The February 2023 issue of IEEE Spectrum is here!

Close bar

The Most Complex 2D Microchip Yet

The new molybdenum disulfide microprocessor has 115 transistors

3 min read
Image credit:  Photo: Stefan Wachter
Image credit: Photo: Stefan Wachter

A three-atom-thick microchip with more than 100 transistors is the most complex microprocessor made from a 2-dimensional material to date, researchers say.

The new device is made of a thin film of molybdenite, or molybdenum disulfide (MoS2), which consists of a sheet of molybdenum atoms sandwiched between two layers of sulfur atoms. A single-molecule layer of molybdenum disulfide is only six-tenths of a nanometer thick. In comparison, the active layer of a silicon microchip is up to about 100 nanometers thick. (A nanometer is a billionth of a meter; the average human hair is about 100,000 nanometers wide.)

Scientists hope two-dimensional materials such as graphene or molybdenite will allow Moore's Law to continue once it becomes impossible to make further progress using silicon. Whereas graphene is an excellent conductor, making it ideal for use in wiring and interconnections, molybdenite is a semiconductor, which means it can serve in the transistor switches that lie at the heart of electronic circuits.

The scientists detailed their findings online April 11 in the journal Nature Communications.

Until now, devices made from 2D materials consisted of up to only three or so transistors, says study senior author Thomas Müller, an electrical engineer at the Vienna University of Technology in Austria. Now, he and his colleagues have created a molybdenite microchip with 115 transistors.

The device was fabricated on a silicon wafer, but the researchers say it could in principle be made on a flexible backing. “I think the most exciting implication of our work is that it opens new possibilities to realize circuits on flexible—or, in general, arbitrary—substrates for novel applications,” Müller says.

The device can execute user-defined programs stored in external memory, perform logical operations and transmit data to its periphery. Although this prototype operates just on single-bit data, the researchers stress their design is readily scalable to multi-bit data. They also note the device is compatible with existing semiconductor manufacturing processes.

The total power consumption of the circuit is roughly 60 microwatts, and it operates at frequencies between 2 kilohertz and 20 kilohertz, the researchers say. “In terms of performance, our device is of course by no means competitive with current silicon-based microprocessors,” Müller says. “It is just a very first step towards a new generation of electronic devices.”

The minimum size of features in this prototype is 2 micrometers. However, “going to 200- or 100-nanometer transistor channel length should be rather straightforward,” Müller says. In the future, by improving the quality of electrical contacts (the electrically conductive circuit components used to make and break electrical connections) “the ultimate scaling limit of 2D transistors is about 1 nanometer,” he adds. “This can not be reached with silicon; the limit there is around 5 nanometers.”

The main limitation now to building larger, more complex microchips has to do with improving the number of transistors they can successfully fabricate, Müller says. Currently their yield for fully functional chips is only a few percent.

The main way to increase transistor yield is to grow more uniform molybdenite films, Müller says. Imperfections in the molybdenite films crop up when they are transferred to their target wafers from the sapphire substrates on which they are grown, he explains. The researchers are now working on ways to directly grow the molybdenite directly on the target wafer to remove this transfer step.

“If the uniformity of the molybdenum disulfide films can be improved, then it should be rather straightforward to increase the complexity of two-dimensional circuits to tens of thousands of transistors,” Müller says.

In order for 2D microchips to incorporate hundreds of millions of transistors, as is seen in modern silicon chips, scientists will have to switch from the n-type metal-oxide–semiconductor (NMOS) design used in this prototype to the lower power-consumption complementary MOS (CMOS) designs use in conventional microchips today, Müller says.

“This probably will require another two-dimensional semiconductor material instead of molybdenum disulfide, but there are plenty—for example, tungsten diselenide,” Müller says.

A version of this article appears in the June 2017 print issue as “The Most Complex 2D Microchip Yet.”

The Conversation (0)

Two Startups Are Bringing Fiber to the Processor

Avicena’s blue microLEDs are the dark horse in a race with Ayar Labs’ laser-based system

5 min read
Diffuse blue light shines from a patterned surface through a ring. A blue cable leads away from it.

Avicena’s microLED chiplets could one day link all the CPUs in a computer cluster together.

Avicena

If a CPU in Seoul sends a byte of data to a processor in Prague, the information covers most of the distance as light, zipping along with no resistance. But put both those processors on the same motherboard, and they’ll need to communicate over energy-sapping copper, which slow the communication speeds possible within computers. Two Silicon Valley startups, Avicena and Ayar Labs, are doing something about that longstanding limit. If they succeed in their attempts to finally bring optical fiber all the way to the processor, it might not just accelerate computing—it might also remake it.

Both companies are developing fiber-connected chiplets, small chips meant to share a high-bandwidth connection with CPUs and other data-hungry silicon in a shared package. They are each ramping up production in 2023, though it may be a couple of years before we see a computer on the market with either product.

Keep Reading ↓Show less