Mr. KASTENMEIER. Thank you very much, Mr. Baumgarten, for your splendid contribution, and we would appreciate it if you would be on hand if possible for comments or some further questions.

Mr. BAUMGARTEN. Thank you, Mr. Chairman, Mr. Edwards.

Mr. KASTENMEIER. At this point, the chair would like to greet its next witnesses, the distinguished president of Intel Corp., representing American Electronics Association, Western Electronics Manufacturers Association, Mr. Andrew Grove.

Mr. Grove will be accompanied by Roger Borovy, who is vice president and general counsel in Intel.

Also, I'm informed that Mr. L. J. Sevin, chairman of Mostek Corp., representing a similar point of view, will also form part of that panel.

Gentlemen, as far as your presentation is concerned, we are in your hands. You may proceed as you wish.

Mr. SEVIN. I am L. J. Sevin, I'm president of the Mostek Corp., and I'd like to point out that I'm from Texas.

Mr. KASTENMEIER. Where in Texas?

Mr. SEVIN. Carrollton, Tex.

We are inserting a slide in the order of presentation. It may come up wrong, we'll change it as it comes up. Mr. Chairman, I have a prepared statement.

Mr. KASTENMEIER. We have your statement. Did you wish to proceed from it?

Mr. SEVIN. Yes, I would.

Mr. KASTENMEIER. Please go forward, sir.

TESTIMONY OF L. J. SEVIN, PRESIDENT, MOSTEK CORP.

Mr. SEVIN. The proposed amendment will amend section 101 of the Copyright Act of 1976 to clarify that copyright protection is available for the imprinted design patterns on semiconductor chips.

At this point, maybe I should comment on the difference between semiconductor and integrated circuits. The semiconductor is merely a description of the material from which integrated circuits are made. All integrated circuits as we know them, and as we're talking here today, are semiconductor integrated circuits; however, all semiconductor circuits are not necessarily integrated circuits, the exception being there may be a semiconductor chip containing one transistor or one diode or one temperature sensing element, something of the like.

The Registrar of Copyrights has denied registration under the present act. She takes the position that these patterns cannot be identified separately from the utilitarian aspects of the chip.

The integrated circuit, a combination of transistors and other electronic circuit elements on a single chip of silicon-silicon being the semiconductor-was invented in the late 1950's. Commercial integrated circuits, known as IC's in our parlance, first were available in 1961.

Mr. KASTENMEIER. At the outset, may I ask you, Mr. Sevin, why the term "semiconductor"; why is it not a full conductor chip? Mr. SEVIN. It means exactly what the term says. The silicon, for example, is neither a perfect, or are they perfect insulators, so they semiconduct. Now, pure silicon does conduct very small amounts of

currents, but it is a pretty good insulator. We have to do things to it to make it conduct electricity, and I hope to be able to clarify some of that as I go through here, and I may or may not succeed. The technology expanded rapidly and in 1971, a then-small Santa Clara company produced an entire computer on a single silicon chip. The "computer-on-a-chip," technically called the microprocessor or microcomputer, is revolutionizing the electronics industry. The February 20, 1979, issue of Time magazine quoted an industry analyst as saying that the microcomputer chip, "Will have more impact on our society in the next 20 years than any other invention." Already, the microcomputer is being used in microwave ovens, refrigerators, electric ranges, cash registers, taxi meters, gas pumps, typewriters, television, et cetera, and by the end of this decade, we expect they will be found in virtually every home and business electronic unit produced.

To meet the 1980 pollution standards, for example, every automobile will have at least one microcomputer.

The integrated circuit was an American development. It was first marketed in 1961, and integrated circuits are already a $5 billion worldwide industry. Only over the last few years has foreign competition become a factor in the state-of-the-art, leading edge portion of the business. The microcomputer started from nothing in 1971. Last year's microcomputer sales were $235 million and are expected to grow 50 percent annually to exceed $800 million by 1981. Continued development of integrated circuit memory chips has reduced the cost of information storage in computers a hundredfold in the last 10 years. In the late 20th and early 21st centuries, integrated circuitry will be as basic to an industrial economy as steel in the 19th and early 20th centuries. Leadership in this technology will be vital to any nation that will be a world leader in economic and military power.

