The vital contribution of manufacturing technology
11/01/2001
 |
While new device technology gets the major press and public acclaim, it takes the creativity and innovation of the manufacturing sector to bring new developments into the commercial world. When new announcements of technology advances come along (like carbon nanotube switches), they often include the proviso: "This is five years (or 10 or even 20 years) from the marketplace." The reason, of course, is that the technology first must be shaped into products that buyers will want, and then we have to figure out how to manufacture them at reasonable cost.
Often new technologies are trumpeted to the skies, but they never seem to reach our offices or living rooms. Or they come onto the market but never achieve their promise (like magnetic bubble memories, which couldn't catch up with the cost/performance curves of alternate storage devices).
While new manufacturing methods do not get much notice, their impact can be just as important. In Japan, they refer to "mother technologies," such as machine tools for metalworking and process tools for wafer processing. National programs in Japan have been focused in such areas because they are so critical to the total national economy. Sematech was initially formed in the US in response to Japan's challenge in chipmaking, which started, not surprisingly, with gaining leadership in lithography, the most critical process for IC manufacturing.
Since that era, the industrialized world has recognized that pushing chipmaking technology is a task of such magnitude that international cooperation can help all countries progress faster. As a result, there are many more interchanges and greater cooperation across the globe in moving to larger wafers, automating chip factories, and other critical manufacturing areas, at least in the pre-commercialization phase.
Today, the manufacturing sector is being challenged not just for processes to make ICs but in an array of related fields that can adapt these technologies to other important developments. Process developers are being pressed to come up with a whole array of equipment to enable new cost-effective products, including low-power devices that can operate at high frequencies for wireless and portable equipment, flat-panel video displays that can offer top quality images at costs to challenge CRTs, and microelectromechanical systems for optical switching.
The semiconductor industry has gotten a remarkable run from CMOS devices on silicon wafers with some polysilicon and silicon dioxide and a few added materials such as tantalum and titantium, plus a few dopants. Thirty years ago some thought bipolar would have to be combined with MOS to meet speed requirements, but shrinking features enabled MOS switches to get steadily faster until they pretty much squashed the bipolar market. The shift to complementary MOS gave the added advantage of lower power.
Now another jump in speed is needed for rf devices in cell phones and wireless equipment. Compound semiconductors have served higher frequency requirements so far, although they are more difficult to process and more costly.
With some laboratory R&D work and innovative process technology, however, ways are being found to extend the CMOS process. Since the manufacturing of CMOS has been so refined, and done in such huge volumes, anything that can be produced by a relatively simple extension will have a huge cost advantage. This is already happening with silicon germanium devices for high frequency, led by IBM.
Now Motorola, and process tool companies such as Aixtron, with enhanced molecular beam epitaxy equipment, are showing how compound semiconductor devices can be made right on top of silicon wafers, along with CMOS circuitry. Some trickery is needed to bridge the mismatches in crystal lattices, but it is being done. These advances also have great promise as optical devices for networks, combining optical and electronic functions on the same chips.
Maskmaking has become critical as the industry goes to subwavelength lithography with reticle enhancement techniques, but great advances are being made here as well, with micromirror-based spacial light modulators with gray-scale capability to provide dose control (Micrion), and ultrastable platforms using air-bearings in vacuum (Etec Div. of Applied Materials).
Lasers have potential for a number of new types of process steps — for super-fast thermal processing as well as speeding chemical reactions within very small areas — and companies such as Ultratech Stepper are exploring these possibilities. Atomic layer chemical vapor deposition (ASM, Genus, Applied Materials) is enabling new, very thin gate dielectrics, with high-k materials such as zirconium dioxide.
These are just a few examples of the many tremendous strides being made in the fundamental manufacturing technology that supports the emerging Information Age. These are often the most significant advances, even though they don't get the publicity given to exotic devices based on buckyballs or quantum computing. Let's keep them coming.
Robert Haavind
Editor in Chief