AsiaFocus
10/01/1997
China's bid to join the submicron era
Elizabeth Schumann, Semi, Mountain View, California
China is now halfway through the second year of its Ninth Five-Year Plan period. During the Eighth Five-Year Plan period (1990–1995), the Chinese electronics industry experienced rapid increases in productive capacity, technological capability, electronics output, and international trade volume. By the final year of the Eighth Five-Year Plan, the total output of the Chinese electronics industry reached RMB 247 billion (US$30 billion), a growth of about 25% over the previous year. In 1995, China experienced a record year in foreign trade of electronic products, with US$17 billion in exports and US$16 billion in imports. It was the first year that China experienced a trade surplus in the electronics sector.
The Chinese electronics industry celebrated another banner year in 1996. According to figures released by China's Ministry of Electronics Industry (MEI), total electronics output reached RMB 298.2 billion (US$35.9 billion) in 1996, a 20.7% increase over the prior year. Total sales revenue for the sector in 1996 was RMB 194.7 billion (US$23.5 billion), up 18% over 1995.
In 1996, the semiconductor sector was very active, with several new semiconductor facilities in the planning, construction, or early production phases. Most of the major new facilities are partially or wholly owned by foreign chip manufacturers (see table).
|
In general, China welcomes the investment of foreign companies, since their involvement intensifies competition while upgrading the overall technical level of IC products in the marketplace. At the same time, however, China is concerned about being overly dependent on foreign suppliers for the chips needed for its own burgeoning electronics industry. At present, Chinese companies produce only about 20–25% of the IC devices demanded by Chinese electronics manufacturers. Because Chinese manufacturers are predominantly involved in the production of lower-end devices, on a value basis this difference is more pronounced. It is estimated that China's IC market is worth about US$5 billion today, but the value of production in China is only about US$0.4 billion. By the year 2000, this relationship should improve somewhat, with a projected IC market of US$15 billion and production value reaching US$3 billion.
Challenges to an active IC sector
According to Tony Liu, VP of the newest Chinese semiconductor fabrication facility, Hua Hong Microelectronics, China had failed in its efforts of the past few years to enter the IC industry because it was not quite ready to support an active IC sector. Speaking at a recent semiconductor conference, Liu cited several factors that challenged the development of this IC sector. One of these was the Chinese electronics market, which, although potentially quite large, is relatively difficult to access and not uniform. A second difficulty is that the existing IC industry in China is geographically quite diffused, which impedes cross-cultivation. Also, Chinese companies lacked sufficient financial resources to fund the aquisition of high-tech, world-class management and the technology required for leading-edge mass production. Finally, a typically long decision cycle time has hampered the development of this industry.
Hua Hong Microelectronics — a domestic solution
In 1996, China launched its largest development project for ICs in the Pudong New Area of Shanghai. This facility, dubbed Hua Hong Microelectronics, is aiming to help stem the flood of imported products now dominating the Chinese market and to lead China into advanced semiconductor manufacturing. With an investment greater than RMB 10 billion (about US$1.2 billion), this would be the largest-ever project undertaken in China's electronics sector, according to published reports. The project includes a production line with an eventual monthly capacity of 20,000 wafer starts, producing advanced logic devices using 0.5-µm design rules on 200-mm wafers. Hua Hong is just one piece of Project 909, sponsored by the Ministry of Electronics Industry (MEI). Other aspects of Project 909 include a 200-mm wafer-manufacturing facility and several IC design companies.
According to Liu, there is great confidence that Hua Hong Microelectronics will be successful, whereas past efforts have failed. First, unlike other Chinese device manufacturers, Hua Hong will focus on addressing the external market, at least initially. Following the example of countries like South Korea, Taiwan, and Singapore, Hua Hong plans to open various facilities around the world, including a sales and marketing office and a design center in the US, and an additional center in Japan.
Second, by locating in Shanghai in the Pudong science park along with other microelectronics industry players, Hua Hong can take advantage of the external economies of scale that have been so important for success in regions such as California's Silicon Valley, Scotland's Silicon Glen, or the Hsinchu Science Park in Taiwan. Though a lack of sufficient resources was a problem in the past,
Hua Hong Microelectronics has about US$1.2 billion available for investment, and is planning an initial public offering in the near future to raise additional funds. High-tech managers are being recruited to the facility with the opportunity for high salaries comparable to those commanded by Silicon Valley personnel. Hua Hong reached an agreement with Japanese semiconductor maker NEC Corp. to ensure that access to advanced processing technology is available. (Under the terms of the agreement, Hua Hong Microelectronics will hold about 70% share, while NEC will retain the balance. NEC will transfer advanced 0.5–0.35 µm technology to Hua Hong and will assume a great deal of the management responsibility for the facility.) And finally, the long decision cycle time problem is being addressed by having high-ranking Chinese government officials directly involved on the board of directors for Hua Hong.
Chinese IC production technology currently lags world-class producers by about seven or eight years. If Hua Hong is successful in its development plans, that technology lag will be reduced to only about two years by the year 2000.
