Today, the state-of-the-art in chip technology resides at the 14nm mark. With increasing chip speed directly related to the density of transistors on a chip’s face it stands to reason that crunching the size of a transistor is critical to maintaining Moore’s pace.
According to IBM researchers, by 2020 the world’s most advanced processors will be littered with transistors that measure no more than 5nm across. However, to reduce the size of transistors any further carbon nanotubes or another material will be needed. "[At 5nm] that's where silicon scaling runs out of steam, and there really is nothing else," noted Wilfried Haensch, IBM's head of carbon nanotube development.
While the 2020 date looms large as the end of silicon-based chips, IBM has been creating and improving carbon nanotube transistors since 1998. Today chips can be made with 10,000 nanotube transistors embedded on their surface, but that’s a far cry from the millions of transistors that are required to process today’s software requests.
For the next six years IBM’s researchers will be hard at work developing processes which will improve the processing speed of carbon nanotube transistors while optimizing the placements of nanotubes onto a chip’s substrate. Running in parallel to those obstacles will be the task of ensuring that manufacturing new nanotube chips will be amenable to the semiconductor industry.
If IBM’s researchers can achieve the amazing feat they’ve set before themselves a whole new world of computing innovation might be unleashed. Carbon nanotube transistors could very well be the key to exaflop computing, accurate and holistic engineering simulations, genomic breakthroughs, and a myriad of other applications that are beyond today’s technologies.
Source: MIT Technology Review