In 1965 Intel co-founder Gordon Moore penned his famous Moore's Law, which states that computer chips will double the number of transistors (mechanism that regulates power in a computer) every two years. For nearly half a century this law has held true and computing power has exponentially increased while costs have plummeted. This has made possible revolutions in personal computers and consumer technology.
For example, comparing the iPhone 5s to the computer guidance system that put men on the moon we find that the iPhone 5s is 1,106 times less expensive, yet has 16,000 times the memory and is 1,300 times faster.
The key to Moore's Law has been miniaturization, with transistor sizes dropping steadily to the point that they are now approaching one thousandth the width of a human hair.
Source: Intel Global Leadership Summit Presentation, page 3
Source: Source: Intel Global Leadership Summit Presentation, page 4
The End of Moore's law
Yet despite all the money and efforts of thousands of engineers and scientists around the world, the fact is that Moore's Law as we know it will at some point come to an end. This is because the physical size of a transistor can only get so small, and at the nanoscopic scale chip makers are working with, quantum physics becomes a problem, with electrons tunneling from one place to another in an uncontrolled manner.
Currently, the latest silicon chip designs are 14 nanometer (nm) in scale (meaning 14 billionths of a meter across). IBM, Intel, and Applied Materials believe they may eventually achieve 7 nm, 5 nm, and 3 nm scales respectively, using silicon.
To continue advancing computer technology past these scales will require alternatives to silicon, and right now carbon seems to be the candidate.
IBM's big graphene bet
IBM has recently announced a $3 billion, five-year effort, spread across labs in New York, California, and Switzerland. The goal of this lofty endeavor is to find ways around current technological hurdles and design chips three generations into the future. The final goal is a 7 nm chip, just a third as large as the company's current 22 nm architecture, with stops at 14 nm and 10 nm along the way.
IBM is working on several approaches to achieving its goals of faster, more efficient, and cheaper computing; however, one of the most promising involves the use of graphene.
Graphene: the wonder material
What's so special about graphene, a substance that's found in pencil lead? It's the strongest material ever discovered, highly flexible, and more conductive than copper.
This last property is critical because industry experts are predicting that past 2015 copper wires (which currently connect transistors on a computer chip) won't be able to get any smaller. Graphene, being just one atom thick and highly conductive, promises to allow future optoelectronic chips to achieve performance orders of magnitude superior to current technology and with far lower heat production, energy consumption, and lower cost.
So can you expect breakthrough graphene-based chips in your next laptop? The answer is probably not.
Intel's CEO Brian Krzanich, when asked about the status of his company's progress with graphene chips, said his R&D teams were making good progress on them, however they were still a few generations away and wouldn't be available this decade.
That is not to say that graphene-augmented chips will not be hitting the market sooner. For example, back in January IBM was able to build a graphene-augmented radio frequency chip that was 10,000 times faster than a standard silicon-based chip. The secret? The transistor channels were made of ultra-conductive graphene, laid down in a new manufacturing process in which standard chip components are assembled first then graphene laid down on top (the reverse of the previous method which damaged the graphene).
Competitors to graphene include the carbon nanotube, which IBM believes show better promise as a direct semiconductor replacement to silicon in the sub-7-nm scale era, due to slightly different physical properties. Carbon nanotube-based chips promise to offer five to ten times improved performance over silicon-based ones.
IBM does, however, see a crucial role for graphene in photonic computing, thanks to its unique optical properties.
Scientists at MIT, Columbia University, IBM's T.J. Watson Research Center, and UC Berkley have discovered that graphene can be made completely transparent or opaque to light with minor variances in voltage. This makes possible graphene optical switching relays that can operate as high as 500 GHz (turn on and off 500 billion times per second).
Because graphene can transmit a wide variety of light frequencies (each frequency can be used to store data), graphene optical networks can achieve data transmission speeds in the petabit/second to exabit/second range. This could revolutionize the telecommunications industry with scientists at the University of Bath (in the UK) postulating Internet speed increases of up to 100-fold.
Foolish bottom line
Companies such as Intel and IBM are still optimizing silicon-based chips, which promise to be with us for several more generations. However, as the physical limitations of silicon are reached, innovative chip makers will need to turn to carbon, in forms such as carbon nanotubes and graphene, to keep revolutionizing the computer and telecommunications industries.
