Chapter 5 Electronics Accuwave Corporation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .50 Expanding the Number of Light Signals in an Optical Fiber AstroPower, Inc. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .52 Manufacturing Technology for High-Performance Optoelectronic Devices Cree Research, Inc. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .55 Processes for Growing Large, Single Silicon Carbide Crystals Cynosure, Inc. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .58 Harnessing Cheap Diode Lasers to Power a Low-Cost Surgical Laser Diamond Semiconductor Group, LLC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .60 Lowering the Cost and Improving the Quality of Computer Chips FSI International, Inc. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .63 A Gas Method to "Dry" Clean Computer-Chip Wafers Galileo Corporation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .65 Low-Cost Night-Vision Technology Hampshire Instruments, Inc. (Joint Venture) . . . . . . . . . . . . . . . . . . . . . . . . .67 Large-Scale Diode-Array Laser Technology for X-Ray Lithography Illinois Superconductor Corporation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .69 Using High-Temperature Superconductivity to Improve Cellular Phone Transmission Light Age, Inc. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .72 Exploiting Alexandrite's Unique Properties for a Less-Expensive, More-Reliable Tunable Laser Lucent Technologies Inc. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .75 Precision Mirrors for Advanced Lithography Multi-Film Venture (Joint Venture) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .78 Joining Several Chips Into One Complex Integrated Circuit Nonvolatile Electronics, Inc. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .80 Computer RAM Chips That Hold Memory When Power Is Off Spire Corporation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .83 A Feedback-Controlled, Metallo-Organic Chemical Vapor Deposition Reactor Thomas Electronics, Inc. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .85 Flat Fluorescent Lamps for Displays 49 Electronics Accuwave Corporation Expanding the Number of Light Signals in an Optical Fiber O ver the last two decades, the use of optical fiber as an alternative to metal wire and cable has exploded. Optical fiber is now the technology of choice for use undersea and for most terrestrial applications of more than the shortest distances. More Light Signals Per Optical Fiber This ATP project with Accuwave Corporation, a small California company specializing in the development of holographic communications systems, created a way to substantially increase the number of signals that can be transmitted in a single strand of fiber-optic cable. The new technology is designed to enable the transmission of 80 or more channels per fiber. If adopted, it could eventually reduce the cost per transmission and save hundreds of millions of dollars over a period of just a few years. A Unique Holography-Based Approach The new technology is based on the concept of wavelength division multiplexing (WDM), which transmits light of more than one wavelength through a single optical fiber, separating the individual wavelengths at the receiver. Such systems must discriminate among the different wavelengths and so are limited by the accuracy of the multiplexing and demultiplexing optics. Accuwave had previously developed a unique approach to WDM using volume holography: holograms "written" in the interior of thick crystals of photorefractive (light-bending) materials. In the demultiplexer crystal, for example, multiwavelength light enters one end . . . multiwavelength light enters one end of the crystal and encounters a series of holographic gratings . . . that separate the light signals of different wavelengths.
of the crystal and encounters a series of holographic gratings -- each tuned to deflect a specific wavelength of light -- that separate the light signals of different wavelengths. Accuwave had demonstrated the individual elements of a system that could multiplex wavelengths more than 10 times better than the current state of the art at visible wavelengths. With its ATP funding, Accuwave extended its technology to the infrared wavelengths used for long-distance telecommunications, and designed a prototype WDM system. Accuwave officials report that ATP funding enabled it to develop WDM for signal transmission, a task it would otherwise have been unable to do. In addition, receiving the ATP award helped the company form important alliances with research partners during the ATP project (not identified here for confidentiality reasons). lengths. Nonetheless, Accuwave continued to work toward completion of its WDM multiplexer, which it believes provides multiplexing capabilities of higher signal accuracy, with more channels per fiber and in a smaller package than the products offered by competitors. Though Accuwave did not succeed in its original commercialization plans for sale of a WDM system in the bulk-signal-transmission market, it launched several component products based on the ATP-funded technology. These include wavelength controllers, wavelength lockers and fiber-optic collimators, all of which are being sold to producers of WDM systems. The company developed contacts with potential telecommunications clients in Europe, Japan and Brazil, as well as the United States, and it planned to introduce its own wavelength multiplexers in the near future. Potential Big Savings in Telecommunications With its potential to increase the number of signals that a single optical fiber can carry, the Accuwave technology could significantly affect the cost of communications via fiber-optic cables, particularly if used for undersea applications. Because of the volume of messages transmitted via this medium, cost savings would be great, even if the number of signals per fiber were doubled. The Accuwave technology has the potential to double and redouble the number of signals per fiber many times Marketing Disappointment Spurs Alternative Commercialization Near the end of the ATP funding period, while Accuwave was trying to raise additional private capital to complete the technical work on its WDM system and sign commercialization agreements with potential customers, another company beat it to market with a competing system operating in the same infrared wave- . . . another company beat it to market with a competing system operating in the same infrared wavelengths. 