John - Re: "Moore's Law"
This "Law", taken from an observation Dr. Moore made in, I believe, 1965, has more or less characterized the semiconductor leading edge for about 30 years.
Every time the boundaries of physics have appeared to limit the application of this law, the "boundaries" have been pushed back.
Lately, the COST to push back these boundaries has escalated faster than the national debt! Thus, a new economic factor must now be addressed. Several people, Intel included, have mentioned $10 BILLION as the cost of a wafer fab to produce 0.18 micron devices. Thus, the laws of physics may be hiding behind the laws of economics.
Currently, the path to 0.18 micron production processes will require photolithography capabilities that do not exist today (except for a small R & D capability). Intel is working with a few institutions to help develop DEEP UV (laser activated) capability for this process - and it will be expensive, ASSUMING that the development will proceed on schedule.
This lithographic capability will be required to produce silicon gates (and other features) with 0.18 micron dimensions, thereby permitting smaller transistors, and more of them, in a given area of silicon.
Couple this with the transition to 300 mm wafers (approximately 12 inch diameter), and you're talking about expensive equipment - and LOTS of it.
So, the physics guys and the financial guys are both going to have to work together to keep the semiconductor industry on the same growth path (Moore's Law) that it has been on since 1947 or 1948.
This road block (temporary, we all hope) will separate the "haves" from the "have nots".
From a physics limitation standpoint, 0.1 micron processes appear to be "doable", assuming that the equipment to produce these features are available, and that a given compnay can afford to buy them. Scaling of devices (i.e., reducing their sizes), reduced doping levels, thinner gate oxides, more metal levels etc. will be required and they will also require reduced voltage levels. The new devices will have supply voltages of 2.5 volts or less - probably less. (This will assist power dissipation issues).
All it will take, for the foreseeable future, is time, effort and money - and lots of each!
Paul |
