Site: NTT Optoelectronics Laboratories,
Nippon Telegraph and Telephone Corp.
3-1, Morinosato Wakamiya
Atsugi-shi, Kanagawa-Ken, 243-01
Date Visited: December 16, 1994
Report Author: M. Warren (assisted by P. Shumate)
Nippon Telephone and Telegraph (NTT) is the major supplier of domestic telecommunications to the Japanese market. As its name implies, it has been somewhat modeled on AT&T in the United States. NTT was a public (government-owned) corporation operating as a regulated monopoly until it was privatized in 1985. Until that time it not only operated the domestic Japanese telephone services, but also influenced standards and set specifications for a number of Japanese electronics companies that manufactured products for NTT in a close relationship (NEC, Fujitsu, Hitachi, Oki, etc.). American trade pressure was partly responsible for a new (after 1981) open procurement policy that allowed foreign and domestic companies that were not part of the original NTT family of vendors to compete for NTT's business. At the time of the JTEC visit, NTT reported an operating revenue of 5.8 trillion yen, an income of 230 billion yen, and 215,600 employees. NTT is the largest company in the world, based on a market value of $130 billion. At the time of the JTEC visit, about 5% of operating revenue (290 billion yen) was spent on R&D. NTT's research staff totals 3,300, and the development staff totals 7,500. There are three Research Center groups: the Telecommunications Network group with 1,500 researchers, the Multimedia System group with 800 to 900 researchers, and the Science and Core Technology group with 1,130 researchers. The JTEC visit was to the Optoelectronics (OE) Laboratories in Atsugi, which are part of the NTT Science and Core Technology group, headed by Dr. Tetsuhiko Ikegami.
The OE labs have about 250 researchers, half of them in Atsugi and the other half in Ibaraki. The Atsugi researchers are organized in two departments. Dr. Junichi Yoshida heads the Integrated Optoelectronics department, which concentrates on semiconductor optical devices for lightwave transmission systems and OEICs. Dr. Takashi Mizutani heads the Photonic Functional Devices department, which includes work on two-dimensional array devices, quantum effect devices, and heteroepitaxial devices. The Atsugi laboratory is also the location of NTT's basic research department, which does fundamental research and includes a number of non-Japanese staff and postdoctoral personnel.
The discussion started with some informal conversation about the size and scope of NTT's research activities. Dr. Ikegami was present for the first half hour of the discussion. He expressed the opinion that after 1985, NTT research activities took a more basic emphasis, but with a recent reorganization there was an effort to refocus research toward applications.
Dr. Tomonori Aoyoma presented NTT's technology transfer activities. The substance of that presentation, with some supplementary discussion of the activities of two NTT subsidiary companies involved in technology transfer, is recounted in the technology transfer section of this report.
Four technical presentations were given as part of a short laboratory tour. The first technical presentation was of a waveguide pin-photodiode for OEIC applications by Dr. Yoshida. This device was part of an effort in OEICs for board-to-board interconnect applications. The OEIC concept is an integrated receiver consisting of a pin photodiode and a HEMT (high electron mobility transistor) on a semi-insulating InP substrate. The initial work was with surface-illuminated photodiodes that were fabricated in 5 channel arrays with -35 dB crosstalk. These devices had insufficient high frequency response, so waveguide pin photodiodes were being investigated for better efficiency at high frequency. Waveguide detectors 12 microns long exhibit 110 GHz cutoff frequencies. A novel aspect of this device is the use of a multimode structure to increase coupling efficiency from optical fibers into the waveguide structure. The multimode structure is realized by using intermediate cladding layers around the InGaAs low bandgap core layer to increase the optical field size while not decreasing the bandwidth of the device. The high bandwidth was realized by fabricating devices with wet etched mushroom mesas that allowed large (6 micron) contact areas to the cladding layers with a very narrow (2 micron) core layer to lower the RC product.
The second technical presentation was on tunable DBR lasers by Dr. Y. Yoshikuni. These devices have two feedback gratings that are periodic, chirped gratings. The grating mirrors, called SSG (super-structure-grating) mirrors, have periodic reflection peaks. The two grating mirrors are phase shifted so the peaks can be aligned at only one frequency for which lasing will occur. The laser has multiple contact areas that act as electrical tuning elements by varying the current injection to different contacts. Small variations in the refractive index in the parts of the laser can result in large changes in wavelength by switching the alignment of the reflection peaks of the grating mirrors. A three-port laser, operating at 1.55 micron, has shown quasi-continuous tuning over 34 nm and can operate at 10 Gbit/s, but the switching speed is around 2 to 3 ns.
