Laying down the future of semiconductor manufacturing
Arthur del Prado
Arthur del Prado has been sitting at the head table of the semiconductor industry since practically day one. The president and chief executive of ASM International, he oversees a global operation with nearly 10,000 employees and about 2.5 million square feet of plant capacity. As such, he has seen his share of innovation and technology disruption.
The tools his company makes include furnaces, reactors, chemical vapor deposition and other process tools on the front end, as well as die attach, wire bonders, encapsulation and automation and test equipment on the back end. As del Prado explains in an interview with Small Times’ David Forman, emerging technologies are poised to impact semiconductor manufacturing once again.
Among them is ALD, or atomic layer deposition - a process whereby individual atomic layers are deposited one at a time. By semiconductor industry standards, it can be slow. But del Prado claims ALD technology, in which Bilthoven, Netherlands-based ASM is a leader, is improving rapidly and may well be the only way to make certain components of chips as feature sizes scale down to new levels.
Q: You recently said you expected atomic layer deposition would be a key enabling technology for the semiconductor industry. How was it developed?
Well let me take a quick stab at some history. We started ALD by accident, more or less. We read about work which was being done in the northern part of Europe, in Finland, which at that time was primarily for use in catalysts.
We thought this was a tremendous application for the semiconductor industry, which was at that stage at micron levels. But when we looked at where the industry was going, which was where our fantasies were going, it really fit with the problems that this industry would be facing - meaning dealing with these really fine feature sizes.
So we brought that to the industry in 1999. We bought the Finnish company called Microchemistry. They were doing nothing but developing ALD layers for totally different applications.
Q: How did the initial rollout go?
We are pretty smart technology people, and we have a way of developing these technologies normally ahead of others who cover the same field. Generally, we have a pretty good knack for guesstimating trends.
But what happened is we missed the 200 mm era in this industry, which really boomed in the mid-90s. We missed that window. We were supplying the industry, but it was one-sies, two-sies everywhere, never high volume. So we decided that 300 mm was going to be our game.
We were quite successful with a number of our products. But today, Moore’s law is not only being pushed forward by lithography. It used to be lithography. Every time when you made a step it was really guided by lithography. That time is over.
The new materials - materials science, I guess - is where a lot of the steps come from now. The combination of many of these developments is helping us tremendously in pushing forward Moore’s law. ALD is going to play a major role in that field. When we started ALD it was kind of disappointing to us because it took a long time before the industry became aware of it.
Q: Why did it take so long?
This is an industry that is full of brilliant people, but if you want to bring something to high volume with new materials, then you find a lot of resistance because the industry is also very conservative. It wants to work with what (already) works. So it took us a lot longer than we anticipated.
The reason it took so long for this technology to enter the industry was because it was too slow, in essence. It did not fit the bill at the larger feature sizes because it was too slow. When the feature sizes were getting to the level that you couldn’t do it anymore (with conventional approaches) that was when (some momentum shifted toward) ALD.
Today is 2006. It’s only last year that we started to push ALD to a higher volume. And now you see that ALD is really catching on. I think we are on the verge of getting the biggest change this year.
This is not because we are so smart but it’s because we were early in estimating what would happen and have the time horizon built into our working methods. ALD is now being applied to many other areas, for example, in microprocessors and DRAM manufacturing. Not every company will follow it at the same nodes, both in transistor formation and also in interconnects. We see a lot of divergence there.
Q: What differentiates ASM from its competitors?
We have a powerful front end group but we also have an extremely powerful back end group. We feel that front end and back end will converge in the future. You may raise your eyebrows and wonder how that may happen. We see very clear indications that front end is going to work with back end processes.
Part of the back end will move to the front end. We don’t even know yet where it will be. We feel it is one of the powerful capabilities of this company because we are the only one who has a meaningful front end and a powerful back end group.
What we see now is it’s not only our own ideas but the partners pushing us in this area to form teams that cross barriers. Can you imagine wafers stacked together? Positioning them and making through-holes. That’s what is going to happen.
There will be a filtration of front end processes into the back end area. And there will be a filtration of the precision of the mechanization capabilities for the back end filtering to the front end.
Q: When will that happen? How long will it take?
Today it’s beginning. How big is that market going to be today, next year, five years from now? We believe that in the next three, four, five years it’s going to be major. But don’t ask us exactly at what level because that is shifting. But it’s coming. It’s definite.
Q: Then why has there been pressure from your investors to separate ASM’s front end and back end divisions into separate companies?
That’s the interesting part. We are battling ignorance.
Q: So the integration of the manufacturing line is the argument you make for keeping them together?
It’s one of the arguments. We built this company globally, and we developed manufacturing capabilities that were fully vertically integrated. Why? Not because we loved it but because we had to. The Far East’s major contributor at that time was low cost labor. But you couldn’t source the parts (for precision machinery). So we had to develop that ourselves.
It’s a major advantage today. The synergy between front end and back end is that we have those manufacturing capabilities now in the Far East that we are putting to the power of front end. So the vertically integrated manufacturing will also be run now for our front end group, which means that from the margin point of view we can really increase our margins dramatically to have very competitive products.
Q: It’s the opposite of outsourcing, so you’re battling people who aren’t used to hearing this story today?
They don’t understand it. The short term investors are investors who are only participants in the company for six months. They want to unlock the value and take off.
What we want is to unlock the value of keeping these lines together over the long term. I was surprised myself how quickly margins can improve if you do design for manufacturability, which you can only do if you are vertically integrated.
The del Prado file
Arthur del Prado is chief executive officer, president and managing director of ASM International N.V. He oversees the operations of the worldwide organization from the company headquarters in Bilthoven, Netherlands.
He has led ASM since he formed the company in 1968. In 1984, he also co-founded ASM Lithography N.V. (ASML) through a joint venture with Philips Electronics N.V.
For many years, Mr. del Prado served as a director of JESSI, the Joint European Submicron Silicon Initiative. He currently serves as a director of its successor MEDEA, the Micro Electronics Development for European Applications, a non-profit project. He has been appointed a member on the board of directors of various European companies and serves on the board of Dujat, the Netherlands-Japanese Trade Federation. He was previously a member of the advisory council of ABN-AMRO bank.
His honors include the 1983 Bi-Annual President’s Award from the Dutch Association of Corporate Board Members and Presidents.