Posts Tagged ‘Moor’s Law’

What About Moor’s Law and Computer Chips?

In the electronics industry people sometime quote Moore’s law as if it was a solid principle of atomic physics.
Well it’s not really about atomic physics, rather it is about the growing rate of electronic chip technology.

Actually, all areas of technology are growing at a rate faster than what you might imagine. Certain parts of modern technology have not developed as fast as they should have for various reasons. One factor being the marketplace and the presence of large commercial interests who do not want radical change in technology.

Any kind of rapid sequential growth might be termed a form of exponential or geometric growth. This is similar to having a bank account where the interest you receive goes back into the account and added to the principal. Doing so, it’s possible to double your original investment within a number of years. Depending on the interest rate, you might double your investment in a period of about 15 years or so. Most people understand how that works.

Now what made Moore’s law so noteworthy is that he proposed the doubling would occur in just 18 months. Of course, that kind of growth does catch your attention and it tends to disrupt whatever technology already existed. Often times the idea is applied to lithography as used to produce computer chips. In this case lithography means a photographic process where a silicon wafer can be etched and exposed to different materials to build transistors on the basic silicon chip. This is a form of photography where an intense light shines through in negative image that is focused down to the silicon wafer. Thus it is possible to make transistors and interconnections that are extremely small.

Electronic logic does not require any specific size, at least at the moment. Even transistors made extremely small are still functional and perform logic. So the whole idea was traduce everything in size to reduce the costs and reduce power consumption.

A benefit from smaller feature size is higher-speed. So as to improve the yield from the silicon foundry, you also increase the speed of the chips and reduce the  cost.

Now then, why are some saying that Moore’s law is coming to its end? Well, there is a principle called diminishing returns. This is usually cited in economic articles. But the idea could apply to almost any area where there is rapid growth followed by a leveling off.

Another example is an oil well.  Somebody drills an l oil well and goes through a lot of effort to try to get the oil well to yield. Then one day the oil will starts improving its yield, more and more oil comes out faster and faster. It is on a geometric expansion. Sad to say, oil wells to eventually dry up and the period of expansion terminates and then the oil well produces less and less until eventually it is not profitable.

Now back to the subject of electronic chips and silicon foundries and lithography. There is a limit as to how small you can make an etching on a silicon chip using conventional photographic techniques. You get to a point where the light wave used must be extremely short and becomes difficult to focus. When you get to that point you start running out of tricks and techniques to make the etchings smaller and smaller.

So what about going to x-ray technology and the use of electron technology to make etchings? Well, it’s not really the same as optical light. Try to control and electron beam and focusing it into a small area and etch something on a silicon surface is not quite as easy as it is with photographic techniques. In fact, the electron voltages can be so high that they become destructive and that’s not really what you want. So it would seem that there is an up word limit as to how far you can go with electronic technology as far as making computer chips smaller and smaller.

Not necessarily so. There are other ways to make things smaller and smaller that have not been fully exploited. For one thing, there is the concept of building things in a three-dimensional array instead of using a two-dimensional surface to etch electronic circuits. True 3-D technology is not yet with this when it comes to making computer chips. But if some new breakthrough should come that makes it possible to build computer chips in three dimensions, then Moore’s law may continue.

So what is the actual limit of how small you can make a transistor ? Some researchers believe it’s possible to make an effective transistor gate down at the molecular level. Thus a transistor might actually be a few molecules tied together. As for me, that’s beyond my ability to comprehend. So I am not qualified to say that it’s impossible.

Anyhow, I don’t believe that Moore’s law has it come to its limit if you take a broader view of what it means. The broader view is it means we can keep on expanding our technology in the foreseeable future.

What do you think?