Smaller computers coming as microchips downsize

COMPUTERS and the microchips that make them work may soon undergo a significant downsizing

COMPUTERS and the microchips that make them work may soon undergo a significant downsizing. A research team at the Tyndall National Institute has found a way to shrink transistors down until their dimensions can be counted in just atoms across, writes DICK AHLSTROM

Tyndall has become the first lab in the world to design and make a “junctionless transistor”, says Prof Jean-Pierre Colinge, who leads Tyndall’s Micro Nanoelectronics Centre, one of four research centres within the institute.

Transistors are the little pieces of technology that allow microprocessors and computers to do what they do. Generally the more you have the better and size matters – the smaller they are the more you can squeeze into a given space.

There has been a relentless drive towards transistor miniaturisation since the first working transistor was built at Bell Laboratories in the US in December 1947.

READ MORE

In 1970, the smallest one made allowed designers to pack 1,000 transistors onto a microchip “the size of a postage stamp”, Colinge says. Now the same-sized chip can hold 32 billion. You could pack 300,000 modern transistors into the area covered by the cross- section of a single human hair.

And now Colinge and his group have taken this much further. His “junctionless transistor” is just dozens of atoms across. His design is about seven times smaller than the current smallest transistor in the latest microchips being produced by Intel, he adds.

Intel’s best is about 35 nanometres (a billionth of a metre) across using the transistor design currently in widespread use. “Everybody agrees when you get to 15 nanometres things don’t work any more,” Colinge says.

“With the new device, there is no limit to the size. You can go, in principle, to five nanometres across.” He has also shown that his new transistor design works perfectly well despite its size.

His inspiration for the new design came from an unlikely source – the theoretical design for a transistor patented all the way back in 1928. “It goes back to the simplest possible way of making a transistor,” he explains.

Transistors work by being able to either halt or allow an electrical current to flow. Today’s transistors use a sandwich of silicon layers. The layers will form “a junction” if they have different polarities, but it is becoming increasingly difficult to make working transistors much smaller than the sizes currently being built.

The challenge is to keep the layers from “smearing” into one another, Colinge says, something that stops the transistor from working properly.

Instead he developed a transistor based on the earlier design, one that looks like a finger sporting a wedding ring. “That structure was first proposed in 1925. It was never made, and later more complex [transistor] designs were introduced,” he says.

This design had a wire running through a ring, an electrical structure that controls the flow of electricity through the wire. The ring can allow or stop the current flowing, but doesn’t need the layers of silicon seen in transistors that use junctions.

He likens the design to holding a garden hose in your fist. Water can flow through the hose, like current running through a wire, but if you squeeze your hand the flow stops. His design uses a silicon nanowire just dozens of atoms across running through the ring that can stop or allow the current flow.

The new design is much easier and cheaper to make, is much smaller and works despite its minute size, Colinge says. Its size also means it will use less energy and will allow the devices in which it will be used to shrink accordingly.