IBM, along with the Georgia Institute of Technology, reached this speed by chilling the chip to what IBM said was nature’s coldest possible temperature, or 451 degrees below zero Fahrenheit.
David Ahlgren, a senior engineering manager in IBM’s labs, said the chip has no practical application on Earth at this freezing temperature.
But it does enable IBM to study and develop future technologies surrounding chips at this speed. The research will enable IBM to explore performance barriers of silicon to make possible a new type of fast, low-energy chips.
In other words, IBM expects it will help it get a jump on future high-speed chip technologies.
We’re saying we’re not limited, Ahlgren said. We understand the limits of the device.
There also are potential applications for this chilly chip technology in deep space, where such extremely cold temperatures naturally occur, Ahlgren said. But that’s not really what we’re interested in doing.
We’ve been trying to push those limits at room temperature and I think any practical applications have to occur at room temperature, he said. Once we push the process even further — making changes in the process as time goes on — we should be able to operate at room temperature.
At room temperature, the chips used in the research operated at about 350 GHz.
The researchers’ work was based on silicon germanium chips, the standard embedded semiconductors used in cell phones, notebooks, Internet backbone components and various handheld devices. This means IBM can manufacture its research chips using its conventional 200-mm wafer facility in Burlington, Vermont, Ahlgren said.
While 451 degrees below zero Fahrenheit doesn’t happen naturally on Earth, the researchers used liquid helium to cryogenically chill the chip.
The term frozen was used to be able to connect with the public, Ahlgren said. It’s not frozen because any silicon chip is solid and you can’t freeze a solid. But rather it refers to the fact we’re taking the temperature down.
The colder the silicon germanium chip, the faster its performance because the low temperatures change the semiconductor properties, he explained.
At lower temperatures, there is less scattering of electrons, which means they travel directly from point A to point B on the chip, in more or less a straight line. At higher temperatures, electrons tend to wander off their paths somewhat, which slows them down.
Exactly when this research will lead to faster mobile chips is difficult to predict. But, at the very least, the research shows silicon germanium will be around for 10 years or so, Ahlgren said.
We’re showing … we know there’s certainly a lot of headroom for improvement, he said.