50 A D V A N C E D T E C H N O L O G Y P R O G R A M PROJECT: To develop holographic-optics technology that will increase (by more than 10 times) the number of signals that can be transmitted through a single optical fiber. Duration: 3/1/1993 -- 3/14/1995 ATP number: 92-01-0055 FUNDING (IN THOUSANDS): ATP $1,987 Company 898 Total $2,885 COMMERCIALIZATION STATUS: In 1996 and 1997, Accuwave introduced three WDM system components: wavelength controllers, wavelength lockers and fiber-optic collimators. The company continued to pursue its original goal of selling WDM products for fiber optics telecommunications applications. OUTLOOK: Despite the heretofore promising prospects for growing applications of this technology in the telecommunications sectors, the commercialization outlook at this time is bleak. As this report was going to press in late 1998, it was learned that the company had ceased operations and was in the process of declaring bankruptcy. While it is possible that the technology will be picked up by other companies and carried forward in the future, at this point there is insufficient information about the likelihood of this to comment further on the outlook. COMPANY: Accuwave Corporation 1651 19th St. Santa Monica, CA 90404 Contact: Neven Karlovac Phone: (310) 449-5540 Number of employees: 5 at project start, 16 at the end of 1997 69% 31% ACCOMPLISHMENTS: Accuwave developed a process for producing photorefractive materials suitable for fiber optics telecommunications applications. The company also:
s received two patents for technologies related to the ATP project: "Photorefractive Systems and Methods (Divisional)" (No. 5,684,611: filed 6/6/1995, granted 11/4/1997) and "Wavelength Stabilized Laser Sources Using Feedback From Volume Holograms" (No. 5,691,989: filed 9/14/1993, granted 11/25/1997); . . . launched several component products based on the ATP-funded technology. s applied for two additional patents for technologies related to the ATP project; s completed pilot production of wavelength division multiplexing (WDM) components designed for incorporation into equipment manufactured by other companies, and introduced the components in 1996; s signed a purchase agreement with a major telecommunications equipment manufacturer; s raised $4 million from venture capitalists and other investors since 1990; and s . . . in late 1998 it was learned that the company had ceased operations and was in the process of declaring bankruptcy. built a plant and ramped up volume production in 1998. over, with the count possibly reaching as many as 80 signals per fiber. In addition to applications in the bulksignal-transmission market, the ATP-funded technology has the potential of providing greater cable bandwidth to homes and offices for use with high-definition TV and to the closed-circuit TV market, particularly for security uses. The company was interested in pursuing these potential applications, but instead used its resources to develop the WDM system for telecommunications applications. The technology also has potential applications in ultranarrow band filters, spectrometers and optical disk memories. As this report was going to press in late 1998, it was learned that the company had ceased operations and was in the process of declaring bankruptcy. It is possible that the technology will be picked up by other companies and carried forward in the future. 51 Electronics AstroPower, Inc. Manufacturing Technology for High-Performance Optoelectronic Devices O ptoelectronic devices -- from light-emitting diodes (LEDs) and solar cells to lasers and detectors -- are abundant in everyday life. Millions of LEDs are used in automobile dashboards and consumer electronic products (clocks, radios, VCRs, CD players, coffee brewers and microwave ovens), as well as in commercial and industrial products such as fax machines, copiers and printers. LEDs That are Four Times as Bright Although LEDs are used in many applications where digital readout is needed, they have limitations. They do not emit much light, so they cannot be seen at a distance. If they produced really bright light, LEDs would be even more widely used than they are already. This ATP project with AstroPower, a small Delaware company incorporated in 1989, developed a new approach to production-scale liquid-phase epitaxy (LPE). The company has fabricated LEDs in a way that significantly increases, by a factor of four, the brightness of the light they emit. A large area solar grade silicon sheet emerging from a silicon growth reactor which incorporates new ATP-funded technology. . . . made significant advances in understanding growth processes for compound semiconductor materials . . . A New Approach LPE is a widely used technique that involves melting a semiconductor material and letting it crystallize on a substrate. AstroPower's novel enhancement, the first technical goal of the project, involved the use of a thermal gradient that promotes the growth of the epitaxial layer laterally much faster than vertically from the substrate. Company researchers made significant advances in understanding growth processes for compound semiconductor materials and in applying LPE to lateral growth over buried reflectors and other components. The technology can be used for volume production of low-cost compound semiconductor devices -- those made from a compound of elements, such as gallium arsenide, rather than a single element. . . . succeeded in designing and assembling a modular prototype production growth system . . .
