Nonvolatile Resistance Switching in Atom Thin Memristors

A research team with members globally has created what seems to be the first working atomically thin memristor.

In the ongoing endeavor of creating smaller and more powerful devices, engineers have been vesting their developmental efforts into multi-integrated nanoscale components and concepts. One recent area of development has been in the area of nonvolatile resistance switching (NVRS,) specifically in atomically thin structures as we approach current device limitations.

Nonvolatile resistance switches differ from our current volatile switches in that they are capable of having high and low modulated resistance states that preserve their resistance values in the absence of a power supply. The NVRS phenomenon can be seen in our current memristor technology. However, development in the nanoscale has encountered issues due to current leakage as our current oxide based memristor devices approach atom thickness.

Recently, a collaborative effort between researchers from universities in China and the US has yielded what seems to be a working atom thin memristor.

Solving Atom-Scale Issues

Unlike our current oxygen-containing memristors, the researchers were able to create a working atomically thin memristor or “atomristor” out of various single-layer transition metal dichalcogenides (TMDs) that did not experience current leakage.

This phenomenon is considered an important breakthrough as our current bulk material memristors have encountered issues that hinder nanoscale integration. Inherently, these atomristors have fundamental differences from our current memristors because they lack a third dimension for the atoms to travel such as the oxygen atoms in traditional bulk memristors.

The researchers’ discovery was relatively unexpected in their study of monolayer TMDs and the reasoning behind why they work isn’t exactly clear. It was hypothesized by the research team member Deji Akinwande that existing defects in the atomically thin crystal lattice move around and tends to clump up when voltage of a particular polarity is run through it. This effect would decrease the resistance over the material, and when the opposite polarity is run through; the defects would scatter causing the resistance to increase.

Image courtesy of Nano Letters.

While researching this phenomenon, the researchers also discovered that it wasn’t exclusive to the molybdenum disulfide (MoS₂). The ability to perform as a 2D memristor appears to be shared between most if not all other TMDs as each TMD tested worked as a memristor. The team believes that the technology can be optimized to create advanced 3d chips by layering arrays of the atomristors onto logic chips.

“The sheer density of memory storage that can be made possible by layering these synthetic atomic sheets onto each other, coupled with integrated transistor design, means we can potentially make computers that learn and remember the same way our brains do,” said Akinwande

The discovery of a 2D memristor has broadened applications due to their size and capability for integration. One potential application that is currently being researched is the ability for the atomristor to function as a zero static power RF switch. The team was able to demonstrate a monolayer switch operating at 50GHz with acceptable loss. With further research into the atomristor RF switches; the authors believe that they can approach a figure of merit scaling to hundreds of THz by reducing device and component area.


The research team has been working with other scientists and theorists to further understand the process that is taking place. The team stated that they believed the atomristors have commercial value that can be integrated into current technologies, however for the time being they are unable to fully optimize the technology until they can further understand the unique process that is taking place in their atomristors. The original research article can be found in the American Chemical Society journal Nano Letters.