Coherent has invested about $20 million, not counting staffing, in the formation of a new semiconductor group that the Santa Clara company hopes will garner a significant chunk of a rapidly growing semiconductor laser market. About $8 million of the $20 million capital investment was spent in the company`s relocation last year from relatively small spaces in southern California to headquarters that are now large enough to support high-volume semiconductor laser manufacturing.
Another $9 million was spent in acquiring Tutcore (Tampere, Finland), which provides Coherent with the epitaxial growth capability for semiconductor manufacturing. Coherent completed the move to its new quarters in the spring of 1997, and the original $20 million investment took its current physical shape in September when the semiconductor group was announced and the staff rapidly grew from 3 to 30.
Coherent`s vision for the new group is to become a major player in markets for high-power diodes, ranging in power from hundreds of milliwatts to tens of watts. Improvements in technology and market demand from applications with high growth potential, such as material processing, are expected to further increase power requirements by one or more orders of magnitude.
"Looking at the markets, we saw that there were a substantial number of diodes that needed to be made. And we couldn`t make those in our facilities in southern California, which were more oriented toward R&D," said Vittorio Fossati-Bellani, who has led several project start-ups at Coherent and was named president of the semiconductor group in September. "That`s why we planned and built a large volume manufacturing facility here."
The missing piece
The markets originally envisioned for the new group did not stray from Coherent`s traditional focus on high-power devices for medical, scientific, and commercial applications. The company had no plans for going into other high-volume market sectors, such as CDs or telecommunications, according to Fossati-Bellani. Last month, however, the Coherent Laser Group entered a telecommunications joint venture, which will call upon the resources of the semiconductor group.
A guideline in building the Silicon Valley facility was to include as much automation and process control as possible and to develop manufacturing processes that would work with a small number of devices initially but that would also be scalable to eventually produce hundreds of thousands of devices, according to Sandra Todd, also with the semiconductor grou¥(see photo).
In 1995, when Coherent started planning for the new facility and also began making semiconductor lasers for internal use, the company was buying epitaxially grown wafers from outside vendors, which was reasonable for internal markets such as manufacturing the Verdi laser. Buying wafers from outside, however, did not fit into the company`s high-volume plans for external markets. Neither did it suffice for the optimal development of new products, which calls for close cooperation between the processing, growing, and packaging aspects of manufacturing, Fossati-Bellani said. "Without the growing there was a very important missing link," he said. "The kernel of the operation was missing."
In seeking out sources for epitaxially grown wafers around that time, Coherent started working with Tutcore, a spinoff of the Technical University in Tampere, Finland. The subsequent acquisition of Tutcore has put Coherent in a reasonable position from which to seek a dominant market presence, he said. "At that point, all the various phases of semiconductor manufacturing [came] under our control," he added. "Now we have to make them work together."
The need to effectively address both internal and external markets was one of the reasons for starting a separate semiconductor group. The strategy behind the group structure at Coherent is that it allows each group to collaborate with others in developing or improving technology. At the same time each group also maintains market confidentiality with external customers that may compete with other Coherent groups.
Coherent`s other groups include the medical group (Palo Alto, CA); the laser group (Santa Clara, CA), which designs and manufactures CO2, tunable-dye, ion, YAG, YLF, and ultrafast lasers; the Auburn group (Auburn, CA), which supplies optics and laser instrumentation; and Lambda Physik (Ft. Lauderdale, FL), which manufactures excimer lasers as well as Nd:YAG-pumped dye lasers and laser dyes.
The group structure at Coherent currently allows the Auburn group, for example, to sell optics in external markets as well as to other Coherent groups. It will also allow the new semiconductor group to sell devices to other Coherent groups while selling externally to suppliers of medical lasers, for instance, that compete with Coherent`s medical group.
A second reason for establishing a separate semiconductor group is that it makes it easier for the company to monitor the return on its semiconductor investment. A third reason for carving semiconductors out of the original laser group is because of differences in growth potentials, technologies, and business types. The semiconductor group expects to produce devices in quantities of tens of thousands and to experience steep growth and high volumes. Consequently, manufacturing controls will differ from those of the laser group, where quantities are expected to run an order of magnitude lower.
