Tetsuro Higashi, Chairman, Tokyo Electron Ltd.

Tetsuro Higashi

Note: This interview was conducted in Japanese, and translated to English.

SI Japan: The semiconductor market has recovered. The general industry forecast is that there will be a slump next year, but do you still have your optimism about end applications using semiconductors?

Higashi: My expectations for the end-application market are still strong. The first major application for semiconductors was mainframes, and by the late 1980s, this had changed to PCs. Today, it has changed yet again, this time to digital appliances, and I believe there is another change on the horizon to personal applications. Each time there is a major change in application, there is an accompanying evolution in technology and investment scale.

Today, technologies are fusing together, and there is no question in my mind that a huge market will emerge. I think the industry has to be aware of this. Even when you consider just the TV, the key product in the home electronics market, the analog era is drawing to an end and the digital era is developing rapidly. This has an enormous impact, and it will trigger the birth of a host of new technologies.

SIJ: And semiconductor manufacturing technologies will change, too.

Higashi: We need new technologies capable of meeting high-level device requirements. For example, an improvement in the relative permissivity of interlayer dielectric film would be a breakthrough for high-density, high-speed performance. As technology becomes increasingly sophisticated, the development of the necessary equipment and materials is underway around the world.

SIJ: I have heard it said that the technological barriers faced by the semiconductor manufacturing industry today are a bit higher than anyone expected.

Higashi : In general, a number of new technologies are being developed, but it takes a long time to bring them into practical use. Perhaps this is amplifying some apprehension about the future. In the past, there was constant innovation and the entire industry was surging forward, but today the atmosphere is much heavier.

SIJ: Do you have any ideas on possible solutions?

Higashi: I think we have to seriously strengthen the complete industry environment, where corporations compete with each other technologically and apply the results in products. And I think launching a new magazine for engineers so that they can stimulate each other is a great idea, too!

Semiconductor manufacturing is beginning to use 65 nm process technology, and a host of breakthroughs will be needed. The industry overall is not as nimble as it once was, though, and seems at times to be dead in the water. There are far too few stories about pushing ahead of the crowd, and pulling the industry along after. I think we need that type of story.

SIJ: There are demands for high device performance, and it's taking time to commercialize the solutions. Worse, the industry is becoming increasingly conservative.

Higashi: I don't think we can continue to meet the needs of shrinking design rules with existing process equipment. It's clear that we need some significantly new approaches in manufacturing technology. Tokyo Electron is committed to participating actively in development projects around the world. Once, semiconductor manufacturers led the process technology. We jointly developed it with them and made breakthroughs.

Today, massive capital investment is required, and semiconductor manufacturers seem to be hesitant to take the plunge. The load imposed by development expenses is also a big one. As a result, a variety of joint development projects have been launched, and it has become evident that one approach is to resolve problems through participation in these projects. Tokyo Electron also wants to participate in a wide range of projects, if possible.

SIJ: Some semiconductor manufacturing equipment companies have expressed concern over the transfer of technology out of Japan.

Higashi: National projects serve to evaluate the industrial capabilities of the nation, and are often very sensitive about releasing technology. It's dangerous to generalize about this problem. We are involved in a variety of projects, and I think it's important to carefully evaluate each one constantly. Unless the technology can be completely disclosed overseas, it is impossible to achieve technological leadership for the industry, in my opinion. This issue must be treated within a strategy designed to lead the world industry. In general, I think saying that nothing can be disclosed overseas is going too far. The industry needs more aggressive discussions about competitiveness.

SIJ: Can Japanese equipment manufacturers make a go of it only dealing with Japanese semiconductor firms?

Higashi: I hope that the Japanese semiconductor manufacturers will recover their strength. Process equipment and materials vendors are making tremendous efforts to maintain global competitiveness, and anything that diminishes their opportunities for success is not good. About 70% of our own business is overseas business. Given this situation, I don't think it's possible to say a company is maintaining competitiveness unless new technology can be utilized overseas as well. Unless we discuss this issue from the macro viewpoint of how to lead the entire industry, I think the Japanese industry will begin to decline. In the past, excessive protectionism damaged Japanese industry and, in my opinion, it is crucial to improve our global competitiveness.

SIJ: Are there any changes visible in the situation? Higashi: Some IDMs have positioned semiconductor chips as the core of its end-application strategy, and are making a major investment into semiconductor design and manufacturing. Strategies must also incorporate applications, and it is essential that products should be differentiated through process technology.

SIJ: Was there a time when process technology was considered unimportant?

Higashi: Quite possibly. Semiconductor manufacturers often didn't really consider the end products, and some changed into semiconductor "parts makers." They have to gain enough strength to be able to niche their products. Equipment manufacturers are participating and hoping to establish technologies to enable them to carve their own niches. Manufacturing process technology is becoming recognized today as a core technology, and now some companies are refusing to make their technology available to the equipment manufacturers. In the past, foundries achieved high-volume production of chips at low cost by using equipment manufacturers' process technology, but the limits to this type of competition have emerged. Foundries are beginning to move into cutting-edge technology now, too.

SIJ: So it will be more difficult for equipment manufacturers to make their products different from the competition?

Higashi: In R&D consortiums and national projects, there have been times when results are shared in the same ratio as participation, and some manufacturers believe this has made it harder for them to make products different from those of the competition. In the MIRAI National Project now, for example, there is a move to make the results available to the equipment manufacturers who are working together to make it succeed. From the viewpoint of the participating semiconductor manufacturers, however, of course it will not result in any product differentiation between participating firms. No doubt they will all look for ways to beat the competition, starting off from a base of the developed technology. The equipment manufacturers have no choice but to stay involved in both individual development competition and joint development projects.

SIJ: And the investment load continues to rise.

Higashi: It's pretty big. They just have to balance it somehow. Equipment development takes a lot of capital these days, and the completed equipment is expensive. These days, high level means complex. The best solution would be a technology that is simple, low-cost and has a major impact on the industry. My impression is that the environment today makes this type of solution unlikely, largely because of the excessive development capital requirements. A single 300 mm fab can cost a semiconductor manufacturer $300 billion, and so it makes sense to minimize risk by extending existing process technology as far as possible. Naturally, they become more conservative, and it becomes harder to evaluate and adopt major new technologies. This stimulates new companies to enter the field. Expectations are quite high for the automotive applications, but management hesitates when faced with that $300 billion investment per fab. The situation would be totally different if there was some expectation that costs would drop. Over the medium and long terms, small-scale investment is certainly safer than large-scale, but the prevailing opinion in the industry seems to be that evolutions of existing technology are safe. I really feel that the industry is caught in this dilemma, and the industry must find a way out of the problem together.

SIJ: Are there any exciting technologies being developed?

Higashi: The day is coming when MEMS will open up a lot of possibilities. At present, it is extremely difficult to commercialize MEMS. In the semiconductor manufacturing equipment and materials industries, the process starts with standardization, then concept definition, and finally dramatic innovation within these limits. MEMS hasn't reached this point yet, though, and will only reach maturity when a variety of fabrication technologies are utilized in a variety of equipment. The strength of Tokyo Electron is in semiconductor manufacturing equipment, and we fully intend to make our products stand out from the competition while developing MEMS-based technologies. We don't plan to use MEMS in the manufacturing systems themselves, but as a peripheral technology. No doubt some of our users will want to utilize our technologies. MEMS must form its own industry, and the same for nanotechnology. The issue for the future is how to best continue the expansion of the semiconductor industry while developing MEMS and nanotechnology applications.