Close Mobile Menu

New, Silicon-Free, Atom-Thin Transistors Could Usher in Tomorrowland

July 18, 2016
by Glen Martin
grapheneforweb

Moore’s Law—as first put forth in 1965 by Berkeley alum and Intel and Fairchild Semiconductor co-founder Gordon E. Moore—postulates that the number of circuits on an integrated circuit will double every two years. Amazingly, the prediction (initially just an observation) has held up in the decades since, leading to computers that are ever smaller and ever more powerful.

But Moore’s Law is now running up against hard limits, due to the physical properties of silicon, the semiconducting material used in computer chips.

Silicon is good stuff. It’s plentiful—the most abundant element on planet Earth after oxygen—cheap, and amenable to industrial-scale semiconductor fabrication. The problem is that silicon-based transistors can only be shrunk so far before they start malfunctioning.

Silicon essentially is a “bulk material,” explains Mervin Zhao, a Cal Ph.D. student in applied science and technology. “With silicon, when the thickness of the transistor is comparable to its width or length, it’s harder to turn on and off,” Zhao says. “And that’s what a transistor is supposed to do—turn off, and then turn on. That’s how computations are made.”

Now, however, a group of Berkeley researchers, including Zhao, have found a way to create ultrathin, silicon-free transistors, using graphene—a spooky form of carbon that is only a single atom thick. In the lab, the researchers cut channels into the graphene then “seeded” the etchings with molybdenum disulfide (MoS2), a semiconducting material. The result: a transistor that Zhao describes as “basically two-dimensional.”

Using graphene to build a transistor, Zhao says, “avoids silicon’s major problem, which is greatly reduced functionality as you get into the nano realm.” But graphene, for all its advantages, is not a semiconductor.

Though graphene is carbon, and hence not a metal, it essentially functions as a metal at the nano scale, says Zhao: It conducts electricity well. “Graphene is always, ‘on,’” says Zhao. “It’s always conducting. We get the semiconducting capacity for our system not from the graphene, as you might expect, but from the MoS2.”

Zhao and his co-authors, including Xian Zhang, a senior scientist with the materials science division at Lawrence Berkeley National Laboratory, published their findings in a recent issue of Nature Nanotechnology.

Zhao notes that the big chip companies have been wrestling with silicon’s shortcomings for some time.

“Companies like Intel are always in the news for their new techniques and developments with silicon,” he says, “and they’ve done and are doing remarkable things that keep Moore’s Law going.” Short of a breakthrough though, Zhao says “a revised Moore’s Law is now in place. It’s not a dramatic decrease yet [in the two-year circuit-doubling cycle], but the slope is different, and it could get steeper.” The new graphene chips could reverse that.

So where could it all lead? Invisible supercomputers so thin that you could wear them like a second skin? Sentient, ant-sized robots smarter than a googolplex of Einsteins? Two-dimensional smart watches that you could paste to your corneas? Zhao hesitates to extrapolate.

“I think what we’ve done points to scalability, to more powerful computers, and to the possible perpetuation of Moore’s Law,” he says. “We’d be pretty happy with that.”

Share this article