Graphene switch combines logic and memory functions in a single device

Researchers at Manchester University in the UK have used graphene to make a new electrically-controlled switching device that supports both memory and logic functions. The device, which exploits graphene’s ability to conduct protons as well as electrons, might also be used in applications that involve an electrode-electrolyte interface, such as reducing carbon dioxide to its component chemical species.

Graphene is a two-dimensional sheet of carbon atoms arranged in a honeycomb-like hexagonal lattice. One unique property of graphene is that electrons move freely through the plane of this sheet at almost ballistic speeds, making it a better conductor than metals. Another fascinating property is that when an electric field is applied between the top and the bottom of the sheet, protons from an adjacent polymer or electrolyte will flow through it in a perpendicular direction.

These flows of particles are not independent, however. Some of the protons bind to the electrons, and this binding process, known as hydrogenation, produces defects that scatter and slow the remaining unbound electrons. When the number of bound electrons reaches a threshold, the material turns into an insulator. The material’s conductivity can then be restored by applying an electric field in the plane of the graphene sheet, injecting more electrons into it.

Controlling the movement of the electrons and protons independently

Researchers led by Marcelo Lozada-Hidalgo have now exploited these proton and electron flows to perform logic and memory operations in a single device for the first time. The device consists of a micron-scale graphene layer sandwiched between proton-conducting electrolytes that are connected to gate electrodes on the top and bottom of the device. Additional electrodes placed at the device’s edges induce electrons to flow through the graphene sheet. This arrangement allows the researchers to simultaneously measure the graphene sheet’s in-plane electrical conductivity and the degree to which protons permeate it in the out-of-plane direction.

Crucially (and unexpectedly, Lozada-Hidalgo says), the two-gate set-up also gave the researchers independent control over proton transport and the electron-proton binding that determines whether graphene is a conductor or an insulator. “We can drive proton transport without hydrogenating graphene or hydrogenate graphene without driving proton transport, or both,” he tells Physics World.

The source of this independent control, he explains, is that both hydrogenation and proton transport depend on the electric field E and the charge density n of the graphene. Using a non-aqueous electrolyte allows both E and n to be extremely high, which effectively distorts the energy profile for these processes. And while E depends on the difference between the top and bottom gate voltages, n depends on their sum, making it possible to tune E and n independently simply by altering the voltages. “Such control is impossible otherwise and is so robust and reproducible that we can exploit it to perform proton-based logic-and-memory operations in graphene,” Lozada-Hidalgo says.

A very different computing platform

In the latest study, which is published in Nature, the researchers demonstrated this capability by using the graphene layer’s conducting or insulating status as a “memory” state. At the same time, they used the proton current to perform a logic operation called the exclusive operation (XOR) that outputs a “1” when the number of inputs with a value of 1 is odd, and a “0” otherwise. Hence, when the top and bottom electrode voltages differed, the XOR operation yielded a 1, and a strong proton current flowed – without changing the state of the memory.

The fact that both logic and memory operations occur in the same device is significant, Lozada-Hidalgo says, because these functions are usually performed by separate circuit elements that are physically isolated from each other within a computer. This can lead to long data-transfer times and high power consumption. “Our work could perhaps enable low-cost analogue computing structures that operate on protons,” he says. “At the very least, it is a very different computing platform that does not require silicon and can be implemented in very simple and potentially cheap devices.”

The fact that it uses protons, rather than electrons as in conventional circuits, could also make it possible to couple these devices with biological systems or electrochemical interfaces, he adds.

“A host of application areas”

While the Manchester researchers have so far only demonstrated these processes in graphene, they say that any 2D crystal could be studied in this way. “This represents a great opportunity for investigating electrode-electrolyte interfaces in a large group of materials and over a parameter space that is inaccessible in classical interfaces,” Lozada-Hidalgo says.

As well as memory-logic devices, he adds, the effect could be of interest in “a host of application areas”, including nanofluidics, catalysis, electrochemistry and surface science. “It’s a new technological capability in our discipline,” he says. “The electrochemical processes at play can be linked to the electronic properties of the 2D crystals because they can induce conductor-insulator phase transitions or strongly dope the materials and their heterostructures. We are looking into these possibilities now.”