IBM breakthrough anticipates changes to chip industry

Intel may be under pressure from ARM chips in the short term, but the chip maker is facing some longer term challenges as the end of Moore’s Law approaches.

On Sunday, IBM announced a chip production breakthrough that forgoes any reliance on silicon semiconductors, allowing the fabrication of chips using carbon nanotube technology.   A team of scientists working for the company claim to have developed a production process to cram over ten thousand transistors onto a chip using carbon nanotubes, offering a potential alternative to silicon devices, and opening up the possibility of developing smaller and more powerful computers into the future.

The chip industry has so far relied on design engineers’ ability to consistently create smaller and faster processors, but as development begins to reach atomic scale any speed or size improvements using silicon based transistors will become more difficult due to the laws of physics.

This has meant investment in alternative types of semiconductor production, with the goal of finding a suitable replacement technology as designers approach ever smaller processes.   Graphene has been another high profile example, and the likes of Samsung and IBM are working on developing the atom-thick material into a working product for chips in the future.  Of course any developments made in the lab are distinctly different to allowing millions of chips to be cheaply and reliably produced on large scale.

Interestingly though IBM appears to have made some steps towards production of a processor using carbon nanotubes, rolled up sheets of carbon that have more beneficial electronic properties to silicon.

So far it has only been possible to produce chips using circuits with only a few hundred carbon nanotube transistors.  This is a miniscule amount compared to the billions of transistors that are currently crammed onto silicon based chips.  However IBM has managed to fit over ten thousand transistors onto a circuit using a new placement technique, which importantly uses conventional production processes.

Although the prospect of silicon being replaced by any other material is still some way off, carbon nanotubes would have a number of advantages such as being able to function more efficiently at smaller, atomic-level, processes, with a five to ten times performance increase the researchers say.

Moore’s Law still has a good few years to go yet, but after a few more generations of chips, a new technology will have to be found.  This will mean rewriting the rulebook as far as architectures are concerned, and it is likely that large shifts of power would happen in the chip industry.