Moore's Law at 50: how his predictions have shaped the future ~ lotfi

Introduction and impact of Moore's Law

It's 50 years since Gordon E Moore, co-founder of the Intel Corporation, made the observation that became known as Moore's Law. In 1965 Electronics magazine had asked him to write an article predicting what would happen in the semiconductor component industry in the subsequent 10 years. Moore was, at that time, director of R&D at Fairchild Semiconductor, and this made him something of an expert in the field.

Moore looked at the elements – transistors, resistors, capacitors, and diodes – being used in chips at the time (approximately 60), and based on their use in the preceding years, came to the conclusion that the industry would double these elements every year for 10 years until they hit 60,000 per chip.

Moore lays down the law

Ten years later and Moore's prediction proved very accurate, leading a colleague to coin the term 'Moore's Law', but at this time Moore revised his prediction to a doubling every two years. Ultimately transistors came to be the dominant element in chips, becoming the most useful measure of an integrated circuit's complexity.

But Moore's Law wasn't just about the quantity of elements and a chip's resulting performance – Moore was also concerned with economics. His original prediction was based upon the number of elements within each chip where cost per component was at a minimum. Interestingly, in the past ten years, increases in transistor numbers have come to be more about cost than performance, with transistors being made smaller in order to keep costs down – although this further miniaturisation has resulted in performance gains in any case. Moore's Law economics in action!

Although Moore's first and second predictions were, initially, a means of chronicling the industry's progress, over time Moore's Law became something of a driving force, encouraging semiconductor manufacturers to keep pace with the Law. Today, there are billions of transistors on chips, and this magnitude has a great deal to do with the existence of Moore's Law. It is said that the semiconductor industry still uses it to guide its planning and to set targets for R&D.

Major impact

This being the case, Moore's Law's impact on our lives and the progress of business and industry cannot be overstated. The way we communicate has changed irrevocably over the past few decades. If Moore's Law hadn't been adopted by the semiconductor industry as a call to arms, would we be working on our own individual computers, making business calls on smartphones or travelling to meetings in today's computer-controlled cars (if we bother to travel at all – videoconferencing has never been more sophisticated)? Unlikely.

And there'd almost certainly be no internet without Moore's Law. Gordon Moore has helped to determine our technological reality and is arguably even more (no pun intended) influential in indirectly shaping our futures than Arthur C Clarke – and that's some achievement.

So, what helped Moore's Law to gather momentum in the early years following the 1965 publication of Electronics magazine? Of course, the invention of the integrated circuit, which instigated and influenced Moore's article has a huge part to play – without it there would be no Moore's Law, I'd be typing this piece on a typewriter and TechRadar Pro would be a print magazine. We have Jack Kirby at Texas Instruments and Robert Noyce at Fairchild Semiconductor to thank for the 'birth' of Moore's Law and for keeping it alive ever since. But there were other contributions to the Law's early development.

Super shrinkage

The invention of DRAM (Dynamic Random Access Memory) in 1967, by Robert Dennard at IBM, made it possible to fabricate single-transistor memory cells. The invention of the excimer laser at the Lebedev Physical Institute in 1970, and the subsequent invention of deep UV excimer laser photolithography by Kanti Jain at IBM in the early 80s, ultimately led to the smallest components in integrated circuits to shrink from 800 nanometers in 1990 to a low of 22 nanometers in 2012.

The invention of flash memory, also in the early 80s, by Fujio Masuoka at Toshiba opened the doors for high-capacity, low-cost memory. And manufacturing costs for chips were driven down thanks to developments in CMP (chemical mechanical planarisation) by IBM and Motorola throughout the 90s. CMP smoothes the surface of chips making them easier and cheaper to manufacture.

More recently, various announcements have made the future or Moore's Law – which Moore himself reckons has a limited shelf life – look pretty secure. Researchers at the Tyndall National Institute in Ireland announced that they had fabricated the junctionless transistor in 2010; researchers at the University of Pittsburgh announced the development of the single-electron transistor, 1.5 nanometres in diameter in 2011; and in 2012 a team at the University of New South Wales announced the development of a transistor comprising a single atom placed in a silicon crystal.

Intel Wafer

Will it last?

How long can Moore's Law go on for? As mentioned above, Moore reckons there'll be an end to it. "It can't continue forever," he told TechWorld.com in 2010. Moore says it'll be "two or three generations" before transistors are the size of atoms and that we'll reach a "fundamental limit" with existing processes before 2030.

Also in 2010, the International Technology Roadmap for Semiconductors predicted that transistor-per-chip growth would slow by 2014, projecting that the Moore's Law doubling would shift from every two years to every three. Of course, the growth of nanotechnology could restore Moore's Law to its doubling-every-two-year predictor.

However it might be sustained, as it did before, Moore's Law will have to evolve to survive. As we've seen, developments in semiconductor technology will continue to occur, but this won't necessarily mean that costs will continue to fall – Moore's original economies.

Other measures of the Law will have to come into play, meaning that it will morph into what some in the semiconductor industry call "more than Moore" and 'Gentleman Scientist' Chris Mack calls Moore's Law 3.0. Mack cites the cell phone camera as an example of 3.0 in action. These cameras incorporate image sensors directly onto digital signal processors using large vertical lines of copper wiring called through-silicon vias, so uniting non-logic functions that would previously have been kept separate from the chips themselves.

Assuming Moore's Law continues to have an impact on technology, what might the future look like? However you want it to look, basically, especially if nanotechnology becomes more influential. Cyber body parts, brain implants, wearable tech that interacts with your body biologically – our lives are set to change again and again thanks to Moore's Law as we keep living in what feels like the future. And as Moore's Law evolves to keep up with the technological evolution, there are no limits to where the semiconductor will take us. Here's to the next 50 years.