Researchers from Queen Mary University of London and Paragraf Limited have demonstrated a significant step forward in the development of graphene-based memristors and unlocking their potential for use in future computing systems and artificial intelligence (AI).
This innovation, published in ACS Advanced Electronic Materials and featured on the cover of this month's issue, has been achieved at wafer scale. It begins to pave the way toward scalable production of graphene-based memristors, which are devices crucial for non-volatile memory and artificial neural networks (ANNs).
Memristors are recognized as potential game-changers in computing, offering the ability to perform analog computations, store data without power, and mimic the synaptic functions of the human brain.
The integration of graphene, a material just one atom thick with the highest electron mobility of any known substance, can enhance these devices dramatically, but has been notoriously difficult to incorporate into electronics in a scalable way until recently.
"Graphene electrodes bring clear benefits to memristor technology," says Dr. Zhichao Weng, Research Scientist at School of Physical and Chemical Sciences at Queen Mary. "They offer not only improved endurance but also exciting new applications, such as light-sensitive synapses and optically tunable memories."
One of the key challenges in memristor development is device degradation, which graphene can help prevent. By blocking chemical pathways that degrade traditional electrodes, graphene could significantly extend the lifetime and reliability of these devices. Its remarkable transparency, transmitting 98% of light, also opens doors to advanced computing applications, particularly in AI and optoelectronics.
This research is a key step on the way to graphene electronics scalability. Historically, producing high-quality graphene compatible with semiconductor processes has been a significant hurdle. Paragraf's proprietary Metal-Organic Chemical Vapor Deposition (MOCVD) process, however, has now made it possible to grow monolayer graphene directly on target substrates.
This scalable approach is already being used in commercial devices like graphene-based Hall effect sensors and field-effect transistors (GFETs).
"The opportunity for graphene to help in creating next generation computing devices that can combine logic and storage in new ways gives opportunities in solving the energy costs of training large language models in AI," says John Tingay, CTO at Paragraf.
"This latest development with Queen Mary University of London to deliver a memristor proof of concept is an important step in extending graphene's use in electronics from magnetic and molecular sensors to proving how it could be used in future logic and memory devices."
The team used a multi-step photolithography process to pattern and integrate the graphene electrodes into memristors, producing reproducible results that point the way to large-scale production.
"Our research not only establishes proof of concept but also confirms graphene's suitability for enhancing memristor performance over other materials," adds Professor Oliver Fenwick, Professor of Electronic Materials at Queen Mary's School of Engineering and Materials Science.