Before I go on, I would like to introduce some of these materials to you. I have a simple glossary of terms. I'll be using the term "wafer." A wafer is a round disk of pure silicon. The substraights that we make our integrated circuits on-here are two examplesone is a wafer prior to being processed, before any integrated circuits have been imprinted, and another is a wafer after the integrated circuit has been printed. These disks are usually 2 to 4 inches in diameter. These happen to be 3 inches in diameter. Another term we'll be using, and have used, is "chip." It's an individual circuit out of a wafer array, after wafer separation. And here's an example of an individual mounted package.

Now, the mask is a basic tool in the manufacture of integrated circuits. Here we have two examples. It's a square section of highly uniform glass or quartz containing an array of patterns that define one of several steps in the manufacture of an integrated circuit. There are several patterns or masks required to define a completed IC. The patterns on the glass are etched into a thin film of chromium or some other metallic substance. Steps in making a mask include the production of a "reticle," which is one of the array patterns, and a somewhat magnification that is a reticle [indicating].

A "master" is a reticle reduced and reproduced many times, and a "working plate" which is virtually the same as the master, is the final mask tool, and it is contact printed from masters.

Now, the term "layout," a layout is a drawing of the patterns contained on a mask, made at several hundred times the actual size of the integrated circuit.

Now, I have a very much expanded blowup of a small portion of an integrated circuit. This is at a different magnification at 100 times final size. It is a completed layout of an integrated circuit. Now, that layout is drawn by computer and it is a mere reproduction of a layout designer's work. I will also explain how that's done, later.

Now, we also have talked about schematic drawings, which are symbolic representation or abstractions of an integrated circuit. Here's an example of a very small portion of an integrated circuit. That is a schematic drawing [indicating].

The actual manufacture of integrated circuits is very difficult to explain and very difficult to understand if you are not thoroughly familiar with chemical processes or not very well versed in physics and chemistry, but I think I can give you a feeling for it, for this whole process in a very few minutes.

First of all, an integrated circuit is a microminiature structure comprised of several key materials. The basic material is highly pure single crystal silicon. The formation of circuit elements consists of selective introduction of controlled amounts of foreign elements into the silicon. We call these impurities, oxidation of the silicon, forming insulation films, and deposition of silicon and metal films to form connecting wires or other operations to introduce other materials to the silicon. All these materials exist in a micro-world that approaches the size of bacteria. This example, this slide, is an actual photograph, produced with a scanning electronic microscope, and it's a photograph of a single transistor in cross section within an integrated circuit. The magnification is about 30,000 times. The dark area at the bottom is pure single crystal silicon. The regions marked "N+ Diffusion" are the selected areas into which controlled amounts of impurities were introduced. The bright white region above is silicon oxide, an insulating layer, the dark regions marked "Poly I" and "Poly II" are deposited silicon films, and the thick topmost gray region is an aluminum film. Now remember, this is one transistor is one transistor, and the magnification is 30,000.

Here is an example of an entire integrated circuit that contains about 5,000 transistors. The magnification now is only about 100; then at a still lower magnification of about four and one-half you see here examples of two different wafers containing different sized circuits.

The patterns on these wafers were introduced through the use of several successive masks in a process very similar to the taking and developing of pictures. The process requires a very clean, particle-free environment containing much complex and expensive processing equipment. Patterns on masks are projected into an emulsion film on the wafers from a light source with a machine called a projection printer. The "picture" of the mask on the wafer is developed through regular dark room developing techniques, the

purpose being to delineate areas in which to etch holes in oxides or to remove unwanted parts of deposited silicon or metal with suitable acids. After each acid etch a long rinse in very pure water is necessary.

To get oxides and silicon films on wafers in the first place, they are put into furnaces at very high temperatures in different gaseous atmospheres. Oxides grow in oxygen, obviously, and silicon films are deposited from the reactions of certain other gases.

Thin aluminum films can be deposited by vaporization of pure aluminum in a vacuum chamber. Here you see wafers being loaded onto a rack prior to being loaded in the vacuum chamber.

And that's basically all there is to it. That, admittedly is a very broad-brush treatment, but I think, as I said, there should be some flavor for the complicated process of producing integrated circuits. Now, I'd like to get to the key part of this presentation, the process of designing the masks. It starts with the circuit design concepts developed by highly skilled and creative engineers, usually electrical engineers, making heavy use of computer analysis and simulation. The product of these circuit design engineers is some form of logic or schematic drawing, examples of which you have before you. The drawing is documentation of the product in abstract or symbolic form.

