Research: Mass production of graphene memristors finally works

Graphene-based switching elements have long been at the top of the wish list of many chip developers. In the case of memristors, production in the laboratory has already worked – a new breakthrough also makes production on an industrial scale possible.

New applications possible

Researchers from Queen Mary University of London and the company Paragraf Limited achieved this success in developing graphene-based memristors. The innovation you see in the trade magazine ACS Advanced Electronic Materials published, was achieved at wafer scale and paves the way for scalable production. Memristors are considered potential game changers in computer technology.

They enable analog calculations, store data without power supply 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, could significantly improve these switching devices. However, until now it has been difficult to integrate graphene into electronics on a large scale.

“Graphene electrodes offer clear advantages for memristor technology,” said Zhichao Weng, a scientist at Queen Mary University’s School of Physical and Chemical Sciences. “Not only do they improve lifespan, but they also enable exciting new applications, such as light-sensitive synapses and optically tunable memories.” Until now, producing high-quality graphene that is compatible with semiconductor processes has been a major hurdle. The new MOCVD (Metal-Organic Chemical Vapor Deposition) process now makes it possible to grow single-crystalline graphene directly on target substrates.

Also in industry

This scalable method is already used in commercial devices such as graphene-based Hall effect sensors and field effect transistors (GFETs). John Tingay, chief technology officer at Paragraf, said: “The ability to use graphene to develop new computing devices that combine logic and memory in innovative ways offers opportunities to reduce the energy costs of training large AI models.” The team used a multi-step photolithography process to integrate the graphene electrodes into the memristors.

These reproducible results show that mass production is possible. Professor Oliver Fenwick, head of research at Queen Mary University, added: “Our research not only confirms the proof of concept, but also that graphene is capable of improving the performance of memristors over other materials.”