Advertiser and

Buyer's Guide
Buyers Guide

Chip Shots blog

Greatest Hits of 2005
Greatest Hits of 2005

Featured Series
Featured Series

Web Sightings

Media Kit

Comments? Suggestions? Send us your feedback.



Novellus Systems' Sass Somekh


Throughout much of his 23 years with the company, Sasson "Sass" Somekh's name was synonymous with Applied Materials. But when Mike Splinter took over the reins at Applied, Somekh's days were numbered. He left the company in 2003 and resurfaced in early 2004 as president of one of Applied's archrivals in the semiconductor equipment business, Novellus Systems. Somekh, chairman/CEO Rick Hill, and executive vice president (and fellow ex-Applied exec) Tom St. Dennis make up the Novellus leadership team in its office of the CEO. Charged with driving new product development and the product-line business units, Somekh has brought a sharpened focus and energy to Novellus's efforts in those areas.

Before he joined Applied, Somekh earned a PhD in electrical engineering at the California Institute of Technology and then worked at Bell Labs and Intel. A holder of more than 50 U.S. patents, he received a SEMI award in 1988 for his work in establishing plasma etch as a chipmaking technology and garnered lifetime achievement accolades from SEMI in 1994. In 1993, he was recognized as the coinventor of Applied's workhorse Precision 5000 CVD system, the first semiconductor manufacturing tool to be placed in the permanent collection of the Smithsonian Institution in Washington, DC.

I talked with Somekh during SEMI's Industry Strategy Symposium (ISS), held in January in Half Moon Bay, CA. We discussed the differences between the Applied he left and the Novellus he works for now and how his new company has evolved since he joined it. He cited examples of new technologies finding their way into the latest toolsets and stressed the ongoing importance of alliances and partnering. Our conversation concluded with a lengthy discussion of his relationship with Hill, during which Somekh explained the psychology behind why he and his new colleague work well together.—TC


MICRO: You're in an unusual position. You've gone from helping to run the number one equipment supplier to joining the executive team of perhaps its greatest competitor. What are some of the key differences you see between Applied Materials' operations and how Novellus operates?

SOMEKH: I know Applied from a year or two ago, and companies change as a function of time, as a function of management, and so on. So the reference I have is really Applied from that time. I think the historic success of Applied really came from two things: Applied was a truly global company in the sense that it recognized the importance of the different regions; it was the first one to go independent in Japan, to set up a large organization in Japan, recruit people from all the regions, and form a management that was multicultural, not just Western but multicultural. It really took advantage of the growth in Asia, for example. That was one major strength.

The second one was technology. Applied has the ability to attract and retain technologists, and promote them to a high level to run major portions of the company. There used to be a time when the office of the president had four people, three of whom were PhDs. The combination of these two things was very powerful for Applied.

Novellus started almost exactly 20 years after Applied with their first product introduction. Novellus's focus was productivity. Bob Graham, I guess, coined "first in productivity." When I came to Novellus, I was really amazed at how many good technologists there were—how much good technology was in the company. But with a company whose basis is in productivity, you're reluctant to introduce technology to the customer before you know that it's production-worthy. A lot of technologies were inside the company that the customers didn't even know about.

Technology is getting so complicated. If you don't work with a customer early on, it's almost impossible to do it. If you are responsible for process development at a customer, sometimes you have to sign up to design rules before you even know how to do it. You really need to interact closely with the suppliers, even in the thinking process, in order to make sure that by the time you're ready to ramp up production, you've worked out all the bugs.

Some of the things that we've done are to open up more to customers and to start working with customers early on in the product development cycle, to take the good technology that exists within the company. I'll give you an example: One of the technologies that we're pursuing now and have shipped already to customers is something we call PDL [pulsed deposition layer]. It's ALD-type step coverage at CVD rates. ALD has a bad name because it's a very slow process. But PDL has CVD rates, so it's pretty high throughput. It was invented at Harvard. We hired, several years ago, the graduate student [who invented it] and developed it in-house to a point where now we can very rapidly develop it into a beautiful tool where it's the only deposition tool that does not deposit in the chamber. It only deposits on the wafer.

