Graphene-like logic microprocessors have promise for Internet of Things

Graphene Flagship in Vienna unveils molybdenum disulphide integrated circuit with potential for low-power electronics

The molybdenum disulphide integrated circuit

While graphene is the undoubted star of ultra-thin materials, other substances that can be separated into crystals only a few atoms thick are also attracting attention. The Graphene Flagship at the Technical University of Vienna (TU Vienna) has now revealed research on an integrated circuit constructed from molybdenum disulphide (MoS2), a material known as a transition metal dichalcogenide that exhibits semiconducting properties.

The processor chip contains 115 transistors, is capable of one-bit logic operations and is scalable to multibit operations that would be more useful for computing, claim the researchers. Because they are so thin and lightweight, they could potentially be used in wearable systems and other ubiquitous computing applications without adding bulk, they add.

“In principle, it’s an advantage to have a thin material for a transistor. The thinner the material, the better the electrostatic control of the transistor channel, and the smaller the power consumption,” said Thomas Mueller (TU Vienna), who led the work.

Mueller believes that such processors might be combined with LEDs to make a new type of smart paper, or could form components of the intelligence of smart sensors for Industry 4.0 or IoT applications. This, however, would entail increasing the complexity of the system. “Our goal is to realise significantly larger circuits that can do much more in terms of useful operations. We want to make a full 8-bit design – or even more bits – on a single chip with smaller feature sizes.” His colleague Stefan Wachter, who co-authored a paper on the research in Nature Communications, added that even increasing the design to two bits would double the complexity of the circuit.

However, the team is already formulating a strategy for increasing the number of transistors on the chip. One idea might be to grow the structures as crystals directly, rather than transferring them from a fabrication process. The medium-term goal is to find a way to produce a chip with tens of thousands of transistors, although this is still well below conventional silicon microchips which can have billions.