low power consumption and interference-controlled operation. The team overcame the limits of two-terminal devices, which have been investigated in previous memristor research, and presented a new direction in terms of device control and power consumption. The paper was featured on the cover of Advanced Materials (impact factor: 21.950), one of the leading journals in materials science, on August 27.
* A word coined from “memory” and “resistor.” Refers to a memory device with two or more resistance states in response to electric stimuli. It is capable of continuously adjusting these states, and non-volatile memory retains the previous resistance state even when the power is cut off.
▲ (Clockwise from top left) Professor Chul-Ho Lee, Professor Gunuk Wang, Seonghoon Jang, and Woong Huh
The team developed a brain-mimicking electronic device comprised of a vertically integrated memristor, which was used to store information on resistance states, and an electrostatically controlled barristor using 2D materials with an atomic-layer thickness. The device emulates the memory and learning characteristics of various synapses in the brain. A three-terminal configuration was used to examine different methods of varying the speed of information storage, memory, and learning. The results can be utilized in low-power neuromorphic applications, image recognition, and machine learning.
* A Schottky diode fabricated by combining metallic graphene and semiconducting 2D material. Because the work function of graphene can be adjusted using an electric field, the device can regulate current flow in a manner similar to transistors.
▲ Schematic representation of the proposed 2D-material-based electronic device and synapse (left), and the cover of the journal (right)
The team said, “We have proposed an electronic device capable of mimicking neural functions based on the intrinsic characteristics of 2D materials. This will facilitate the development of low-power, high-efficiency neuromorphic-based AI technology.”
This study was supported by the Basic Research Project in Science and Engineering offered by the National Research Foundation, the KU-KIST School Project, and a Korea University Future Research Grant (KU FRG).
- Title of Paper: Synaptic Barristor Based on Phase-Engineered 2D Heterojunctions
- Authors: Eleven authors including Woong Huh (Korea University, first author), Seonghoon Jang (Korea University, co-first author), Gunuk Wang (Korea University, co-corresponding author), and Chul-Ho Lee (Korea University, corresponding author)