So it's this ability to take the good technology we have inside, use the productivity mentality within the company, the history of the equipment, and be able to move very rapidly to engage with the customer early on in the development phase so that the customer can program this into their products, as opposed to waiting until it's bulletproof. At that point the customer says, "I had to make a decision a year ago to do that."

Novellus is the only company besides Applied that is able to master a number of areas. So there's CVD, we're the leader in electroplating, and we have outstanding product in PVD as well. And we have product in the area of CMP. So we're able to master these multiple technologies and, in every product that we have, we can demonstrate both the technology and the productivity.

MICRO: Anything else surprise you about the company?

SOMEKH: One of the other things we have done is that once we recognize that we have a hole in our roadmap, we move quickly and do an acquisition. We have acquired Angstron [Systems] and have filled a hole we had. The hole we had is the PVD chamber. The holocathode magnetron is a unique chamber—it has very good step coverage. So we had a strong belief that PVD will do a good job at 45 nm.

MICRO: That it was extendable.

SOMEKH: Yeah. I think Rick Hill mentioned several years back that this technology will be extendable to 45 nm. But customers said, "hey, we want to buy a platform and we want to know that it has the ultimate capability," and the ultimate capability is ALD. So even though we might use PVD for 45 nm, it's important to have ALD, and that was a hole in our roadmap because we're somewhat smaller and cannot afford to cover every base. But at that point we moved very rapidly, acquired the company, integrated the thing, and this month we're shipping our first integrated Angstron chamber—on the Inova platform—lock, stock, and barrel. It's striking technology, the ion-induced ALD. It gives a superb barrier—a 10-? barrier is already a good barrier, you know. Ten angstroms. That's not that many atoms. The reason it's so good is this ion-induced technology, which makes a high-density film. ALD is beautiful because it deposits layer by layer on a substrate, but because it deposits layer by layer, it's very sensitive to the surface. So if the surface is not right, that layer by layer doesn't build nicely, doesn't have the right adhesion, it's different on this surface than that surface, it's different here than there. So by flooding it with very soft ions, we sort of activate it, clean the surface, and make sure we have very good film.

This technology allows us to do very thin barriers, and we're also the only ones that can do ALD copper. So the combination of the two is absolutely striking. Maybe it'll be used for 45 nm, maybe it'll be used afterwards, because 45 nm will be the transition period for a whole variety of processes. Luckily we can do 45 nm with PVD, but we can also get striking results with ALD, so we'll see how that goes.

I was really surprised with the great deal of good technology. We've opened up to customers, we have more engagement with customers than we traditionally had as a productivity-focused company. I think customers worldwide are now exposed to the good technology we have in every product category.

Another example is in the area of tungsten. Although tungsten is not that many layers on the wafer, it's still a very important technology. We have ALD nucleation. That is how you start nucleating the film. We have low-resistivity tungsten that almost approaches the bulk resistivity of tungsten. Before you deposit tungsten, you have to put down what's called a liner barrier. Right now, people do it in a PVD-CVD combination, sometimes in two different systems. We introduced a concept we call tungsten-nitride PNL [pulsed nucleation layer], which is an ALD process that allows us to put down a thin layer of tungsten nitride as the liner barrier. We do it in the same system as the tungsten, so we call the system direct-fill. In some cases, one system does the job of three different systems.

Another area of innovation is PECVD. The Vector is a very productive platform, very small and very inexpensive, and now we're combining deposition with a Vector with a UV cure, which was announced in November. That allows us to achieve a porous low-k value of 2.5, which we strongly believe is an integratable film. It's something you can stack up, and you can package it.