A layout designer takes over at this point and turns an abstraction into an engineering drawing of a mask. Here is a small section of a mask layout [indicating]. Now, this is done mostly by hand, and it involves much trial and error and is one of the most difficult and time-consuming parts of the development. The largest drawings are usually done at a magnification of 500 to 1,000 times and have to be done in sections because of the practical limitation of drafting tables. Just how small a part of an integrated circuit that is will become apparent to you in a later slide.

These drawings also have to be dimensionally quite accurate so that the drawing is done on a very stable mylar film. There is an example of mylar films that Mr. Borovoy will show you.

Special computer graphic software provides for the entry of the drawing-I'm sorry-here's a layout designer at work and a closeup shot of the layout designer's work.

Special computer graphics software provides for the entry of the drawing into a computer through the use of a machine called a digitizer. The digitizer has a specially constructed table having a grid of many fine wires hidden under the surface. The drawing is converted into a series of points against some common reference. An electromagnetic coil with cross hairs communicates with the hidden grid wires to pinpoint exact locations.

The drawing can be displayed on a TV screen for editing purposes. The product of the digitizer is a magnetic tape which can be used to drive a computer controlled drafting machine for engineering checks. The small drawing I passed out is a computer controlled drawing, and it is simply nothing more than a reproduction of an original mask drawing at a different magnification.

Finally, a machine called a pattern generator which reproduces the mask in a thin emulsion film on a glass or quartz plate in a photographic dark room. This will become the reticle that I intro

duced earlier. A few more similar steps will result in the working masks with arrays of circuits on them.

Now, why is this layout important? The "chips" we have been discussing are becoming larger, more complex and increasingly expensive to develop. Computer memory chips in 1971 were approximately one-tenth of an inch on each side and contained about 2,000 transistors. The state-of-the-art memory chips being developed in 1979 measure about 5/32 by 1/4 inch and contain approximately 70,000 transistors. A microcomputer chip of 1971 also contained about 2,000 transistors, while the state-of-the-art microcomputer chip of 1979 contains over 30,000 transistors. It is more powerful than the IBM 1401, which was the workhorse business computer of the 1960's. The chip layout design on each of these circuits almost fills a room 20 feet square. Five designer years were required for the memory and 10 designer years were required to transform the microcomputer into a chip layout. In either case this time is by far the longest part of the development cycle of new products, longer than the engineering time that went into conceptualizing the product, and much longer than the time taken by the routine computer operations necessary to produce a mask.

Who are layout designers? Layout designers are creative persons and not just draftsmen. They must have some training in electronic circuitry, usually they are electronic technicians. They must have a strong ability to visualize from abstraction and must be able to plan ahead mentally much as does a good chess player. The designer must be able to cram 70,000 or more transistors and their intricate rabbit-warren connections into an absolutely minimum area in order to minimize the chip size because that is directly related to cost of the product. Layout design is a skill that has successfully resisted 12 years of attempts at computerization. It requires a level of human ingenuity that will not be computerized for at least another 25 years, in my opinion, maybe longer-maybe

never.

At the same time as the cost of creating new layout designs is skyrocketing, the technology for copying them is improving. Ironically, electronic equipment making use of these very microcomputer chips is being used by chip pirates to copy them. Better lenses, better photomicrographic techniques, better control electronics are becoming available for taking superb blowup pictures of the tiny chip. As soon as the company which did the original design puts a chip on the market, the chip pirate purchases it, removes any impeding coatings on the chip surface, and sends it to a photomicrographics specialist to make a blowup photo of the layout design. Typically, a blowup 800 to 1,000 times the original chip size is used. The chip pirate, or a commercial business which offers chip copying as a service, electronically traces the photographic blowup and feed the design information into a computer in exactly the same way as the original digitizer did from the original layout drawings. The techniques and equipment are exactly the

same.

To demonstrate that chip pirates do exist, I have a slide of one of my company's products. It's a 16,000 bit random access memory. I'd like to do one thing on this slide: There are a couple of small rectangular areas on either corner of the chip here. I don't know if