Conclusion
In order to achieve its goals, China must rely on the technological know-how of foreign semiconductor manufacturers, while at the same time taking steps to protect its large market from foreign domination. Thus, while foreign direct investment, such as the semiconductor facilities mentioned previously, is welcomed in China, some restrictions still apply. Wholly owned foreign firms, such as the Motorola (China) facility, are subject to regulations that require them to export a significant portion of their output. Joint venture approval is often restricted to those companies that promise a certain level of technology transfer. This means that the potential exists for China's total semiconductor productive capacity to exceed its internal market needs someday, as foreign firms take advantage of land and labor resources in China to build semiconductors for export, and Chinese firms build a semiconductor infrastructure to meets the country's growing domestic market needs.
Elizabeth Schumann is a senior market analyst at Semi. Semi, 805 East Middlefield Road, Mountain View, CA 94043-4080; ph 415/940-7905, fax 415/967-5375.
A tale of two cities — Singapore and Seoul
The National University of Singapore (NUS) first made news in semiconductor manufacturing when it announced the purchase of a Helios II synchrotron for x-ray lithography. Seoul National University (SNU) has long been a premier training ground for technologists in the booming Korean semiconductor industry. However, Korea is now following the lead of the US in cutting back funding for science and technology, while Singapore pushes ahead, driven by its own vision of its future.
The National University of Singapore
According to Bernard Tan, head of the department of physics and chairman of the Steering Committee for the synchrotron program at NUS, the synchrotron is intended to improve the standards of science conducted at NUS, enhance student interest in physics and materials science, and have an economic impact on the country. The machine will be housed on campus in a building where up to 10 beam lines will be available to industry, while 10 others will be used by university scientists doing "strategic research" expected to have economic impact. The machine will be equipped with a "wiggler" that, though not immediately useful for lithography, will facilitate programs in protein crystallography and other scientific areas. While industrial participation in the project is still being negotiated, Tan expects one to three steppers to be installed soon after the synchrotron is delivered. Nonsemiconductor lithography such as LIGA for deep resists and 3-D micro-machined structures is also on the agenda.
The synchrotron project is also associated with the Institute of Materials Research and Engineering (IMRE), headed by C. Fong Shih. IMRE has a three-tier strategy, with collaborations driven by both the university and industry and "global networking." The core programs include electronic, photonic, and biological materials; nanotechnology; advanced polymers; ceramics and coatings; and instrumentation development. Shih notes that manufacturing accounts for 25% of the Singapore GDP and that materials contribute 30–50% of the cost of products. "The government recognizes that the future competitiveness of Singapore's industry depends on the ability to do focused research," according to Shih. "Our long-term government commitment to R&D will attract the best foreign talent, as the US did in past generations. Singapore's science and technology funding will continue to double every four years. We hope to make Singapore the R&D hub for all Asia, turning our disadvantage in area and population into an advantage in useful knowledge, just as Edison found advantage in his deafness when he invented the phonograph."
Antony Bourdillon of the department of materials science and IMRE sees synchrotron lithography as key to advancing Singapore's role as a manufacturer of high-value semiconductors. "Singapore has nine fabs being built or in operation now, and room for sixteen more. What happens if the world goes to x-ray lithography?" he asks rhetorically. "NUS will have done the R&D in a timely fashion, if that happens. If not, we will still have a nice tool for relevant science."
The NUS professors are aware of the need for industrial scale-mask manufacturing to support x-ray lithography, but they believe that IBM has "licked the problem at 0.18 microns," according to Tan. Building an x-ray mask house is not appropriate for a university, but those who are interested do consult experts at NUS.
Among the other unique facilities at NUS is a 2.5-MeV Van de Graaf accelerator with a 100-nanometer dia. beam, used for "nuclear microscopy." Since this beam penetrates 100 microns below a surface, Rutherford scattering and other elemental analysis techniques can be done deep in a film stack with submicron resolution.
Seoul National University
Seoul National University has long been the foremost university in South Korea, supporting semiconductor programs elsewhere through the Inter-University Semiconductor Research Center (ISRC). The center operates a 1200-m2 cleanroom, complete with electron lithography and an Ultratech 1100 stepper. It trains 200 students a year in semiconductor design and processing. While more students study processing, the design track — which appears to require creativity — has a higher prestige, according to Young-June Park, director of ISRC. "Most of the processing students are not electrical engineers, and there seems to be an oversupply of materials scientists and process engineers in Korea. The design students are mostly electrical engineers and there is presently a shortage in this area," reports Park.
The ISRC cleanroom and fab were established 10 years ago using funds from a loan by the International Bank for Reconstruction and Development. "Ten years ago, Korea was a poor country," remarks J-H. Lee of the department of physics. "There is still no mechanism for the government to fund equipment like that at ISRC."
Consequently, Seoul National University cannot contemplate leading-edge programs like Singapore's synchrotron facility. ISRC does, however, allow students to do their own projects in the fab and research nonstandard devices and processes. The current ISRC fab is capable of completely processing a 1-micron CMOS wafer, while the proposed Singapore facility emphasizes only lithography and characterization. Still, the Seoul facility probably addresses current industrial needs better than the nascent Singapore center. Problems, if any, may appear in a few years as technological change makes current facilities and business plans obsolete. The big Korean companies are committed to large-revenue generic products like DRAMs, while Singapore seems more interested in foundry facilities and large-value-added processes. Semiconductor manufacturing may fracture into two different business models in the future, with Korea emphasizing low-cost production of a few home-designed devices and Singapore working toward flexible manufacturing of leading-edge chips designed elsewhere. Each nation uses its limited resources to promote its own vision of a successful future. Both of them make sense. — M.D.L.