A third presentation, by Dr. H. Mori, was on fabrication of InP laser diodes on silicon substrates. These 1.54 micron lasers were fabricated by sophisticated heteroepitaxy techniques that allow the growth of the lasers on substrates containing LSI silicon circuitry. The fabrication sequence uses epitaxial silicon surfaces (called ESS) on the silicon wafer to provide the desired double atomic stepped surface for high-quality III-V layers. The native oxide can be removed by a simple HF dip instead of the high-temperature processing used previously. The quality of the material is indicated by InP layers on silicon, with 70% of the photoluminescence peak intensity of InP layers on InP substrates. The laser diodes fabricated by this process have 50 mA thresholds and operate CW at room temperature. A demonstration was provided of a video signal being transmitted by one of these lasers.
The last technical presentation was by Dr. T. Kurokawa on current efforts in vertical-cavity surface-emitting laser (VCSEL) technology at NTT. This work is an outgrowth of earlier efforts on vertical-cavity, optically controlled switching devices called EARS (exciton absorption reflection switches). NTT's work is now concentrating on VCSELs and smart pixel concepts integrating VCSELs with photodiodes and FETs. The recent VCSEL work was on 850 nm substrate-emitting lasers grown by MOVPE on AlGaAs substrates. The approximately 10% aluminum concentration substrates were purchased from Hitachi. Substrate-emitting VCSELs at 980 nm on GaAs are produced by a number of groups, but 850 nm VCSELs have always been top-emitting devices because of the absorption of the GaAs substrate. The unusual choice of substrate-emission for 850 nm devices is claimed to have better thermal dissipation, access to the VCSEL emission from both top and bottom, and easier fabrication. The devices have been fabricated in 8 x 8 arrays with deep, mesa-isolated devices planarized with polyimide. Their performance is claimed to be identical to similar devices grown on GaAs. The work on smart pixels consists of integration of a metal-semiconductor-metal (MSM) photodiode, a metal-semiconductor FET (MESFET), and a VCSEL. Fabrication of the devices utilizes MOVPE and MBE growth on GaAs. An InGaP selective etch stop layer facilitates the integration of the different devices. Optical switching operations have been demonstrated at 250 MHz. The power dissipation for a pixel was quoted at 50 mW.
The technology presentations focused on device-related work. There was a short discussion during the meeting of systems efforts at NTT. NTT is heavily involved in Japanese multimedia activities. It was stated that 1995 would be a threshold year for multimedia research in Japan.
NTT is a service company and does not have manufacturing activities, although there is no law by which NTT is prevented from manufacturing its own hardware. Due to the nature of the company (NTT is still the biggest telephone operating company in Japan), public opinion does not approve of NTT having manufacturing facilities. This lack of manufacturing facilities is a situation similar to that of other national PTTs and the U.S. Regional Bell Operating Companies (RBOCs) and Bellcore. This is a major difficulty for NTT and most similar organizations that carry out leading-edge research on telecommunications technologies and systems. In order to test new system concepts, prototype devices and hardware are often needed. The 1984 divestiture agreement in the United States is particularly rigid about even prototyping by the RBOCs and Bellcore.
There are three procurement tracks for NTT to deal with manufacturers. Track I is direct purchase of existing catalog items that are commonly available. Track II is purchase of products that already exist and that must be modified to meet NTT's use or specifications. Track III is a route for R&D flow (tech transfer) to manufacturers in the form of product prototypes and specifications. This track is for products that are not already available and that NTT must develop itself and then transfer to a manufacturer. Technology transfer from NTT to other companies can also be performed through NTT Advanced Technology Corporation (NATC). NATC was started in 1976 to assist, as a consultant, the transfer of NTT technologies to small companies. This is accomplished without NATC actually fabricating devices or building systems.
