Site: Sanyo Microelectronics Research Center
1-18-13 Hashiridani
Hirakata City
Osaka 573, Japan

Date Visited: November 17, 1994

Report Author: S. Forrest

ATTENDEES

JTEC:

D. Crawford
S. Forrest
R. Hickernell
F. Leonberger
C. Uyehara

HOSTS:

Dr. Keiichi Yodoshi
Senior Manager
K. Yagi
H. Hanafusa
M. Shono

BACKGROUND

Sanyo is one of the consumer electronic giants of Japan, with over 1.5 trillion yen(~$15 billion) in annual sales in 1993 and 25,000 employees. As is generally the case for such companies, it has a broad range of products in many different areas. In particular, Sanyo holds the world's largest market share of a-Si solar cells, claiming to be the first company, in 1981, to use a-Si to power hand-held calculators. It continues to improve a-Si efficiency for use in homes, automobiles, power plants, and so forth. To accommodate its breadth of interests, Sanyo has five "research centers" that feed their technologies into eight business headquarters. Within each business headquarters is a product development center where the new technologies are produced. The Microelectronics Research Center works with the Semiconductor Development Center to transition photonic technologies.

DISCUSSION

Much of the discussion focused on mechanisms of technology transfer between the various corporate segments and on the visible laser activities of the Microelectronics Research Center. Sanyo has aggressively pursued the CD laser business for several years. The company's particular strength in this area has been the development of the shortest-wavelength CD-type laser, which operates at 635 nm wavelength. The rationale for this device is that it can be used in quadruple density optical disks. At this wavelength, a 0.8 microns spot resolution is achieved, allowing for a 135 min. video disk as compared to the current 60 min. disk. The laser is a buried ridge structure grown by MOCVD in the GaInAlP materials system. To achieve the short wavelengths, the growth is done on (100) GaAs wafers with 7 degrees to 9 degrees misorientation to the [011] direction. This misorientation, which appears to be a technique pioneered by Sanyo, enlarges the band gap and results in excellent surface morphology. Typically, the misorientation shortens the wavelength by 15 nm from structures grown directly on (100) GaAs. Typical operating characteristics of these lasers is a threshold current of 35 mA, 5 mW CW power, and a lifetime (MTTF) of 10 4 hours at 50 deg. C under 5 mW output operation.

Using similar device structures in the AlGaAs system, Sanyo has developed a large family of high-power, short-wavelength laser diodes. For example, in the 630 nm wavelength range, lasers with powers approaching 20 mW are available, while at around 780 nm, 50 mW lasers are achieved, and in the 800 to 870 nm range, 150 mW output powers are achieved. Sanyo recently developed a new 780 nm CD laser. This laser is packed in a very simple, molded-resin, lead frame package that has no window or covering for the laser. Reliability does not seem to be adversely impacted by this packaging style. Due to its commitment to lasers, Sanyo claimed at the time of the JTEC visit that it held 30% of the domestic Japanese market in CD optical pickup heads. The company sells both the lasers and the full pickup mechanisms. Sanyo's latest application for the high-power devices is intersatellite communications.

In addition to high-power lasers, Sanyo is also developing VCSELs, although with a less clear idea of their ultimate application. Sanyo has been working in VCSEL technology for ten years; in 1989 it announced a GaAlAs device with a CW threshold current of 5.2 mA.

The new laser diode research is focused into four areas: (1) Short wavelength lasers consisting of the II-VI materials; (2) Laser-based OEICs used for optical interconnects and optical communications; (3) High-power/high-efficiency diodes for laser printers; (4) Low-noise/high-coherence lasers for CD player applications. There was no InP research being pursued at Sanyo at the time of the JTEC visit.

There are several other areas of research at the Microelectronics Research Center. The most notable is the center's AMLCD work. Attractive 3 - 5 in. displays are in production, with 10 in. displays under development in the laboratory. Sanyo representatives also showed the JTEC team the operation of an LCD projection screen with 1.5 million pixels. It appeared that Sanyo was also beginning to investigate organic electroluminescent materials for display applications.

Technology transfer at Sanyo appears to occur by several different means. In general, it involves a three-level process starting with the generation of new technologies in the research center. This technology is shifted to the development center for final development, and then finally placed into manufacturing. Sometimes, although not always, the transfer of technology is accompanied by the physical transfer between the research and development centers of members of the engineering staff. In one example transfer, it was suggested that a particular project might begin with a joint effort of ten researchers. Once the concept was proven in the lab, perhaps three of these engineers would move to the development center to facilitate technology transfer. Since the research and development centers are physically separated, this move entails the movement of both the researcher and his family. (One does not have to worry about referring to a Japanese engineer as "he," since there are virtually no women engineers in the organizations the team visited.) Researchers move willingly, since they feel a sense of commitment and ownership of the technology. In the final stage, perhaps one or two of the engineers in the development center accompany the technology to manufacturing. Manufacturing and development generally are located in the same facility.

Engineers who go all the way from research into manufacturing do not generally return to the research environment, but rather go into the marketing or business side of the company. Those who go only to the development stage either stay there, or sometimes return to the research center, where they take on the next research challenge. Total time from research to manufacturing is of course highly variable, depending on the technology. For example, a laser technology transfer might take five years in the research stage, followed by two years in development; whereas projects in VLSI design may take only six months at both levels.

It needs to be emphasized that this movement of personnel to effect technology transfer is only an example and generally applies to those immature technologies that are not yet firmly established in the development centers. In that case, the engineers from the research center are required to efficiently build a technology base in the development-oriented locations. A counter example would be in VLSI, where only paper plans are typically transferred between the research and development centers, since the technical foundation for that technology already exists in the development center.

There seems to be very little collaborative work between the Microelectronics Research Center and universities.

SUMMARY

Sanyo seems to be a company that carefully plans the markets in which it wants to play, then focuses its efforts strongly and over a long term to achieve its goals. It has worked over ten years to develop a large range of short-wavelength laser products. Due to this dedication, it now has captured a significant fraction of the CD laser market. Indeed, its short-wavelength products are poised to take over other markets traditionally occupied by He-Ne lasers. For example, Sanyo is now producing very sophisticated laser pointers and bar code readers, the latter promising to be a sizable market. The short-wavelength diode laser may soon replace the He-Ne laser completely, due to the longer lifetime and compactness of the laser diode. All of these developments appear to be the result of a relatively simple corporate structure and an effective and flexible process of technology transfer.


Published: February 1996; WTEC Hyper-Librarian