What’s Next In R&D?

Imec CEO sees AI, auto, EUV and DNA storage.

popularity

Luc Van den hove, president and chief executive of Imec, sat down with Semiconductor Engineering to discuss R&D challenges and what’s next in the arena. The Belgium R&D organization is working on AI, DNA storage, EUV, semiconductors and other technologies. What follows are excerpts of that conversation.

SE: Moore’s Law is slowing down. And it is becoming more expensive to move from one node to the next. Yet, R&D is not slowing down, right?

Van den hove: It’s getting harder and harder, so we need to do more and more R&D. There is one phenomenon that is happening. As the problems get harder and more complex, it’s even more important to align the value chain around those R&D challenges. It’s not a problem with one company anymore. Because things take time to develop, you have to make sure the entire value chain is aligned, so that you can go the same direction. And that’s a big part of the role we are trying to play. We are trying to make sure we bring all of the key stakeholders together and align them on the roadmap.

SE: What’s changed in the R&D arena?

Van den hove: In the past, we had a technology roadmap. We had the applications. Then, we have this phase of design technology co-optimization. Now, we are getting into system technology co-optimization. So that requires even more of a need to align the value chain. We need to work with the suppliers of the materials and equipment as well as the foundries, IDMs and fabless companies. Now, we work with even the systems companies. That’s why are engaging with people from car manufacturers, Google and Microsoft. It’s very important to bring the entire value chain together. We have to link the system to the technology.

SE: Over the years, Imec has been able to address these changes and grown in the process, right?

Van den hove: We are continuing to grow. We are doing extremely well. The breadth, of course, is increasing. Our critical mass is increasing. We are now at 3,500 people. We also link with the universities, so we can bring in a lot of new know-how and skills.

SE: There are a multitude of different applications that are emerging, such as AI, machine learning, biotech, sensors and many others. How does Imec decide what to pursue?

Van den hove: It’s a top-down multiple strategy approach that we take here. On one hand, we build roadmaps. We try to build strong offerings in some of those domains. We try to identify where our technology can be leveraged to the maximum. One of those domains is related to life sciences and healthcare. Around that topic, we have a lot of assets. We also have a very strong biotech and pharma community here. That is certainly one of our priority focus areas. But also, automotive is very important to Europe. That’s another area we are focusing on. We also have projects around IoT in general. We have a focus around smart cities.

SE: It’s impossible to work on everything, right?

Van den hove: Of course, we cannot do everything at Imec. We are quite selective in trying to identify those projects where we can leverage our unique technologies. We should not do projects that other teams can easily do. There is so much to do, so let’s focus on the really tough challenges.

SE: Imec is known for its R&D efforts in the semiconductor field. Imec is also working on other technologies, such as DNA storage. What’s that about?

Van den hove: I would consider this as a long-term project. We are not going to solve these challenges in a few years. It’s maybe five or ten years. What we need today is a solution for a long-term archive for the massive amounts of data that we generate in the world. I’m sure you have had this problem. You want to retrieve a picture that you took two to three years ago. You stored it in your disk, but you cannot read it anymore. Archiving all that information is going to be important. DNA is a very interesting medium to do that. We can store a few petabytes of information in just one gram of DNA. You can imagine storing the information in entire world in a fraction of what digital can do.

SE: What are the challenges here?

Van den hove: The process is to synthesize the DNA, and to write and build up the DNA strands in the right sequence. It’s not an easy process. We will have to accelerate this and we will have to massively parallelize these reactions through these very sophisticated systems.

SE: What about AI?

Van den hove: AI and machine learning are becoming extremely important. But again, we are linking this to the hardware with new processor configurations that can enable AI, for example, at the sensor nodes or at the edge.

SE: Apparently, there is a major shortage of engineering talent. Attracting new engineering talent is very difficult in the semiconductor and other industries. Any thoughts?

Van den hove: Indeed, it is a big challenge. The amount of engineers we need to tackle these big problems is huge. This is one of the roles that Imec is trying to play. We are trying to attract more people to get interested in this domain. And we are succeeding pretty well in doing that. For Imec, I cannot say this is our number one problem. With our links to many universities, we can attract good talent. But it’s certainly not straightforward for the industry. As a whole, it is a challenge.

SE: How is Imec addressing this issue?

Van den hove: We are working very closely with universities. Locally, we are working with schools to focus on STEM, that is, science and technology. It is an area that is high on our radar.

SE: What worries you or keeps you up at night?

Van den hove: Of course, there are always challenges. There is a lot of consolidation in the industry. The risks in the industry are increasing. If one company makes a mistake, it may have catastrophic consequences.

SE: Does Imec compete with the other research institutes in the world?

Van den hove: Of course, there are some very strong research organizations. They are also focusing on certain aspects of what we are doing. We also want to form partnerships with those other research centers, because we can’t do everything alone. The challenges are so big that we better align and partner with others, because it’s getting so difficult. If you look at everything that will become possible with AI and data, we need a lot of solutions to store, connect, compute and make sense out of it. You better align these efforts. Competing at the research level doesn’t make too much sense. There is enough competition on the application side.

SE: Last but not least, what about extreme ultraviolet (EUV) lithography?

Van den hove: I’ve been saying this for many, many years: ‘EUV is coming and there is a lot of progress.’ Of course, it has made tremendous progress over the last couple of years. But now, we are really there. The tools are shipping. All of the major companies have committed to EUV.

SE: What are the remaining challenges?

Van den hove: There are still a couple of worries with the resists and the pellicle. But also there, we are making progress. So with EUV, it’s not a matter anymore of whether it will happen or even when it will happen. It is happening now.



1 comments

GL says:

If it’s happening now, don’t waste any more time on EUV.

Leave a Reply