AstroPower's second technical goal was to develop the technology to automate the new LPE growth process in integrated factory-scale fabrication equipment. Company researchers succeeded in designing and assembling a modular prototype production growth system that has already significantly shortened production scale-up times for currently fabricated products, as well as for potential products under consideration by customers. 52 A D V A N C E D T E C H N O L O G Y P R O G R A M PROJECT: To develop new crystal growth methods and high-throughput manufacturing technology for fabricating light detectors and emitters with integrated reflecting mirrors. Duration : 7/15/1992 to 7/14/1995 ATP number: 91-01-0142 FUNDING (IN THOUSANDS): ATP $1,423 Company 1,580 Total $3,003 tors), achieving a fourfold increase in brightness;
s completed scale-up of liquid-phase epitaxy (LPE)-growth technology to a high-throughput, production-scale process; s significantly shortened production scale-up times for specific products, compared with previous manufacturing processes; constructed a demonstration production facility to implement the technology; and conducted an initial public offering of stock in February 1998, raising $16.7 million. COMMERCIALIZATION STATUS: Direct commercialization of ultrabright red LEDs, a proposed initial goal of the project, did not occur, mainly due to economic and market developments. Knowledge of new crystal growth methods acquired during this project contributed, however, to the enhancement of fabrication methods for the company's Silicon-FilmTM solar cell and for other compound semiconductor devices. OUTLOOK: AstroPower has applied the ATP-funded crystal growth technology to its current manufacturing processes, improving productivity and lowering costs. It also plans to use the technology for several breakthrough devices when appropriate market size has been achieved; if such markets develop substantially, the outlook is promising. Two significant products that are nearing introduction are combustion sensors based on gallium-phosphorus compounds, and avalanche photodiodes and detectors based on indiumgallium-arsenic-antimony compounds. COMPANY: AstroPower, Inc. Solar Park, 461 Wyoming Road Newark, DE 19716-2000 Contact: James B. McNeely Phone: (302) 366-0400 Number of employees: 86 at project start; 160 at the end of 1997 s 47% 53% . . . the company's product lines have all grown rapidly in recent years, with much of the growth attributed to knowledge developed in the ATP-funded project.
product lines have all grown rapidly in recent years, and they attribute much of the growth to the ATP project. All of AstroPower's compound semiconductor-based products incorporate epitaxial growth in their fabrication. This includes their flagship product, the SiliconFilmTM solar cell. Silicon-FilmTM is a continuous production process to manufacture crystalline silicon sheets and layers. s ACCOMPLISHMENTS: The company achieved the goals of the ATP project: developing new epitaxial growth methods, as well as new processes for plant-scale industrial production operations. Evidence of the company's achievements are that it:
s received four patents related to the ATP project technology; "Columnar-Grained Polycrystalline Solar Cell and Process of Manufacture" (No. 5,336,335: filed 10/9/1992, granted 8/9/1994) "Hetero-Epitaxial Growth of Non-Lattice Matched Semiconductors" (No. 5,356,509: filed 10/16/1992, granted 10/18/1994) "Columnar-Grained Polycrystalline Solar Cell and Process of Manufacture" (No. 5,496,416: filed 8/5/1994, granted 3/5/1996) "Semiconductor Device Structures Incorporating "Buried" Mirrors and/or "Buried" Metal Electrodes" (No. 5,828,088: filed 9/5/1996, granted 10/27/1998) s demonstrated the application of the new epitaxial production technology to optoelectronic device structures that have integrated reflecting mirrors for enhancing light output (an ultrabright light-emitting diode (LED) with buried reflec- Market Developments Upset Initial Commercialization Plans Commercialization of the enhanced compound semiconductor devices in high volumes has not yet happened. An initial goal, to produce high volumes of red LEDs, has been stymied by market developments. The Japanese have come to dominate the market for red LEDs, which have become a commodity product. Although AstroPower has a technical advantage in producing the devices, the value of this market to the company is quite small, since the cost of entering the market is too high to make such a venture profitable. Use of the Technology for Current Product Lines Knowledge developed in the ATP-funded project, especially advances in understanding epitaxy technology, has proven useful across all company production activities, AstroPower officials say. They report that the company's Shortened Production Scale-Up Times The success of the ATP-funded project ensures that new and innovative optoelectronic devices will have significantly shorter production scale-up times than were possible before the project. The establishment of a technology that permits low-cost, high-throughput synthesis of compound semiconductor structures is potentially useful for many optoelectronic device products. It can be used, for example, in making specialty devices on a job-order basis using gallium arsenide, gallium arsenide-on-silicon, indium phosphorus and a host of other unexplored alloys. These devices are used in the fabrication of common products like detectors, solar cells, sensors and light-emitting products. The new technology can also be used in the production of highly sophisticated devices such as vertical cavity surface emitting lasers and resonant optical cavity detectors with back reflectors. AstroPower intends to incorporate this technology in a number of breakthrough devices that it can produce in sufficiently large quantities when appropriate market size has been achieved. Two significant applications are nearing product introduction. The first is combustion sensors, based on gallium phosphorus compounds, that can be used for flame control in internal combustion engines and utility burners. The second is avalanche photo- 53 diodes and detectors, based on indium-gallium-arsenic-antimony and indium-arsenicantimony-phosphide compounds, that can be used for light direction and range instruments, collision avoidance, atmospheric gas measurements, weather prediction, spectroscopy, blood gas analysis and noninvasive medical analysis. These two products are currently in pilot production and are being tested by NASA, the Air Force and industrial companies. Cross-sectional photomicrograph of a light emitting diode showing device active layers and buried mirror overgrowth. An initial goal, to produce high volumes of red LEDs, has been stymied by market developments . . . red LEDs have become a commodity product.
