High-speed transistor channel developed using a core-shell nanowire structure

Hello, forums people. Here's your dose of current events for the day:

A research group led by Naoki f**ata, International Center for Materials Nanoarchitectonics, National Institute for Materials Science (NIMS), and a research group at Georgia Institute of Technology jointly developed a double-layered nanowire, consisting of a germanium (Ge) core and a silicon (Si) shell, which is a promising material for high-speed transistor channels. In addition, the groups verified that the Si layer, which was doped with impurities, and the Ge layer, which transports carriers, were not intermixed, and that carriers were generated in the Ge layer. These results suggest that the new nanowire may effectively suppress the scattering of impurities, which had been a problem with conventional nanowires, thereby taking a major step toward the realization of a next-generation high-speed transistor.

Link: user link on phys.org

Interesting.

Making a note of that for sure.

Heheheh

ScottyMahoney, you said about two hundred two words about this. Motor mouth.

As long as the density factor is solved (packing too many transistors on such a small silicon wafer) this sound promising.

Increasing clock speed is one of the two major ways computers can be sped up (the other is increasing density but we're coming close to hitting a wall with that - quantum effects start to come into play once you make the transistors so small) but with increased speed comes increased heat and heat dissipation requirements are getting crazy as it is.

Unless computers all switch to ARM processors.

The point about the heat is good. I couldn't find anything in the article that addressed the issue of thermal effects causing limitations.

I know in flash memory chips, quantum tunneling is a part of how they actually work. I did not realize that the scale of CPUs had gotten to the point where electrons would have too high a probability of tunneling across circuit paths to become unreliable. Are we really at that point? It just doesn't seem likely.

ARM processors are great in mobile devices (and Raspberry Pi!) but lack the power that's expected from desktop workstations. I wouldn't want to render 3D or video on an ARM just yet. Maybe that will change.

Although the article doesn't mention it, the reduction in impurities will also have the effect of reducing the resistance in the component, which in turn reduces heat. So, simply put, it does that as a matter of course.