New transistor could slash global digital power budget by 5%

A new device in uno de los inventos más pequeños pero más grandiosos del siglo XX, el transistor, podría ayudar a alimentar el creciente apetito del mundo por la memoria digital, à la vez que reduciría hasta a 5% la energía de su dieta hambrienta de energy.

After years of innovations by Christian Binek of the University of Nebraska-Lincoln and Jonathan Bird and Keke He of the University at Buffalo, physicists recently teamed up to create the first magnetoelectric transistor.

In addition to reducing the power consumption of any microelectronics that incorporate it, the design of the equipment could reduce the number of transistors needed to store certain data by up to 75%said Nebraska physicist Peter Dowben, which would lead to smaller devices.

Moreover, it could equip these microelectronics with a steel memory that remembers exactly where their users left it, even after being turned off or suddenly losing power.

Several million transistors line the surface of every modern integrated circuit, or microchip, which is fabricated in staggering numbers – about 1 trillion in 2020 alone – from the industry’s favorite semiconductor material, silicon. By regulating the flow of electrical current through a microchip, the tiny transistor acts as a nanoscopic on-off switch that is essential for writing, reading and storing data such as the 1s and 0s of digital technology.

But silicon-based microchips are approaching their practical limits. These limitations compel the semiconductor industry to research and fund as many promising alternatives as possible.

There is a limit to what can be reduced. Basically we’re talking about 25 silicon atoms wide. And the heat is generated with each device on one (integrated circuit), so you can no longer transport enough heat to make everything work.

This situation occurs when the demand for digital memory and the energy required have exploded. The use of microchips in televisions, vehicles and other technologies has only increased demand.

You need something that works differently than a silicon transistor, to drastically reduce power consumption.

Typical silicon transistors consist of multiple terminals. Two of them, called source and drain, serve as starting and ending points for electrons flowing through a circuit. Above this channel is another terminal, the gate.

Applying a voltage between the gate and the source can determine whether electric current flows with low or high resistance, resulting in a buildup or absence of electron charges coded as 1s or 0s respectively. random access, the way most computer applications are based, requires constant power just to maintain those binary states.

So instead of relying on electric charge as the basis of their approach, the team turned to a magnetically related property of electrons that points up or down and can be read, such as charge electric, like a 1 or a 0.

The team knew that electrons passing through graphene, an ultra-strong material only one atom thick, can maintain their original spin orientations over relatively long distances, an interesting property to demonstrate the potential of a transistor based on on spintronics. In fact, controlling the orientation of these spins, using much less power than a conventional transistor, was a much more difficult prospect.

To do this, the researchers had to coat the graphene with the right material. Fortunately, Binek had already spent years studying and modifying precisely this material, chromium oxide. Chromium oxide is magnetoelectric, which means that the spins of atoms on its surface can change up and down, or vice versa, by applying a small amount of temporary voltage which consumes energy.

When a positive voltage is applied, the spins of the underlying chromium oxide point upwards, eventually forcing the spin orientation of the graphene electric current to shift to the left, producing a detectable signal in the process. . In contrast, a negative voltage causes the chromium oxide spins to point downwards, shifting the spin orientation of the graphene current to the right and generating a signal that is clearly distinguishable from the other.

As promising and functional as the team’s demo is, Dowben said there are many alternatives to graphene that share its one-atom thickness but also have properties better suited to a magnetoelectric transistor. The race to coat chromium oxide with these other 2D candidates is already underway, he said, marking “not something, but the start of something.”

Dowben recounted some of the team’s key developments. The realization that magnetoelectric materials could be a viable approach. The identification of chromium oxide. Modifying it, both to control its rotation with voltage rather than power-hungry magnetism, and to ensure that it performs well above room temperature, because, as Dowben said, “if you want to compete in the semiconductor industry, you can’t just work in Nebraska in the winter. He has to work in Saudi Arabia in the summer“. Then there were the theory-based computer simulations and several early-stage prototypes.

More information: onlinelibrary.wiley.com (English text).

Going through unl.edu

Leave a Comment