MICRO: You're almost talking like a materials supplier at times. There seems to be a transformation of the big tool companies, where you have to be more and more of the materials solutions provider. Most of what you've just gone through has as much to do with the films as it has with the machinery around the films, and that's an interesting evolution that's been going on.

SOMEKH: Right. There's the equipment, and there's the material, and there's the process. So you have to combine all of them together. We're not as much material as we are process or solution or, as I prefer to call it, process provider. It used to be you give a film and say, "here's the film, I did my job on the process." Then we said, "no, that's not enough, you've got to make sure it doesn't have particle defects, whatever." But it doesn't stop there. You give somebody a film, say, "here's the film, great particle performance, yeah", but he says, "I can't etch it. Or even if I can etch it, I can't fill it."

You need to make sure that the process you're providing, is integratable, because we're really selling process, not tools. Integratable is the key word. Sometimes the processes are standard materials, sometimes they use designer materials, but in all cases, these processes need to be integratable. So when I talk about low-k materials, the first thing I'm saying is the k-value, and the second thing is that we believe that it's integratable.

MICRO: Speaking of process integration, what kinds of activities are you doing in the Customer Integration Center (CIC)?

SOMEKH: We're doing three types of activities in CIC. The first and foremost is to work with customers on resolving integration issues. The second one is testing that our stuff is integratable, by making test structures and testing them. The third thing is looking at aspects of the technology; for example, there's an issue with copper that as you go to smaller and smaller dimensions, copper resistivity starts going up because there's more scattering.

There are a number of effects that take place when you go to small dimensions. First, the sidewalls are not smooth, and it gives you scattering and it slows the electrons down, and it gives you increased resistivity. The other one is you have more scattering from the grain boundary because the grain tends to be smaller. Because when you deposit copper, you deposit pretty small grains. But then you anneal it, and the way you anneal it, there's overburden, which is the thick copper before you wipe it off. . . and that thick copper, as you anneal it, starts forming the large grain and these large grains propagate into the small trenches and eventually stay. As dimensions get smaller and smaller, these large grains don't propagate that easily.

MICRO: So they don't fit into that space.

SOMEKH: Right. They can fit in one dimension but, depending on the aspect ratio, there's a problem. So these are some of the studies on the principles of the physics and the chemistry, that's the third thing we do in the CIC.

MICRO: How much similar work is going on with your customers at their  own sites?

SOMEKH: We do that as well. An advanced customer wants to have differentiation in technology, they don't want to sit and say, "why don't you come and give us everything?" They want to get the best out of everybody and come out with their own differentiated technology. There's a great deal of first-cycle learning that also happens at the customer site.

MICRO: Has that type of learning also increased the customer's willingness to invite you in. . .as an expert in that part of the process?

SOMEKH: I think so. Since we opened up to the customers, we have shown a lot of internal, good technology. The customers' appreciation for our technology has increased, both the interest in the specific technology and also in the fact that they see us more as a partner that can put a great deal of weight, innovation, and drive behind some of the issues that they have. I think our success with UV cure, where the rest of the industry was going somewhere else and now everybody's switching to UV cure, is a major indication of the good technology that we have, and also of the traction that we're getting with customers.

MICRO: What about in terms of alliances or partnerships with fellow equipment and material suppliers. How's that progressing?

SOMEKH: There are a number of alliances and collaborations that are very important in our industry. First, you see it at the level of our customers, where there are alliances for developing technology. Then we talk about the need to have interaction between chip manufacturers and equipment manufacturers early on to do that. From our side, there are two kinds of interactions, [including] one with our supplier.

The first thing I did after joining Novellus was to visit customers, but the second thing was to visit suppliers. Subsystems, material systems are very important. We want to develop a culture and cherish our interaction with them and their contributions. Then the third thing is the horizontal collaboration, which makes a great deal of sense, and the CIC is really the focal point of that.

