Site: Nippon Sheet Glass Co., Ltd. (NSG)
8-1, 5-Chome, Nishi-Hashimoto
Sagamihara, Kanagawa-ken, 229 Japan
Date Visited: November 18, 1994
Report Author: D. Keck
Nippon Sheet Glass Co., Ltd. (NSG), founded in 1918, is part of the Sumitomo Group. It manufactures products in both the flat glass and specialty glass markets. The company has 5 main production facilities in Japan and 3 production facilities for automotive glass and SELFOC lenses in the United States. In addition, it has production facilities in Mexico, Malaysia, and Taiwan, and it provides technical assistance to China. NSG sales in 1994 were down about 20% from a high of $2.5 billion in 1992, leveling off at about $2 billion. NSG employs about 3,800 employees and is capitalized at about $400 million.
NSG in a period of economic downturn has been suffering the same economic difficulties as steel and other heavy industries, since much of its output goes to the automotive market. The company has done considerable restructuring over the past three years by reducing labor costs and by introducing automation in processing.
NSG's Optics Division, while suffering, does have several areas that are growing, such as SLAs, lens arrays for copy machines, fax machines, and printers. Mitsubishi Rayon is a big competitor, with plastic lenses as well as CCDs using conventional optics for the fax machine market. NSG's sales of SLAs to Oki and others for LED printers are good; although ink jets still have the largest market share, the LED printer market is increasing, and Oki has recently introduced its LED printer in the United States.
NSG has central research facilities in Itami (Osaka area materials research), a research facility in Tsukuba (see following site report), and its Technical Center at Sagamihara. The company is beginning to staff its R&D facilities not as a percent of sales but a percent of net profits; at the time of the JTEC visit, it aimed at 10 to 15% of net profit. In all, NSG has a total of several hundred R&D personnel, with 20% doing optoelectronics work. There are about 10 to 20 engineers at the Sagamihara Technical Center.
At the Sagamihara facility and supported by the Technical Center, NSG produces auto glass, LCD glass, and memory disk glass. In 1995 the company learned to fabricate high-frequency antennas in windows, relying on getting the electromagnetic design from the various universities with whom it has relationships. NSG has developed coating technology for UV and IR filters for automotive windows. Another technology it has developed and has in production is the ability to injection-mold the seal around automotive windows so that the windows come ready to install.
NSG is developing LCD windows (UMUĻ) for architecture as well as automobiles. It obtains the LCD polymer, Talique, from Raychem. These windows are very expensive and therefore may be used initially for smaller windows.
NSG has developed technology for directly depositing solar batteries on automotive windows, in particular sunroof windows, using polycrystalline silicon technology. The batteries can produce 50 to 60 watts/sq. meters. These were in production at "low levels."
The company's main thrust into optoelectronics has been the Selfoc microlens. Lenses are now fabricated using both thallium and lithium as ion-exchange materials. Thallium produces larger NA but is more dispersive, while lithium is restricted to small NA and is less dispersive. NSG now produces an image relay rod for medical and industrial applications; this uses a lithium ion exchange, and rods 30 cm in length are possible. The focusing period is a few centimeters, and resolution of 12 lp/mm is typical. Rod diameters between 0.35 and 3 mm are fabricated. The medical applications for this technology are growing, while the industrial market is leveling off.
In the microlens area, NSG's competitors are NEG, with a low-performing but low-cost ball lens, and Hoya, with high-performance but high-priced molded aspheric lenses. NSG tried to produce a compact disk lens, but plastic shut the company out with lower cost.
NSG has now added to its product line an axial gradient lens; this gives added power to a ground and polished lens. Axial gradient lenses are typically used as collimating lenses for such applications as printers.
At the time of the JTEC visit, the company was developing its planar lenses for liquid crystal displays and projectors. These are made by depositing a photoresist on metal coated substrates, exposing an array of holes, and then ion-exchanging through the holes. This produces a regular array of lenslets for enhanced light throughput in projection LCDs. Lens usage will depend heavily on lens cost. Nevertheless, NSG is in production with its lens plates, which have a few dozen 4 x 5 units with 300,000 lenslets in each.
NSG is collaborating with IOT (Germany) to produce ion exchange waveguide couplers. It is jointly developing a 16-port coupler that is viewed as the maximum port count. There was no Japanese market until the NTT announcement concerning FTTB for 1995. The company is planning to provide couplers to the subscriber loop market. At the time of the JTEC visit, it anticipated a 2 x 16 coupler that would be priced at less than $50/port. NSG's production uses a wafer of a few inches, made of a special silicate glass containing fluorine. Formerly it used ion-exchanged thallium, but now it uses silver. Thallium is still used for multimode couplers where a higher NA is required.
NSG is remaining flexible in the waveguide coupler area and has arranged to distribute fused-biconic-taper (FBT) couplers made by Sifam (SFO), a UK company advised by the University of Southhampton. FBTs are used in EDFAs. Cable companies like Fujikura are presently the main users, but Oki also has a good 980 nm laser.