The technology transfer process from NTT to outside companies can take place in four different ways. The first is through NATC. The second is through joint R&D with manufacturers as described by Track III procurement. This approach is mostly for large companies, and licensing is required for sale of products utilizing NTT technology to third parties. The third is conventional patent licensing, and the fourth is conventional software licensing. Another form of technology transfer exists: NEL (NTT Electronics Technology Corporation) is a subsidiary of NTT that does low-volume manufacturing of LSIs and some semiconductor devices. It was set up in 1982 to fabricate devices and systems mainly for smaller companies that lack fabrication capabilities, and to produce prototypes for NTT labs to purchase for trials. The role of universities and collaborative research also was discussed. NTT support for university activities includes the funding of about 20 postdoctoral positions and 20 to 30 visiting professors. There are typically 10 to 20 sponsored research cases at one time with an upper limit of 5 million yen(~$50,000) each. NTT retains the patent rights in its sponsored research in the form of nonexclusive, royalty-free licenses. There are some grants to researchers without contracts (i.e., gifts) that are limited to 1 million yen(~$10,000) each.
One interesting question related to technology transfer was whether NTT was required to transfer technology to its new competitors. The answer was yes. The example given was sales by an NTT vendor, such as NEC, of NTT-developed equipment. In such a case, NTT would collect a royalty and could have some control over the timing of the approval of the sale. New domestic competitors to NTT were listed as DDI (Daini Denda Inc.), which relies on a microwave transmission network; Nihon Telecom, which is laying fiber in railway rights-of-way; and Nihon Kosoku Tsushin, which is using highway rights-of-way for fiber.
NATC, owned by NTT, Japanese banks, and a variety of major electronics companies, was established in 1976 as Nippon Telecommunication Engineering Co., Ltd. It was renamed NTT Technology Transfer Corp. in 1985, and finally NTT Advanced Technology Corp. in 1990. As of April 1994, NTAC was capitalized at about $5 million, and it employed 680 people, most of them in Musashino. Drawing on NTT technologies and facilities, NATC assists in transferring technology to small companies that otherwise could not easily get at it, provides technology consulting, provides education and training, assists in designing manufacturing facilities, provides system integration, conducts technology surveys and evaluations, and assists in the sales of its own products. A benefit for NTT is that NATC provides strategic feedback to its research personnel on markets and the need for new technologies. Income from NATC's activities in 1992 was about 60 billion yen(~$600 million). As of the end of 1992, NATC had carried out a cumulative total of 1,260 technology transfers and 1,310 consulting engagements. A sampling of recent activities includes the following:
NEL, 82% owned by NTT, is located in Tokyo. Its principle function is to fabricate and sell devices based on technologies developed at NTT research facilities. According to figures available at the time of the JTEC visit, NEL has 478 employees, and sales revenues in 1992 were 6.5 billion yen(~$65 million). The product line includes 0.5 micron and 0.8 micron CMOS and Bi-CMOS LSI; bipolar and GaAs ICs and MMICs for up to 10 Gbit/s; laser modules at 980 and 1480 nm for amplifier pumps; 1310, 1550, and 1650 nm lasers for transmission and transmission monitoring; laser-driver modules; and a broad array of subsystem boards. For devices, the company offer a variety of packages, as well as custom-designed packages on request. NEL conducts sales through independent agents in the United States and Europe. Its primary role seems to be to supply small quantities of specialized components to system vendors that do not have that production capability themselves. This presumably opens up some system procurement to smaller players.
Through these two unregulated subsidiaries, NTT has found an extraordinarily effective means of obtaining prototypes for its own use, while also making its technology available to others. NEL is similar to Fulcrum, an unregulated subsidiary of British Telecom that specializes in system prototypes. NATC has expanded into a broad range of technology transfer and consulting activities, and with low capitalization (no labs) and a small number of people, it produces a strong bottom line ($750,000 per person).
The NTT facilities at Atsugi are very impressive and among the finest in the world. The short technical presentations also were very impressive and well illustrated NTT's capabilities in optoelectronic devices. The atmosphere is probably much like Bell Labs before divestiture. Some of the NTT management people indicated anxiety over whether they would be able to continue on their present course with the changing competitive environment for telecommunications in Japan. Dr. Aoyama expressed the view that not only the competitive environment, but also the discussion of NTT divestiture that will take place this year, will have large influence on the company's future direction.
NTT Atsugi R&D Center brochure
NTT Research and Development 1994 Review of Activities
Imagination & Technology NTT R&D information brochure
NATC NTT Advanced Technology Corp. Company Profile
Viewgraphs from presentation on NTT technology transfer processes
Numerous reprints of publications by groups that made technical presentations