Company Growth At the beginning of the ATP project in 1992, AstroPower had annual product sales of $1 million. By 1997, sales had grown to $16 million. And in February 1998, AstroPower successfully conducted an initial public offering of stock, raising $16.7 million. AstroPower is convinced that had it not conducted the ATP-funded project, its growth experience (as measured by product sales) would have been set back by three years, the length of the ATP project. This belief is based on the use of improved epitaxial growth technology across all of its product lines, its application of manufacturing automation processes to all of its manufacturing operations, and to the overgrowth of semiconductor materials on dissimilar substrates as well as on mirrors, insulators, and conducting planes. Without the ATP funds, AstroPower says it would not have carried out the project. Potential Large Economywide Benefits AstroPower noted at the beginning of its ATP project in 1992 that it expected in a project like this that products might take as long as 10 years to move from initial technology development to new product sales. The demonstration production facility AstroPower developed is capable of producing millions of LEDs or other LPE-based optoelectronic devices per month. When sufficient demand for the new products emerges, AstroPower plans to construct an optoelectronic semiconductor chip-manufacturing facility for new products made possible by the innovative LPE-growth technology. Benefits are already accruing to purchasers of the company's solar cells, which have higher quality and cost less than they did before the ATP project. If the company succeeds in bringing to market additional products that use the new technology, even more benefits will accrue to its customers. Because of substantial uncertainty about these events, it is too speculative at this time to try to predict the magnitude of these future benefits. Benefits are already accruing to purchasers of the company's solar cells, which have higher quality and cost less than they did before the ATP project. . . . had it not conducted the ATP-funded project, its growth experience . . . would have been set back by three years.
54 A D V A N C E D T E C H N O L O G Y P R O G R A M Electronics Cree Research, Inc. Processes for Growing Large, Single Silicon Carbide Crystals M ost computer chips today consist of tiny electrical and electronic components on a thin slice of silicon crystal. As many as 5 million discrete components can be placed on a piece of crystal less than 2 inches square. Silicon crystal chips, however, are quite sensitive to heat. Electricity passing through a chip's super-thin connecting wires creates heat, just as it does in the heating element of a toaster. If too much heat builds up, the chip loses its functionality. Cree's LED chips are used by Siemens A.G. for back lighting for this dashboard. Beating the Heat in Electronic Devices This ATP project with Cree Research, a small company in North Carolina's Research Triangle Park, made significant progress in the development of an alternative raw material for making crystal slices -- silicon carbide. This material belongs to a class of semiconductors having "wide bandgap," which means they are relatively insensitive to increased temperatures. Silicon carbide's thermal conductivity is greater than that of copper, so it rapidly dissipates heat. It is impervious to most chemicals and highly resistant to radiation. Silicon carbide is extremely hard -- it is used as grit in common sandpaper -- indicating that devices made with the substance can operate under extreme pressure. It also possesses high field strength and high saturation drift velocity, characteristics suggesting that devices made of . . . full-color LED displays become possible with the existence of blue LEDs, as blue was a missing primary color.
