Emerging Technology - CMOSTop J Definition J Advantages J Controversy J Players J Moore’s Law J Conclusion
Emerging Technology - CMOS
Tom Hodson, team of one, for BUS 656
"This is the big one." -- John Kelly, IBM Vice President
What exactly is CMOS? It stands for Complementary Metal Oxide Semiconductor, and it will be coming to an electronic device near you within the next few years. The emerging technology was announced by IBM on September 22, 1997, at the same time an eager GBS student and S.O.B. was searching for an interesting project topic. The new technology was widely hailed in the press. USA Today put the "breakthrough" on its front page. Electronic News decided on the catchy headline, "Copper strikes gold for IBM." The financial markets agreed, boosting IBM’s share price by 5% on the day of the announcement.(1)
So what is all the hype about? To understand the CMOS technology, it is first necessary to understand how a microprocessor chip functions. The foundation for the microprocessor chip is a thin piece of silicon. On the silicon chip resides millions of microscopic transistors. Each transistor acts as switch, assuming either the "on" or "off" position, depending on the voltage. Of course, voltage requires electricity. Electric current is carried between each transistor on what - for the purpose of this high level explanation - can be considered a very tiny wire. Currently, that wire is made of aluminum. IBM announced that it will instead use copper, and the aforementioned hype ensued.
Why is this emerging technology considered so important? After all, the wires in every home are made of copper. Why shouldn’t wires on a microchip be the same? The answer, in short, is because the average home is not made of silicon. When copper is applied to silicon, the tiny atoms pass through the porous surface and damage the chip. IBM made two important discoveries. First, they found (or created, they aren’t exactly telling) a substance that can be applied between the copper and silicon to prevent damage. Second, they designed a manufacturing process to accomplish that task. Thus, we have the complementary metal oxide semiconductor.
The Technological Advantages é Top
The value of using copper over aluminum is not at all related to the cost of raw materials. In such microscopic quantities, the cost is irrelevant. What is important is the size of the chip, and the amount of power it consumes. Compared to aluminum, copper is a superior conductor of electricity. The increased conductivity allows for a smaller amounts of "wire" on the chip, thus reducing its size. In addition, less electricity is lost in the process, so the copper chips will consume less power. IBM predicts performance gains of up to 40 percent and cost reductions of 20% to 30% from the new manufacturing technology.(2) The cost a chip is a relatively small part of the cost of computer, however, so one shouldn’t expect $200 PC’s anytime soon.
The potential applications of copper-based chips are limited only by the imagination. Any end product that currently uses a microchip - from PCs to automobiles to cellular phones - may be using the copper chips in the future. According to IBM:
"Advances in miniaturization allow for computer intelligence to be built into products that are smaller, integrate more complex functions, use less power and require less cooling. Devices such as mobile phones and personal digital assistants (PDAs), as well as computers, may become lighter and more powerful while untold new products may become possible for the first time."(3)
Of course, not everyone agrees with IBM’s version of the story. Dan Hutcheson, president of VLSI Research, expressed an unambiguous opinion on the announcement. "It was such a farce," said Hutcheson, who then noted his admiration for IBM’s public relations staff.(4) According to Hutcheson, several other companies, including Motorola and Intel, already have the copper technology. Will Swearingen, a marketing executive with Motorola, seconded that opinion, noting, "They made the announcement on Monday and their stock jumps six points in the same day. We were there five months ago."(5) Indeed, Sematech, a technology-based research consortium which includes IBM and Motorola, announced a copper-based chip in August.
Is IBM claiming credit for reinventing the wheel? Not necessarily. Remember earlier that the significance of IBM’s discovery could be broken into two parts. The first part involved finding a substance that could serve as a buffer between copper and silicon. That finding, according to the critics, is less-than-Earth-shattering. The second and more important part of the discovery regards the design of a manufacturing process to mass-produce the new copper-based chips. It is here that IBM has leapfrogged the competition. Jon Dahm, a director at Sematech, best described the advantage, saying, "It’s an incredible accomplishment, getting it ready for real, live manufacturing of products for sale."(6)
Generations of microchips are commonly referenced by the thickness of the "wiring" described in a unit of measurement called microns. Intel, which currently manufactures the most advanced chips in the world, recently switched from .35 to .25-micron technology. To get a feel for just how thin that wiring is, a human hair is roughly 100 microns wide. According to IBM, once you get below .25 microns, aluminum just doesn’t work. "Aluminum is running out of gas," says IBM’s Bill O’Leary. "Below .25, aluminum resistance is too high. Aluminum is fighting you."(7) The new CMOS technology would eventually allow IBM to create wiring at the .05-micron level.
Go long on IBM and short Intel, right? Not exactly. Intel responds that aluminum has taken them to the .25-micron level, and is expected to carry them to the .18-micron level. According to Intel, copper will not be necessary until the .13-micron level. Significantly, products based on the .13-micron chip are not expected until 2002. That gives Intel nearly a full four years to figure out how to solve the copper problem - more than enough time for a firm that spends $ xx billion on research and development every year.(8)
There are other industry players affected by the copper technology. Technology equipment manufacturers that supply Intel and IBM with the machines to use in their factories stand to benefit. Among them are Semitool, Novellus Systems, and Applied Materials. In addition, electronic manufacturers such as Sony will use the smaller and faster chips to enhance their product. So too will phone manufacturers like Motorola and Ericsson.
Implications for Moore’s Law é Top
No discussion of microprocessor technology could be considered complete without an obligatory reference to Intel cofounder Gordon Moore and the "law" named after him. In 1965, Moore speculated that the number of transistors that could be fitted on a chip would double every 18 months.(9) As the chart below shows, Moore was right. Moore is also a billionaire now, which should count for something.
When all is said and done, the new CMOS technology will keep the microprocessor industry on pace to meet the expectations of Moore’s law. Considering that those expectations have been met every 18 months for the last 30 years, IBM’s announcement hardly seems like a cause for celebration. But then again, Wall Street is known for its irrational exuberance.
It's not likely that consumers will be clamoring for copper chips anytime soon. End users buy computers and cell phones and Walkmans and Palmpilots. They want products that meet their needs. What happens on the inside of the casing is of little interest if the gadget of choice works as advertised.
Manufacturers that use microchips to produce those products, however, need to make a strategic choice. Is the copper chip a better choice for the laptop or cell-phone manufacturer concerned about battery life? If it is, can IBM continue to deliver in the future as Intel has in the past? Will IBM's lower cost of production translate into higher profits, or a price war that benefits original equipment manufacturers and consumers? A representative from Intel probably had it right when he said, "It's the merits of the products that slug it out in the marketplace." (10)
If the marketplace is indeed the boxing ring, one important point is abundantly clear: this matchup will not live up to the pre-fight hype.