Home > All news > Industry news > Atomically Thin Memory Resistors Usher an New Era of Neuromorphic Computing
芯达茂F广告位 芯达茂F广告位

Atomically Thin Memory Resistors Usher an New Era of Neuromorphic Computing

The University of Kansas and the University of Houston have reportedly launched a $1.8 million grant from the National Science Foundation's Future Semiconductor Project 2 (FuSe2) to develop atomically adjustable memory resistors, called memristors, for advanced computing systems that mimic the human brain. In addition, the program will train future technical talent for the U.S. semiconductor industry.

The FuSe2 project was established in 2023 and focuses on solving key challenges in future semiconductor R&D. Industry partners include Micron, Intel and Samsung.

The project was led by Judy Wu, Distinguished Professor of Physics and Astronomy at the University of Kansas, and her team includes Hartwin Peelaers, associate professor of physics and astronomy at the University of Kansas, and Francisco Robles of the University of Houston. Their research goal is to solve the important challenge of "brain-inspired computing".

This next-generation computing approach is designed to achieve high speed and energy efficiency, mimicking the way the human brain works, and optimizing computing for artificial intelligence. The system consists of a series of memristors that mimic artificial synapses and neurons.

Wu and her colleagues have adopted a co-design approach that blends materials design, fabrication, and testing to enable precise manipulation of oxide-semiconductor memristors at the atomic level to give brain-inspired circuits the functionality they need.

Their research will address a long-standing challenge in materials research: the ability to accurately stack a small number of atomic layers to achieve the functionality and large-area uniformity required for future semiconductor devices. These atomic layers are ten times thinner than the nanoscale (by comparison, a piece of paper is about 100,000 nanometers thick).

"The innovation that received this grant is an ultra-thin memristor based on ultra-wide bandgap semiconductors, such as gallium oxide, whose electronic structure is controlled by atomic-precision theoretical simulations," said Judy Wu. "This is the result of a collaboration between three of our researchers, using a co-design approach to achieve an simulation-guided memristor design."

Figure: Atomic-level thin memory resistors: opening a new era of brain-like computing (Source: KU News)

Figure: Atomic-level thin memory resistors: opening a new era of brain-like computing (Source: KU News)

It is worth mentioning that Judy Wu and his team were the first to demonstrate a memory with a thickness of less than 2 nanometers.

"This is one of our important inventions," Wu says, "and we have mastered a unique technique that allows us to stack just a few layers of atoms on top of each other precisely."

By achieving thin films with thicknesses ranging from 0.1 to less than 2 nanometers, the research team at the University of Kansas is breaking new ground in science.

"We were able to stack selected atomic layers on top of each other," Wu added, "and our core goal was to develop 'tunable memristors' that can act as neurons and synapses in brain-inspired circuits to support brain-inspired computing." With this circuit, we hope to use this circuit to simulate the brain's ability to think, calculate, make decisions, and recognize patterns, and achieve high speed and energy efficiency.”

The research will also lead to a STEM talent development program within the two universities, with a view to developing diverse and highly skilled talent for the semiconductor industry through education and outreach activities. The team includes professors Heather Domjan, Dr. Long Hae-young, and members such as Teresa MacDonald, Eleanor Gardner, and Carolyn Cocken.

"This program has a special focus on developing talent in the semiconductor industry," noted Judy Wu, "and we plan to run a week-long workshop every summer for three years. The event will focus on the children of immigrants who, although U.S. citizens, face a lack of educational resources due to their parents' seasonal work, and have limited access to semiconductors, high-tech microelectronics, and future cleanroom micromanufacturing.”

Wu and her team will recruit immigrant family students from Kansas, Texas, and beyond to provide learning opportunities for students with limited financial circumstances.

"We want to introduce them to the field of high-tech semiconductors to spark their interest in the field," Wu said, "Many of our PhD students are now working at companies such as Intel, Honeywell, Thar Semiconductor, Blue Source, Lockheed Martin and ASM – and these students can follow similar career paths in the future." In addition, we plan to develop a series of online hands-on activities to expand our social impact.”

 

Original:Atomically Thin Memory Resistors Will Optimize Semiconductors for Neuromorphic Computing

Related news recommendations

Login

Registration

Login
{{codeText}}
Login
{{codeText}}
Submit
Close
Subscribe
ITEM
Comparison Clear all