it can be smaller and more efficient than those made of silicon. Cree and others have shown that, even at red-hot temperatures, silicon carbide devices maintain functionality. Some of them, in fact, have continued to operate at 650 degrees Celsius. The wide bandgap also allows silicon carbide devices to operate at shorter wavelengths, enabling the creation of blue lightemitting diodes (LEDs) that could not be made from silicon. Moreover, full-color LED displays become possible with the existence of blue LEDs, as blue was a missing primary color. Growing Large Crystals to Reduce Costs Cree was founded in 1987 to commercialize silicon carbide and began by making LEDs on a silicon carbide substrate. Prior to its ATP project, Cree was already the world leader in silicon carbide technology and had been making 1-inch-diameter silicon carbide crystals. But progress in the development of devices based on silicon carbide had been stymied by difficulties in growing large, high-quality single crystals, a bottleneck that led Cree to pursue more research. During the ATP project, Cree advanced silicon carbide technology by developing methods to greatly reduce the amount of imperfections in crystals and to increase their size to two inches or greater in diameter. Larger-diameter crystals result in lower production costs, which 55 PROJECT: To substantially reduce the cost and improve the durability of light-emitting diodes (LEDs) and other electronic and optoelectronic devices by increasing the quality and size (to 2 inches or more) of silicon carbide (SiC) single crystals. Duration: 6/15/1992 -- 6/14/1994 ATP number: 91-01-0256 FUNDING (IN THOUSANDS): ATP $1,957 Company 435 Total $2,392 The Real Color DisplayTM, a moving sign which is capable of displaying the full range of colors, made possible by the use of blue LEDs. s formed Real Color Displays, a wholly owned subsidiary, to exploit this technology for fullcolor LED displays; s received a $6 million order in September 1996 from Siemens for blue LEDs; and s 82% 18% supplied the SiC wafers for components in the SiC solid-state transmitter used by Westinghouse Electric to make the first U.S. commercial-scale high-definition TV (HDTV) broadcast in April 1996. COMMERCIALIZATION STATUS: The larger SiC wafers, made with the ATPfunded technology, are being used in the fabrication of blue LEDs sold to many industrial customers. The wafers are also being provided in limited quantities for development projects in government and industry research laboratories. ACCOMPLISHMENTS: Cree essentially met or exceeded all of the technical milestones. Successful development of the technology is indicated by the fact that the company: applied for one patent on technology related to the ATP project;
s s presented several papers at professional conferences; s raised $13.2 million via an initial public stock offering in February 1993; made high-quality, 2-inch SiC wafers, greatly opening up the blue LED and SiC wafer markets; s s raised approximately $17.5 million in a private stock offering in September 1995; increased annual revenues from $3 million at the start of the ATP project in 1992 to $7.5 million at the end of the ATP award period in 1994; received $5.8 million from the Defense Advanced Research Projects Agency in May 1995 for further development of silicon carbide growth processes to support production of 3-inch wafers; s OUTLOOK: The improved processing technology makes the outlook for the commercial use of SiC crystals highly promising. The cost of producing blue LEDs has already been reduced substantially, and the expected widespread commercial availability of larger-diameter SiC wafers promises a new range of applications, including HDTV transmitters. Benefits in the form of lower costs and higher quality will accrue to industrial users of blue LEDs and SiC wafers, as well as to consumers who use devices containing these two Cree products. COMPANY: Cree Research, Inc. 2810 Meridian Parkway, Suite 176 Durham, NC 27713 Contact: Calvin Carter Phone: (919) 361-5709 Number of employees: 41 at project start, 210 at the end of 1997 s . . . devices that were impractical to make with pure silicon can be made with silicon carbide.
are crucial to opening markets for silicon carbide devices. The company also developed ways to significantly improve the doping (adding impurities to achieve desired properties) and epitaxial deposition (growing one crystal layer on another) processes for silicon carbide. Improving doping uniformity directly increases production yield and thus reduces costs. Cree's success with the ATP project enables the fabrication of electronic devices that can operate at much higher temperatures and withstand high power levels. Silicon carbide components used in experimental high-definition television (HDTV) transmission, for instance, delivered more power, lasted longer and cost less to produce than conventional silicon-based components. Now equipment that was costly to manufacture (owing to the need for heat-dissipation systems) can be produced less expensively, and devices that were 56 A D V A N C E D T E C H N O L O G Y P R O G R A M Cree's success with the ATP project enables the fabrication of electronic devices that can operate at much higher temperatures and withstand high power levels. impractical to make with pure silicon can be made with silicon carbide. New Products: Blue LEDs and Silicon Carbide Wafers The ATP project has been highly productive for Cree and the economy at large. The company has used the new technology to produce larger silicon carbide wafers to use in its fabrication process for blue LEDs. It is also offering the larger silicon carbide wafers for sale to other companies. Cree is using the ATP-funded technology to reduce the cost of producing blue LEDs, and their sales have increased substantially. Production cost is primarily a function of the number of wafers processed. If wafer size can be increased dramatically, the cost per device will decrease dramatically because so many more devices can be made on a wafer. The silicon carbide wafer technology is also aimed at markets for other blue light-emitting optoelectronic devices, optical disk storage, microwave communications, and blue and ultraviolet laser diodes, as well as high-temperature, high-power and high-frequency semiconductors. Benefits for the Economy Benefits from the new silicon carbide technology are already accruing to customers who have bought large volumes of blue LEDs or silicon carbide wafers to use in their own production. Performance measures (resistance, power output, sensitivity to light, operating temperature) for silicon carbide devices are frequently large, relative to available alternatives. Economic benefits from these performance improvements spill over to other producers involved in fabrication and assembly before a wafer-based product reaches the end user. The total of these incremental benefits is expected to be much larger than the profits Cree receives for selling the silicon carbide wafers. Cree's private success has led to public benefit, which is expected to grow as the number of applications for larger silicon carbide wafers increases. Westinghouse, for example, used Cree's silicon carbide wafers in fabricating components for the transmitter it used in the first commercial-level HDTV broadcast in the United States, in 1996. Westinghouse said its transmitter can deliver