ChipWatch: The End Of CMOS And It's Effect On The World
[Guest post by Matt Grimshaw]
By Matt Grimshaw, Editorial Director of the Semiconductor Technical Journal Future Fab International.
In continuation of the topics I started three weeks ago this is the last part in the trilogy, but also the most controversial. If I don't get stoned to death by a raging mob of PhD's then this may well prove to be the transistor gate that broke the camels back.
I'll be the first one to admit that this one is a bit of a stretch as an issue for the next down cycle as the end of the Complimentary Metal/Oxide Semiconductor more commonly referred to as CMOS is probably 15-20 years out, and beyond that in mainstream markets with lower priced devices. The technology is robust and mature and will be around for the foreseeable future.
The reason for its demise is the reason for the success of the semiconductor industry - the oft mentioned Moore's Law. For decades the law has held true but now we are reaching the physical limits of what the technology is capable. Cutting edge chips have features that are 32nm or below - to put that in perspective a strand of DNA is just 12nm in width and it will not be long before even that is surpassed.
Simply put the industry is hitting the physical limitations of the materials and structures it's been using and must now find another way to make progress. This is something that causes a fair bit of concern for everyone in the industry. Why you might ask? because it's a change of such magnitude it may well alter the world's economy, and beyond that, the fabric of society itself.
The basis for this extrapolation is founded in the fact that since the Integrated Circuit (IC) became prevalent in modern society it became the foundation of a technological revolution that is more commonly known as the information age although I'm sure some parents out there would call it the age of too much information. Everything is now getting smart - the Internet is the depository for more information than it's possible to absorb and is fast evolving into the foundation for many aspects of our way of life.
However, if you think beyond the Internet, and PCs and begin to think of how much of our modern surroundings are solely reliant on semiconductor technology - TV's aeroplanes, computers, satellites, cell phones, cars, infrastructure - even the stock markets - you begin to get an idea of how big a shift it will be to move beyond what is known.
All industries are touched and made possible in their modern forms by CMOS based devices. The switch is similar in magnitude to switching from an oil based economy to one based on hydrogen - indeed fuel cells and electric cars would not be possible without IC's either.
Our society is very reliant on not just the technology but the economics of Moore's Law, a cycle of exponential growth in functionality tied to an ever decreasing cost per function. In other words, as devices become more complex and powerful, their relative cost comes down, allowing more people to afford them and hence swelling the market size.
Beyond CMOS not enough is known to predict if Moore's Law or one similar will continue to drive the economy of modern society in the same manner, and this is the basis for the concern. We have all come to expect that the latest gadget will be twice as fast, cost half as much and will reshape how we operate...think of the world before cell phones, or laptops or GPS and you get the drift.
One of the major factors is that we cannot see what could be possible with some of the replacements for CMOS as they are fairly outlandish to the uninitiated. Google "Spintronics" and you'll see what I mean.
For now, CMOS will continue its march towards making itself obsolete but sooner or later decisions will have to be made as to what will succeed it...keep an eye out for the news as it will have an impact on you. In a future post I'll go into what could step into the void.