MICRO: You mentioned that seeing your suppliers was very important when you started at the company. One thing I learned about Novellus years ago was the idea of systems integration being a different way to put your tools together, an approach that was quicker but also meant more reliance on the subsystems suppliers because they had to get it right before you put it together in your manufacturing facility. So that's more important than ever.

SOMEKH: People used to say that they measure a person by the people he or she hires. I think the same thing applies to suppliers. You're so dependent on the suppliers, and a good supplier versus not such a good supplier makes a big difference for your success.

I used to be responsible for engineering and manufacturing for one of the systems at one time in my career [at Applied]. I started reading all these Japanese books about cycle-time reduction, just-in-time deliveries, and so on. And I did a number of things. We built a factory with no warehouse. That was a platform. The first platform in, and there was no warehouse, it was direct delivery from the supplier to the factory. And I took a group of suppliers, and we went all over the country and visited different suppliers, talked, and compared notes. For about 15 years in a row during Semicon, I would have a supplier party in my garden at home, and all the key suppliers would be invited. If one day a supplier didn't get an invitation, it became a big issue. (Laughs) So that was an appreciation of the importance of suppliers.

MICRO: Where do you think your suppliers still need to improve?

SOMEKH: We're in a cost-driven food chain. People will do whatever it takes to reduce their costs, and I think suppliers need to realize that we are [doing this]. I think they do, because we as suppliers are in the tip of the bullwhip. We're swinging more than our customers, and our suppliers are swinging more than us, in year-over-year changes in business level, so we really need to accommodate that, and we need to figure out smart ways to ride the waves without getting destroyed by themÉ. Besides the fact that we need to ride the waves, we've got to think about cost reduction and productivity enhancement.

The last item for suppliers is quality. We've got to continuously worry about quality. We need it because we sell the most expensive industrial machinery in the world and [a lack of quality] is just unacceptable.

I have this car, a Prius—the Toyota hybrid car. It is superb from an engineering point of view. It's beautiful! I fit in it much better than I fit in the Q45. The seats are comfortable, every aspect of the car was thought through. It has all the latest electronics, Bluetooth enabled, remote key entry. I leave the key in my pocket, it knows me, it opens the car. I dial from the car and call out, but there's no phone in the car. It connects to my cell phone in my pocket, and it calls to my pocket. It calls out. It has GPS, voice recognition, everything. And it gets 45 miles per gallon. Just superb.

So then I drive it for six months. I can't stop admiring it. I sold like 10 of them to friends and family by showing it and doing test drives and so on. Then one day the display dies. And then it puts out a message—the system is not connected. This car costs $25,000, and you expect it to be perfect. You're really disappointed that you had one failure after six months, it's just unheard of.

Our customers pay $5 million for a piece of equipment, and they have every right to be disappointed when it keeps failing. I mean if it fails once a week, we feel really good about it. So I think quality, when it comes to reliability, particles, whatever, all the things that we can improve from a quality point of view, is a major differentiation for us, and it's a major differentiation for our suppliers.

MICRO: In your supply chain, are you relying on a single source in a lot of cases or do you have multiple sources across the board?

SOMEKH: I think that it varies from one subassembly to another, but we work very closely with our suppliers and we have a great deal of loyalty to them.

MICRO: What do you hear from your customers? What are they asking you for? Obviously the cost reduction is going to come back around.

SOMEKH: With the customers, the technology race continues. I spend a great deal of my time in the company focused on two aspects: one is the product business units, products in the business units, and products in the company; and the other is interaction with customers on the technology and joint-development-project level, as opposed to the commercial and other levels, in order to do that. The challenge for us as a company is to substantially increase market share in the areas where we have developing businesses, like PVD and CMP. We need to do that through product excellence and through this close interaction with customers.

MICRO: One regional market area that would have lots of room for growth would be Japan, which is traditionally not a Novellus stronghold, but more of an Applied stronghold.

SOMEKH: Right.