"Semiconductor laser manufacturing is kind of a hybrid," Fosseti-Bellani said. "It`s not low-volume laser manufacturing, and we of laser extraction say that this is high volume. If you look at semiconductor criteria, [however], this is very low volume."
Current semiconductor laser devices tend to be used for specialized and sophisticated applications ranging from diode pumping to material processing and medical applications that still do not command the volumes needed to drive down the prices of semiconductor lasers. "Those are expensive systems," he said. "And the devices that are in [the systems] are expensive devices." The prices of semiconductor laser devices range from a few hundred to several thousand dollars, he noted.
Demand for devices
The California group basically receives wafers from Finland and processes them, which includes photolithography, etching, cutting, packaging, and so on. In addition to supplying unmounted chips and bars from this stage of the process for needs internal to Coherent, the semiconductor group also supplies the unmounted devices to packagers who make specialty modules. Coherent had not initially intended to enter that market, but the market is substantial and is currently underserved because not all of the major laser diode manufacturers supply it, Fossati-Bellani said.
The magnitude of the market for unmounted bars is in the range of tens of thousands of devices, and in terms of dollars may be in the range of tens of millions. Tutcore used to serve that market as an independent organization. But the market has grown sufficiently that it now requires the additional manufacturing capabilities in Santa Clara.
The first diode system made so far by Coherent is a fiber array package (FAP) device, which is a bar whose emission is coupled into an array of 19 fibers that can be connected as a fiber bundle to a transport fiber for pumping or delivering applications. The simple system is an individual 100-µm, nominally 1-W device (.5 W if red) that gets coupled into a 100-µm fiber. The initial customer was internal (for the Verdi), but the devices are sold externally as well.
The theoretical production limit of the equipment in the semiconductor group—based on an estimated maximum of 10,000 wafers per year from the production reactors in Finland—would be about 25 million devices a year, if all devices were the same (at 100 bars per wafer and 25 devices per bar). The lack of a market that is yet that big, as well as device variation and the limitations of reasonable production efficiencies, make the 25-million number an unreasonable target in practical terms. The number does, however, indicate the level of investment in this technology as well as the capacity for growth.
The actual production capacity of the system in its current configuration is in the range of tens of thousands of devices, depending upon the mix of bars and single emitters being produced. And the current demand is in the range of thousands of devices, he said.
Coherent`s market estimates tend to sound conservative because of the company`s scientific focus, which doesn`t include massive markets such as consumer electronics and telecommunications. With that in mind, the three major markets for the new semiconductor group are diode pumping, medical, and material processing. Diode pumping (mostly at 800 nm) represents a $20 million commercial market with a much larger military segment, Fossati-Bellani said.
Medical is smaller than diode pumping but is expected to surpass it quickly. A major market potential in medical lies in photodynamic therapy, which has received approval at 630 nm, he said. Several companies, including Coherent, are racing to build a high-power product at that wavelength. And it is rumored that one company will announce a product soon. Newer-generation drugs will use currently available laser technology at 650 and 670 nm. That market is currently served by dye lasers pumped by other big lasers, an inefficient technology that is limiting market size. The development of diode devices, especially if packaged in a doctor-friendly manner, should have a significant impact.
The third market, material processing, represents the largest of the three categories. Fossati-Bellani estimated the size of this market to be in the hundreds of millions of dollars. Diodes are covering only $10 million or $20 million at present, but that should expand as technological improvements provide brighter and more powerful diodes, he said.
One more startup
Prior to taking over the semiconductor group, Fossati-Bellani has led several start-ups at Coherent. In 1979, he started scientific sales for Coherent in Italy and built up that organization before moving on. In 1984, he started Coherent`s ultrafast operation, which has become a substantial part of the scientific business. In the late 1980s he was chartered with bringing the diode-pumped solid-state technology to Coherent. Now he is looking forward to the challenge in semiconductor lasers.
"It`s fun to start something new," he said. "My interest is in building things where there was not anything before. [And] I think the diode business is even more important than my previous business ventures. That`s why I`m interested in promoting this and in making it succeed."