three times more power, has longer life and costs less to produce than conventional silicon-based transmitters. Although the number of HDTV transmitters that will use silicon carbide wafers is unknown at this time, widespread use of this technology in HDTV broadcasting could produce large general economic benefits if it speeds commercialization of HDTV. The low-cost blue light emitting diode (LED) produced with new silicon carbide crystal technology. ATP Advantages Cree reports it was attracted to the ATP as a funding source for the development of the bulk crystal and epitaxial growth technologies because the company could retain its process technology knowledge. The ATP award also helped Cree form alliances with research partners and speed the development work, enabling the company to get results about 18 months sooner than it would otherwise have been able to do. During the course of its twoyear ATP project, Cree also grew significantly. Company officials say the success of the ATP-funded project was primarily responsible for a subsequent award of $5.8 million from the Defense Advanced Research Projects Agency (DARPA) to further develop silicon carbide growth processes to produce 3-inch wafers. If wafer size can be increased to 3 inches, the cost per device will drop even further. This DARPA project got under way in May 1995. . . . silicon carbide wafers . . . used in the first commercial-level HDTV broadcast in the United States . . . 57 Electronics Cynosure, Inc. Harnessing Cheap Diode Lasers to Power a Low-Cost Surgical Laser S urgery is performed tens of millions of times a year in the United States, and it is usually a painful, risky procedure for the patient. It is also risky for the surgeon in terms of malpractice liability. Patients, surgeons and health insurance companies are constantly looking for new, less-invasive procedures to replace conventional surgery. Laser surgery is a prime candidate. One problem that limits this approach, however, is the price of equipment. A typical 100-watt surgical laser costs about $700 to $1,000 per watt of laser output, or about $70,000 to $100,000. Photomicrograph of an array of multi-level diffractive lenses, fabricated with a 193 nanometer excimer laser. A Laser for Lower-Cost, Less-Invasive Surgery This ATP project with Cynosure, founded in 1991, was designed to develop a smaller, lessexpensive laser source for surgery and other applications. The idea behind the Cynosure laser system -- which was expected to sell for about $150 to $200 per watt of laser light delivered at the end of a surgical optical fiber -- is based on harnessing the light from an array of 200 semiconductor, or diode, lasers. The problem with this approach in the past has been the difficulty of exactly aligning all 200 beams before they go into the diffractive optics transformer that collimates them into one tight, powerful beam. Minor inaccuracies in the alignment of the individual lasers can greatly degrade the performance of the system. Cynosure's innovation was to develop an automated system to custom-mill arrays of 200
58 corrective lenses to match arrays of 200 diode lasers. In such a system, diagnostic equipment measures the alignment error of each laser beam and feeds the results to a computer, which drives a powerful laser that mills the lens array in less than 10 minutes. The result is a customized lenslet array that corrects the beams before they enter the transformer. Barriers to Commercialization Cynosure successfully designed and built a customized lenslet array to correct the beams The researchers, however, failed to build a system that could generate the target power level . . .
T E C H N O L O G Y from an array of 200 diode lasers. The researchers, however, failed to build a system that could generate the target power level -- 20 watts of laser light from a medical optical fiber -- because the company was unable to secure an adequate, low-cost supply of a lowtech component: a collimating array. The intended supplier, which was the sole source of the collimating array, stopped making the device and sold its production division. The new owner also chose not to produce the array. To make use of some of the technology developed in the ATP project, Cynosure is collaborating with the Lincoln Laboratory at Massachusetts Institute of Technology and using about $100,000 from the Small Business Technology Transfer Program to develop a "low-cost diode-laser system for treatment of arrhythmia" for the National Heart, Lung and A D V A N C E D P R O G R A M PROJECT: To design an optical system for collecting, aligning and combining beams from an array of semiconductor lasers into one powerful beam, an achievement that will lead to the development of smaller, cheaper lasers for surgery and other applications. Duration: 5/1/1993 -- 4/30/1995 ATP number: 92-01-0136 FUNDING (IN THOUSANDS): ATP $1,965 Company 2,067 Total $4,032 s is collaborating with Lincoln Laboratory and using funds from the Small Business Technology Transfer Program to develop a "lowcost diode-laser system for treatment of arrhythmia," based on the ATP technology, for the National Heart, Lung and Blood Institute. 49% 51% ACCOMPLISHMENTS: Cynosure designed and built a fault-tolerant optical system for a diode-laser array but was unable during the project to obtain a laser beam with the targeted 20 watts of output from a medical optical fiber. Later, the company achieved this goal with an alternative approach built, in part, on the knowledge developed during the ATP project. The company:
s COMMERCIALIZATION STATUS: Commercialization was stymied by Cynosure's inability to secure the supply of a critical part at an affordable price. Since the ATP project ended, the company has taken a different, less-sophisticated approach to building a commercializable medical laser, using its own funds. That device has achieved the 20-watt ATP goal, and the company is scaling it to achieve 200 watts output. Commercial lasers are scheduled for market introduction in the near future. OUTLOOK: The benefits originally expected from commercialization of the ATP-funded technology should be realized via commercialization of the alternative technology that built on the technical knowledge developed in the ATP project. COMPANY: Cynosure, Inc. 10 Elizabeth Drive Chelmsford, MA 01824 Contact: Horace Furumoto Phone: (978) 256-4200 Number of employees: 30 at project start, 120 at the end of 1997 Informal collaborator: Massachusetts Institute of Technology, Lincoln Laboratory using fiber-coupled lasers, which are manufactured using standard optical fabrication methods and readily available components. The company expects this approach will not only reduce the cost of medical lasers but will also cost less than the diffractive optics-combiner approach envisioned by the ATP project. By significantly reducing the cost of surgical lasers, the Cynosure technology would enable wider use of minimally invasive surgery, reducing hospitalization times and lowering health-care costs. For example, gall bladder removal by conventional surgery requires a received one patent for technology related to the ATP project: "Fault-Tolerant Optical System Using Diode Laser Array" (No. 5,369,659: filed 12/7/1993, granted 11/29/1994); The company's switch to a different technological approach using readily available parts to concentrate the laser beams allowed commercialization to resume.