MICRO: It goes back to your mention of globalism as being one of Applied's strengths. Would it be fair to say that at Novellus, you're getting the point across that the company has to be increasingly global?

SOMEKH: Clearly that's what we intend to do, and we had a major milestone in the company. We hired Fusen Chen to be responsible for Asia operations, and I think this is a major step forward for us as a companyÉ.I hired him at Applied Materials 10 or 11 years back, and he had come from a customer. He's very knowledgeable about integration. He ran the PVD organization at Applied in recent years and has helped grow it to a very formidable position.

Fusen's joining us is an indication of a number of things. It highlights our focus on globalism and on Asia that we have, because a large portion of our revenue is really Asian. And it marks a much greater sophistication. Let me phrase it more accurately, it really allows us to ratchet the capability of the company between the different sections of the company, whether it's the operations in the field like Asia and the business unit and have an interaction that allows the company to get stronger and stronger to do that. I'm really thrilled with this assignment for Fusen, and I think it'll strengthen our position in Asia and, through this interaction with the business unit, strengthen the company as a whole. Because you need to have good products. You need to have a strong presence close to the customer, and Fusen will allow us to achieve that, and then ratchet both things up.

MICRO: Can you describe your working relationship with Rick Hill?

SOMEKH: That's a deep question. When you bring somebody in at a high level, what are the chances that people will work well together? If they work well together it becomes like a good marriage, and if they don't work well and they're both high-level could be very risky. So luckily Tom St. Dennis was here; I hired Tom (at Applied), so he was the guy who suggested what would work and what wouldn't work, and he was the guy who was able to tell how well it would work. But it's really amazing.

First, let me tell you a quick thing. If you try to look at different people, Psychology 101 kind of divides people into different drawers. You have people who are very structured. These are the organized guys that always have a staff meeting with agenda ahead of time and minutes, and it's all scheduled and so on. Then you have these very innovative, expressive guys that, come Monday morning, say, "hey guys, let's have a staff meeting, I had some good ideas over the weekend." So these are two types of personalities, and usually a person is either structured or expressive. But you cannot have a person who is strongly both.

On the other hand you have a person who's the driver. The first thing on his mind is, "how quickly can we do this? Why isn't it happening?" This is the guy who comes in first to work in the morning, leaves last, goes home, does e-mails, whatever. Then you have the guy who's more of an integrator, who worries about how people feel as opposed to what they do, who worries about how to do it, who will do it, and so on. He's more of the amiable kind of guy.

So the four types of personalities in the psychology book are driver, amiable, then analytical and expressive. You can have a person who is either analytical or expressive, and then that person can be either a driver or an amiable guy.

How these different personalities interact with each other is an interesting thing in and of itself. For example, suppose your boss is a driver and an analytical person. You come to him and say, "I have this new idea. I want to spend a million dollars to do this, this, and that." The first answer he'd say is no. Because he's very structured. Then you convince him and he might say yes. So for a structured guy, "no" is "maybe" and "yes" is "yes." You can take it and go to the bank.

But say you go to your supervisor, and he's a very expressive guy. And you say, "I'd like to spend a million dollars." He says "yes," meaning, "let's talk about it." You go spend the money; he says, "what did you do?" For this guy, "yes" is "maybe" and "no" is "no." You go to a driver and this is, no time for this; "yes" is "yes," "no" is "no," that's it. You go to the amiable guy who's more of the integrator, more of the politician, who worries what everybody else would say, so for that person "yes" is "maybe" and "no" is "maybe."

It's very important in the interaction to try to understand what the personalities are. If you have two guys at the top who have the same personalities, that helps, because they have the same language. It's not like one wants to do things and the other wants to worry about what the world would do and how it will happen. Or one guy every morning has another idea.

