You are here

Is Nanotechnology a Wellspring of Innovation or a Blind Alley?

It’s that time of the year when everyone recaps what occurred during the current year and looks forward to possible developments and trends for the coming year. Dr. Wally Rhines, CEO of Mentor Graphics and industry luminary,  posted an article in EETimes on 11/28/12 looking at previous advances in the semiconductor industry and how, if these trends continue, they might impact our industry looking 40 years out in time. Dr. Rhines brought some amazing metrics to light:

  1. Mature commodity industries like oil, aluminum and cotton in the last ten years had growth rates of only 1.1% to 2.5% while computer unit growth was 11.5% and semiconductor unit growth was 8.4% compounded.
  2. Demand for consumables like automobiles, food and housing is finite while the thirst for information is practically limitless, which in turn drives the demand for more silicon devices and the communications bandwidth to support them.
  3. Shrinking device feature sizes and the ever-lowering transistor costs have provided a significant underpinning to the unit growth in the semiconductor industry. Lower costs translate to empowering more demand as what was once expensive becomes affordable by mass markets.
  4. Continuing current learning curve dynamics would allow the cost per transistor to decline another 99.999994% by 2052 due to the continued influence of Moore’s Law.
  5. Improvements in transistor count due to continuing Moore’s Law would result in a 58x increase in transistor budget available to silicon designers over just the next 10 years. Continued through the 40 year time horizon envisioned in the article the available transistor budget would be truly mind-boggling.

All these metrics to describe the impacts on our industry of continued scaling and the increase in device capabilities due to Moore’s Law and improvements in process technology are amazing and a little bit humbling when you come right down to it. These advances allow richer feature sets and more functionality in the devices we have come to take for granted in our industry and are used to fuel the innovation that has carried us to this point in time.

However, in looking at the metrics that apply directly to the semiconductor industry, transistor costs and increases in transistor budget at every process node, you might say that these improvements were the result of continuing innovation in the industry.  In looking deeper at them, you might also be led to the conclusion the continuing innovation was due to the fact that the invention of the transistor created a ‘basic capability’ that materials scientists and designers then took to new levels of performance and functionality. When we look at the semiconductor industry today what impression does one get?

We see an industry that is built on innovation and the constant improvements in process technology that create feature sets that start out as premium features but later become commonplace through reductions in transistor costs. We also see an industry that uses these improvements to drive existing applications and to create new ones as the right mix of features, functionality and price become available. This business model has generated trillions of dollars over the last 30-40 years, improving the quality of life for most of the people on this planet in some way.

Can we lay the reason for this increase in living standards solely at the feet of the innovators among us?  I don’t think we can attribute this success only to the innovators. I believe the innovators could not innovate if they did not have access to a basic set of capabilities from which to build. The innovators in our industry took these basic capabilities and employed them in innovative ways that were not immediately or readily apparent when they started.  

Many of them were not successful and failed, but a very few of them succeeded and have pushed technologies, companies and markets to their current positions. In looking out at the next ten years, many would say that the industry is maturing and we are reaching a point where we will see a slowing or even an end to scaling for transistor sizes. A smaller number would even say we are reaching an endpoint for innovation or innovated technologies.

I believe nothing could be further from the truth! Thinking about the industry today leads me to believe we are just entering a period of time when we are about to re-encounter a new set of ‘basic capabilities’ that will shape and shake our industry for the next 40 years. I refer to Nanotechnology in the forms of Carbon Nano Tubes (CNTs) and Graphene, just to name two. If you look at the announcements of advances, discoveries, improvements, breakthroughs and just basic research being done in this field, you get the sense that we are, once again, developing a set of basic capabilities that will be used to solve a whole host of problems confronting the semiconductor industry today.

We are at the stage of developing the basics in using these materials to create the point solutions need to keep the industry moving. Because of this, no one can point to any one of these discoveries and say that here lies the solution. Most of these developments are simply lab curiosities today and may never be viable in the commercial marketplace. However, given their sheer number, only a few need to make it to mass production to have a profound impact on our society and industry.

Going back to my initial comments on ‘basic capabilities’ the first thing to point out is that they are just that – basic. No one today can predict which specific problems developments in the nanotech field may produce, but one thing is certain, there will be tangible, useable developments that will come out of these efforts. The second thing to point out about ‘basic capabilities’ is unlike the beginning of the semiconductor industry with the transistor, there is a roadmap for developers and innovators to follow that points us toward specific issues that need to be solved. In fact many of the issues that need to be solved are already known in the industry.

Some are starting to say that transistor costs may stop declining at the 14nm node or even that they will, for the first time in history, actually increase. This may or not be true: we need to get there first before we really know for certain. However, if it proves to be true, we have a potential means to solve the problem by making the silicon we use to create our amazing solutions better in the form of employing nanotechnology to improve the characteristics of the silicon. Or more simply put: If silicon is running out of gas, use better silicon! This will not happen overnight, but at least the industry has a vehicle it can use to bridge the looming performance gap to arrive where we need to be to continue increasing device performance along with reducing cost.

However, all that will be true only if people continue to focus on innovating new solutions and using these solutions to provide answers to the pressing questions of the day. It will take a continuing will to keep investing in basic science and discovery and then developing these technologies that allow the innovators to keep innovating. We need to continue to do this even though a particular approach might turn out to be a blind alley. There will be many blind alleys before there is one breakthrough that is useful. It’s part of the process of discovery and cannot be shortcut. Just look at the following four announcements covered in the mainstream press of basic discoveries to get a sense of how big developments in the nanotech field could be.

  1. Research discovery could revolutionize semiconductor manufacture [Referenced by SemiMD Dec 4th -]
  2. Funneling the Sun’s Energy [Referenced by Low-Power High-Performance on Nov 27th -
  3. Cork the key to unlocking the potential of graphene. [ECN, 12/04/12]
  4. James' bond: A graphene/nanotube hybrid. [ECN, 11/29/12]

As these four example announcements from just the last three weeks show, the work being done in the nanotech field is broad and diverse. There were many, many more to choose from, but these seemed to typify the diversity in possible applications when fully developed. This last announcement is perhaps the best in showing how basic capabilities are being created in the lab today. Now it remains to be seen how innovators can take these capabilities and ‘run’ with them to solve some rather pressing problems in our industry. My money is on the innovators in the long run!