4- to 6-inch incision that results in four to seven days of hospitalization and a month of recovery time. When the removal is done by laser via a fiberoptic scope inserted through a small incision (a procedure already in widespread use), the patient is hospitalized for only two or three days and recovers much faster. Less-costly medical lasers would likely increase gall bladder removal by laser. Funding from the ATP allowed Cynosure to perform research and development work it would otherwise have been unable to do. The award enabled it to hire highly qualified optical physicists to conduct the research on diffractive optics, and to develop the technical capability needed for future manufacture of diffractive optics devices. Cynosure is currently considering licensing this technology to a company whose core business is diffractive optics. In addition, the availability of highly sophisticated optical diagnostic equipment allowed Cynosure to better understand and test the fiber-coupled equipment it is developing for the commercial sector. s s published a paper on its research findings; was ranked number 112 in the 1996 Inc. magazine list of the 500 fastest-growing private companies in America;
s increased its sales from $626,000 in 1991 to more than $23 million in 1997; and Blood Institute. The company is proposing to extend the scope of the project to include other conditions, besides arrhythmia, that can be treated with minimally invasive surgery. This new project is based in part on the demonstration that the ATP-funded technology, as modified by the company, is capable of delivering 10 watts of power into a 100-micron fiber-optic tube. Alternative Approach After the ATP project, Cynosure investigated alternative techniques, based on commercially available components, to channel the many beams from diode-laser arrays into a surgical optical fiber. The company found this can be done by grinding a hyperbolic lens onto the end of a small optical fiber, fitting one such fiber to each diode and stacking the fiber-cou- pled diodes into a two-dimensional array, as the ATP proposal had suggested. The fibers take the place of the diffractive optics in the proposed ATP laser system, with the tiny lenses directing the output from the diode array into a single fiber. The company's switch to a different technological approach using readily available parts to concentrate the laser beams allowed commercialization to resume. Commercial lasers are now scheduled to be available in the near future. Mission Accomplished Lower-cost, higher-power medical diode lasers are a necessity for minimally invasive surgery, and it is said that necessity is the mother of invention. Cynosure invented the approach 59 Electronics Diamond Semiconductor Group, LLC (DSG) Lowering the Cost and Improving the Quality of Computer Chips B illions of integrated circuits -- the tiny chips that run personal computers and thousands of other electronic devices -- are fabricated every year in the United States through ion beam implantation, a technique for introducing carefully controlled impurities, or dopants, into specific locations on the semiconductor wafers from which chips are cut. Dopants control the electrical properties of the semiconductor, forming the transistors and other microscopic components of each chip. . . . the new approach enables the production of about 2.5 times as many chips from a single wafer as the 200-mm technology can make.
