Launching ASTRO:
Taiwan chipmakers debate R&D consortium

10/01/1999

When the big get bigger, the small get nervous. That's the case for Taiwan's chipmakers currently negotiating the establishment of a semiconductor equipment and materials research consortium.

First discussed by Taiwan's Ministry of Economic Affairs, the project - now named ASTRO for Advanced Semiconductor Technology Research Organization - is being spearheaded by the Electronics Research & Service Organization (ERSO), a division of Taiwan's government-funded Industrial Technology Research Institute (ITRI). The vision is to establish an entity similar to Sematech, which began as a consortium for US-based companies in the 1980s.

However, the ASTRO project has been delayed due to concerns among potential members over the size and influence of one member - Taiwan Semiconductor Manufacturing Co. (TSMC). With its recent acquisition of a 30% stake in Acer Semiconductor Manufacturing Corp. (ASMC), TSMC has become disproportionately large compared with Taiwan's other chipmakers. The Acer deal gives TSMC four out of nine board seats, as well as control of ASMC's marketing and production facilities. TSMC is also a major shareholder in Vanguard Semiconductor, a leading DRAM maker in Taiwan. "Many companies are still showing interest in joining the organization but recently it is somewhat delayed due to these acquisitions," according to Jyuo-Min Shyu, deputy general director of ERSO. "There are some concerns but we are still working on it," he says. TSMC declined to comment on the ASTRO plans, noting that the project is still in the discussion phase.

Other potential members, Winbond Electronics and United Microelectronics Corp., could not be reached for comment by press time.

The champion of the ASTRO project is Genda Hu, ERSO's general director and president of the Taiwan Semiconductor Industry Association. While he declined to discuss the status of the project in detail, Hu confirmed that "an intense period of activities" was underway. "My expectation is by [mid-September] there may be something concrete to be reported," he says.

If ASTRO does get up and running, it will use the core team already in place at ERSO's deep submicron laboratory in Hsinchu. A team of 80 researchers are working in this group, focusing on interconnect technologies such as copper and low-k. The deep submicron lab is a five-year project, currently in its fourth year. ERSO's Shyu says the project is working in cooperation with more than 10 equipment and materials vendors from Japan, Europe, and the US.

He says the greatest challenge for such a project is managing and dealing with people from different cultures. A similar challenge awaits ASTRO, which will have to balance the needs of a diverse group of potential competitors together under one body. "You need to be very open but you also need to respect other peoples' intellectual property rights," Shyu says.

As with Sematech, the driving force for the formation of ASTRO is the desire by chipmakers and equipment and materials vendors to reduce the high costs of developing next-generation IC processing technologies. While the Taiwan government financially seeded and supported Taiwan's chip industry, its budget is limited when it comes to encouraging development research into IC processing. For example, the five-year deep submicron project only has total funding of around US$80 million - small beans when it comes to developing advanced wafer-processing methods.

For ERSO, spending millions of dollars on the latest gear for lithography, etching, and so on is not feasible. "We need to have a consortium to share the R&D cost," says Shyu.

In another development at ITRI, a new body called the IC Design Service Center was officially established July 1st, charged with bringing together the disparate chip design resources of all the ITRI divisions. Previously ERSO, in addition to other ITRI divisions such as the Computer and Communications Laboratories (CCL) and Opto-Electronics & Systems Laboratories (OESL), operated their own independent IC design groups.

A key concern was that the intellectual property produced was not reusable outside ITRI or even within the organization.

"Many people still need to be educated on how to make their IP reusable internally and how to exchange with [outside] people," says Shyu, who is heading up the new design center.

At present about 46 CAD and design engineers are staffing the center, with one-third of those being newly recruited.

The move to centralize ITRI's IC design efforts is part of a wider push by Taiwan to leverage its store of intellectual property, especially in the system-on-a-chip (SOC) area. At an industry strategy symposium held in May by the Ministry of Economic Affairs, it was decided to pursue the establishment of an IP consortium to beef up efforts in this area.

"Taiwan is relatively weak in system design," says Shyu. "Without strong system expertise how can you do SOC? We have to do it step by step. We need to partner with the US system companies."

Although Taiwan boasts more than 100 IC design houses, very few are true IP providers - Unichip and Faraday Technology are two such examples.

Craig Addison,
Asia Correspondent