NSG has a fiber collimator product consisting of a Selfoc lens aligned to a single-mode fiber. This is used for making isolators and laser modules. NSG does all of the assembly itself. It also buys laser and detector modules from various manufacturers, adds a Selfoc lens, and packages. It has developed its own machines to do the alignment and assembly. When used with a fiber collimator, these units allow NSG to sell LAN components. The company plans to introduce these modules into fiber channel, and so forth.
NSG is producing high NA (0.5) step index fiber with epoxy coating for short-distance communication applications. Fiber loss is 12 to 15 dB/km. This fiber is intended for LED use over 100 m distances. The fiber is a multicomponent glass made by a double crucible process. Production capability is 500 km/month. Showa is a competitor.
The same technology is being used for illumination fibers for displays, refrigerators, and so forth. The fiber NA = 0.5, with core and cladding diameters of 240 and 250 microns, respectively. The stated price was around 10 cents/meter with production of a few hundred km/month. Plastic clad silica fiber is the company's competition; however, all plastic fiber has losses of 100 to 300 dB/km. Mitsubishi Rayon is using deuterated plastics and has reported losses of 60 to 70 dB/km and prices of about $0.10/m. For plastic, NSG saw the issues as reliability, viz flames and heat, and spectral flatness. NSG's spectral flatness was typified as 30 and 60 dB/km at 800 nm and 500 nm, respectively. To date, NSG has not talked to any auto company. It would like to introduce this product in the United States.
NSG is producing flat glass for LCD substrates. It uses a float process with a subsequent grinding to get flatness. Thicknesses in production at the time of the JTEC visit were 0.55, 0.7, and 1.1 mm. With NSG's extensive glass coating technology it can easily apply the ITO coating. Presently it adds a SiO 2 layer to avoid Na migration. This is sputtered on in a multistage assembly line vacuum chamber. Moving holders with eight glass sheets on two sides are moved past the sputtering heads. The company is also experimenting with using a special fluoric-silica solution bath to make the surface anti-alkali layer. NSG has operated a subsidiary, Nanox, at Fukushima to produce small LCD panels.
NSG is supplying glass disks to an optical storage maker; this customer likes the stability of glass. NSG uses the same composition as for its LCD product. These disks are ion exchanged for strength and must be polished for flatness. The company tests reliability by spinning the disk at 24,000 rpm and looking for breakage.
NSG is beginning to develop a precision grinding and polishing capability for creating unique shapes and hopes this will become a new business. It is using commercial Speedfam equipment that it modifies.
A core competency for NSG is the lamination of plastic between glass without bubbles. This technology was developed years ago after the son of an auto company owner was killed in an auto accident. This accelerated the development of a laminated glass for automobile safety.
As part of the Real World Computing (RWC) project, NSG has an Optical Bus Interconnect System (OBIS) program. The entire project was funded at $50 million/year over ten years; of this amount, NSG receives about $300,000 per year. The goal is to make an optical motherboard.
Dr. Koichi Nishizawa showed the JTEC representative NSG's Selfoc lens production facilities. Here a few hundred thousand single element lenses/month and about the same number of lens arrays/month are produced. Lens arrays are produced in large sheets and then precision cut and polished. NSG performs an MTF for each array on an automated line. All arrays are 100% visually inspected but only sampled on optical performance. The majority of workers were temporary employees obtained from the relatively skilled pool in the area. Several other manufacturing companies such as Kawasaki are in this area and access the same pool of workers.
Regarding technology transfer, in the past Nippon Sheet Glass maintained research separate from production. Presently it is beginning to move people more to accomplish transfers using a team approach. The teams are led by production people, never by research people. Dr. Nishizawa stated that the company is beginning to adopt the "Gold Medal Project" system gaining favor in Japan, and that Sharp, Hitachi and others are using. Basically, a few projects are chosen and have the attention of the company's top person. The project leader is hand-picked and has one year to achieve the goal established by the top company person. The leader takes a risk but is given all the necessary resources to achieve the project goal.
Dr. Nishizawa described his view of government funding in Japan by using an historical perspective. In 1980, MITI funded the Optical Measurement and Control System program. This was an 8-year project funded at 10 billion yen with 11 companies: OITDA (the founder), Toshiba, Hitachi, Mitsubishi Electric, NEC, Fujitsu (electronics), Fujikura, Furukawa, Sumitomo (cable), Oki, and NSG. Later Fuji Electric, Shimazu Corporation, and Yokogawa Electric were added.
Dr. Nishizawa thought the MITI targets were unclear after 1990, and he had heard that some companies were reluctant to join. At the time, he viewed MITI as changing to staged projects, beginning with a feasibility phase where many projects are picked but few continue to the next phase. He believes this will apply successfully to next-generation projects.
He also cited a Science and Technology Agency program for funding 3 - 4 distinguished scientists and their groups each year. These will probably be restricted to medium-size companies. The program initially focused on nuclear and space programs, but now it covers biotechnology and optoelectronics. Funding is typically for 5 - 10 scientists at about $1 million/year. Examples of named scientists are Professor Nishizawa at Tohoku University, who has a semiconductor project, and Dr. Yoshida at Nikon, who has an X-ray reflector project.