approach enables the production of about 2.5 times as many chips from a single wafer as the 200-mm technology can make. The use of DSG's new technology in production equipment makes it possible to lower the cost and improve the quality of computer chips and other integrated circuits. Worker holding the world's first 300 millimeter silicon wafer populated with electronic components using the wide beam ion implantation technology. Ion Beam Implantation for 300-mm Wafers With chip components getting smaller and denser, the need for more accurate control of dopant implantation has risen. At the same time, competitive manufacturing has driven the size of production wafers up, making increased accuracy problematic because of the difficulty in precisely scanning the implantation beam across the wafer. This ATP project allowed Diamond Semiconductor Group (DSG), a two-person start-up company when it applied to the ATP, to develop a new and better way to implant dopants on large silicon crystal wafers measuring 300 mm or more in diameter, compared with the previous industry standard of 200 mm. Because the area of a 300-mm wafer is 2.25 times that of a 200-mm wafer and some waste always occurs at wafer edges, the new Multiple Advantages of Wide-Beam Technology A key innovation in the new technology is passing the wafer under a 350-mm-wide ion beam for implantation, rather than scanning the ion beam across the wafer. The broad beam is very stable and therefore highly accurate. The new equipment incorporating this technology is also significantly simpler than earlier machines and so is cheaper to build and maintain and is more reliable. Use of the DSG technology has already improved fabrication quality substantially relative to the existing industrywide standard. It doubled the mean time between failures, which means that on average, failures occur only half as often as with current equipment. The DSG technology also lowers fabrication costs by allowing implant equipment to be designed to work on one wafer at a time. Although it seems counter-intuitive, single- 60 A D V A N C E D T E C H N O L O G Y P R O G R A M PROJECT: To develop a novel approach for introducing dopants -- substances that alter the electrical properties of semiconductor materials -- into large semiconductor wafers to enable faster, less-costly fabrication of larger wafers with smaller, more-densely packed components. Duration: 3/1/1993 -- 6/30/1994 ATP number: 92-01-0115 FUNDING (IN THOUSANDS): ATP $1,326 Company 393 Total $1,719 selling the equipment to commercial customers; and
s licensed its technology to Mitsui Electronics and Shipbuilding for a flat-panel display application, after U.S. companies declined the licensing opportunity. DSG used $6.1 million from Mitsui to develop a 650-mm flat-panel component for displays. In 1997, Mitsui signed its first contract to supply the displays to a customer. COMMERCIALIZATION STATUS: The technology has been commercialized in one application and is very near commercialization for a second application. Chip manufacturers using the Varian SHC80 implant system (which incorporates the technology) are producing larger (300-mm) wafers than before (200-mm) and making them faster, with higher quality and at lower cost. OUTLOOK: The outlook is excellent. Varian is already selling semiconductor fabrication equipment that incorporates the new technology, and a flatpanel display application is under way. The technology generates cost savings not only for companies using it to make computer chips but also for those who ultimately buy the chips and the products containing them. The benefits directly captured by DSG will likely be only a small fraction of the total net benefits the technology generates for the economy. COMPANY: Diamond Semiconductor Group, LLC (DSG) 30 Blackburn Center Gloucester, MA 01930 Contact: Manny Sieradzki Phone: (978) 281-4223 Number of employees: 9 at project start, 25 at the end of 1997 Informal collaborator: Varian Associates Inc. The new equipment . . . is cheaper to build and maintain and is more reliable. 77% 23% ACCOMPLISHMENTS: DSG developed broad-beam ion-implantation technology (now embodied in Varian's SHC80 Serial High-Current Implanter) that successfully implanted the first commercially viable 300-mm semiconductor wafer. The new technology doubled the existing industrywide mean time between failures and provided additional ways to increase the quality and reduce the cost of chip fabrication. The company:
s received two patents for technology related to the ATP project: "Compact High-Current Broad-Beam Ion Implanter" (No. 5,350,926: filed 3/11/1993, granted 9/27/1994) and "High Speed Movement of Workpieces in Vacuum Processing" (No. 5,486,080: filed 6/30/1994, granted 1/23/1996); 40 percent to 50 percent of the market. Most of the equipment currently sold is for 200-mm wafers, and Varian was the first to market equipment that handles 300-mm wafers. Over the next five years, industry analysts say, the majority of implanters sold will be for 300-mm wafers. All 300-mm-wafer ion implanters currently manufactured by Varian include the DSG technology, and those produced in the future are expected to, as well. DSG is also developing the technology for another application: flat-panel displays, such as those used in notebook computers. The company has completed the development work through a licensing agreement with Mitsui Electronics and Shipbuilding, which invested $6.1 million in the effort. In late 1997, Mitsui announced it had already won a contract to supply the panels to a customer. Prior to licensing the technology to Mitsui, DSG attempted to interest U.S. flat-panel display companies in it. But most of this industry is off shore, and there were no interested parties in the United States. s applied for two additional patents for technologies related to the ATP project; licensed the technology developed during the ATP project to Varian, which incorporated it in its SHC80 implant system and is actively
s wafer processing is actually an advantage. Fewer wafers are lost if equipment fails, compared with current technology. The latter involves clamping 13 to 17 wafers to a large wheel, which then rotates at about 1,200 rpm under the ion beam. One failure may result in 13 to 17 unacceptable wafers. With singlewafer processing, only one wafer would be lost. In addition, single-wafer processing enables ion implantation to be coordinated much better with other fabrication steps, most of which are also performed one wafer at a time. Licensing for Two Different Applications The ATP project is already a commercial success. DSG licensed the technology to Varian Associates, an ion-implant equipment manufacturer, which has incorporated the new technology into products now being sold. Worldwide sales of ion implanters total $1 billion to $1.2 billion per year, and Varian has Benefits All Along the Supply Chain DSG's broad-beam technology enables the generation of substantial economic benefits. Varian sells its ion implanters to chip-fabrication companies such as Intel, Motorola and Texas Instruments. These companies, in turn, sell their chips to manufacturers that use computer chips in their products -- computer companies like Apple, Gateway, HewlettPackard and IBM, as well as firms that make automobiles, appl