Look at the merger between Chrysler and Mercedes. It was hailed at the time as a merger of equals, and it just collapsed in no time. All the Chrysler guys were thrown out, Mercedes was sued, the whole thing. There was an incredible article that fits what I was describing here. The Chrysler guys would say, "We don't know what they want from us. They bring us into this room, they call it a war room, they want us to do a five-year plan!" And the Mercedes guys would say, "We tell them, it's only Wednesday. Why do you want to change what we decided on Monday?" (Laughs) So you have that issue when the different types of personalities are involved.

Rick's a very expressive driver. He wants to see things done quickly and he's an idea kind of guy. I happen to be very much the same. We exchange ideas, we think very much alike in different areas, we build on each other's ideas. . .There's a creativity that happens when we interact; we're both action oriented, we both want to see it quickly and so on. While a marriage like this could be very risky, since people are set in their ways, it so happens that it's just outstanding, and we're both tickled pink with it. Tom St. Dennis was the guy, the marriage counselor, who knew it was going to work out ahead of time, so he blessed it and that was good. . . .

I've been telling customers that through this transition, I feel like I had to go through a divorce. I got remarried, I have a new family, and now I have these two kids I need to bring up again: PVD and CMP, because these are the two products I personally shepherded at Applied. But the good news is, I feel 15 years younger.

MICRO: I was going to say that first, you had a separation from Applied, and then the divorce came, because it happened in two stages. It's interesting though, with the CMP group especially, where Applied has achieved such a dominant position except for certain regions where Ebara has some strength . . .Having to take that under your wing and then reposition it a little bit and then add some new stuff to it will be very challenging.

One of the other groups I want to discuss is the surface treatment (group). Surface treatment also involves wet processes, and I wonder if that's really a gap in Novellus's portfolio as well. It also raises a question about the future of small- and medium-sized companies in this business, since most of the cleaning companies are small- and medium-sized companies, except for DNS.

SOMEKH: Right, except DNS, and Tokyo Electron also has a presence in the cleaning side. . . .I mean you have a very good point. Applied has this product called Oasis. It hasn't grown substantially yet. And we actually have cleaning technology as part of CMP. But the challenge for us now is to focus on PVD and CMP, and bring the market shares up.

In spite of this vision for integration of dry and wet, it really hasn't happened. Mattson was the first company to jump into that, only to get rid of the wet later on. So we're very careful in jumping into that. The strip tools are used extensively. None of them at this point is really integrated with wet clean. The attempt to integrate it through the Oasis thing hasn't gone anywhere yet. So at this point, we're happy with what we're doing, and we have our priorities very, very clear within the company.

Another thing to put in perspective is innovation. We talked about new technology. Technology is innovated by people. You need to build the atmosphere to have a great deal of innovation to attract and retain innovative people when you do acquisitions, to be able to bring innovative people in and allow them to flourish within the company. That's one of the things that some companies do well and other companies, as they grow, it becomes more and more difficult to do. Take the famous story about Xerox. It all started with innovation but when they grew to be very large, while they were innovating inside, they weren't commercializing it, and other companies ending up commercializing it instead.

You need to know how to innovate and how to commercialize it and how not to view new products and new areas as rounding errors in your P&L statement and just forget about them. This is one of the things we're doing well at Novellus. I talked about the great deal of technology we have in the company. I talked about the acquisitions that we've made in order to plug holes in our roadmap, like the Angstron acquisition, and our ability to attract people from all over the world. We reach an agreement on how to bring their ideas to fruition, and they don't need to be employees and they don't need to own companies to do that; they just need to be innovative, with good ideas.

I see that also as a major strength, that mindset, and then the support for it from top management. Both Rick and I are highly technical and value this kind of interaction. We really encourage people to be innovative and allow them to turn [their innovations] into products. That's another thing I'm really delighted about.

MicroHome | Search | Current Issue | MicroArchives
Buyers Guide | Media Kit

Questions/comments about MICRO Magazine? E-mail us at

© 2007 Tom Cheyney
All rights reserved.