Swaminathan (Sam) Sivakumar is an Intel Senior Fellow in the Technology and Manufacturing Group and Director of Lithography in the Portland Technology Development organization at Intel Corporation, leading the definition, development and deployment of Intel’s next-generation lithography processes. He is also responsible for developing resolution-enhancement technique
s and optical proximity correction technologies in support of Intel’s leading-edge patterning solutions. Chennai 36 sat down for an interview with him…
Congratulations sir on being awarded the DAA ! How do you feel on this occasion?
It feels both fantastic and humbling to be recognised in this way, especially by your teachers, as the famous saying goes in Sanskrit “Aacharya devo bhavaha” which means your teachers are like Gods. When you’re recognised by your alma mater for accomplishments in your field, it feels wonderful. It’s a great honor to be considered for the award.
It’s been more than 30 years since you graduated IITM. Can you share some incident or memory that stood out during your stay here?
I think without question, my IIT experiences were the highlight of my life. Everything that I have experienced here, from the wonderful friendship and comradery, working on different problem sets together with classmates, preparing for exams, all this was a lot of fun. A lot of these memories come back when I think about them. The Mardi Gras (which is now Saarang) was the major event of the year when everyone got together and enjoyed. We had all kinds of tournaments too, like the inter hostel roller hockey tournaments, intra hostel cricket tournament among the hostel wings. Whether it was studying for the exams or having fun, we were in it together.
You’ve been working in the field of lithography for 25 years. What drew you into the field when you joined?
I was intending to focus on electrical engineering from the beginning; transistors seemed to have captured my interest. I was interested in compound semiconductors and I had decided during my time at IIT that’s what i wanted to work on. I got an assistantship at the University of Illinois, which was one of the top institutions in that field at that time. But when I started working with transistors, the process of making the transistors fascinated me more than their electrical performance. I developed a keen interest in their fabrication, especially the optical process of lithography, which is one of the key processing steps. This change in my focus gradually happened as I finished my graduate school, and I started working on the process of manufacturing the transistor. So that’s how I landed in lithography and have been working on it for all of my career. It has been extremely interesting and rewarding throughout.
Did you have any idea that you would be working in lithography when you were in IITM?
I was pretty sure I would do something in the field of transistors and microprocessors. When I first started at IIT, I was more focused on computer hardware. My interests gradually changed more towards solid state devices and eventually their fabrication.
Sir, you became involved in industrial research and development at Intel right from the time in grad school. Could you tell us the differences in research in an academic setting such as a university and that in an industrial company like Intel?
Some differences are obvious. The focus of research in academia is answering fundamental questions about applied or basic sciences. The research in industry is almost always focused on creating some product that can be sold. That is the difference at a very basic level. I think two big things in industrial
research, budget and timeline, are something we don’t learn in school. You learn how strict timelines are when you’re working in an industrial setting. For example, the microprocessors that we make at Intel have to lineup with the product releases of computer manufacturers like HP, Dell or Lenovo, and they always time their products to be sold at times like the back-to-school season. So the computers would be in the stores by July, for example, so parents can buy them for their children, or maybe in time for Christmas, Deepavali or Pongal. The timeline of the market was extremely important so we have to plan our products such that they reach the manufacturers on time, so that they can make their laptops or desktops and sell it in the market. There are different groups working on different aspects of the product and they have to synchronise at some point in time which presents a different set of challenges that you have to look at. Working against the clock and getting things done on time is the single most important thing at an industrial work setting. This kind of schedule driven activity is something you don’t learn in academia.
You have introduced and developed EUV lithography at Intel. Could you tell us more about it and how it is different from conventional lithographic processes ?
Sure. lithography is basically like taking a photograph. You have a circuit on a radical, which is a piece of glass. And so if you look at a microprocessor, you may have millions of transistors, all wired. The wires are defined on this piece of glass, then you image this through pretty sophisticated optics and print this on the wafer. That’s how you get the chip created.
The smaller the features that you need to define, the lower the wavelength that you need to use. Lithography has moved progressively from 436 nanometers of light to 365nm to 248nm to 193nm. As we went down in wavelength we could define smaller and smaller features. The final step was to move to EUV which was 13.5 nm. It was an inevitable step in the progression over the last 25-30 years. Each of those steps involves a lot of development; from chemicals to different types of lenses, everything in the optics changes to deal with that specific type of wavelength. Each step has been a giant technical challenge and EUV more so than the previous ones because it is essentially a soft X-ray and it penetrates a lot of things that normal light would be blocked by. So the way we designed the optics was completely different. It was much more of a radical step than any of the previous ones. People have been working on it for 20- 30 years so it has been a long time coming.
Recently IITM launched India’s first indigenously fabricated processor called Shakti. Why do you think it took our country so long to start efforts in fabricating our own processors?
I think the main challenge with fabricating high-end integrated circuits is money. It costs a huge amount of money to make a fab where you can actually make computer chips. If you start from scratch, a top-end modern microprocessor today would probably cost you two billion dollars just to build the factory. It’s also heavily water-intensive.
Above everything else, India has a lot of human capital, a lot of smart people, so something like software, which requires a relatively small initial investment, you need a bunch of computers and some smart people to program, is a much more natural fit to Indian strengths than building computer chips.
The amount of financial capital needed to make cutting-edge microprocessors is so enormous that the activation energy to get over the first step for fabricating microprocessors was considerably higher than for software, which is why India is a powerhouse worldwide in software, whereas on the hardware side,
it’s a lot harder to break in when you have the US, and more importantly now, Korea and Taiwan with established chip manufacturing lines.
What is the future of semiconductor lithography and more particularly, what innovations at Intel are you most excited about?
This question gets asked around a lot, how long are things gonna keep shrinking? I think that the end is coming soon. We still have a few generations left, but the smallest features on a transistor are already 2 or 3 atoms wide, so there’s only so many more times you can shrink stuff before you run out of atoms. So though there will be scaling for a few more years, we are talking about being a lot smarter about how you scale, not just physical dimensional scaling, but there’s a lot of talk about 3-D stacking where you put one chip on top of another and then wire them. There’s a lot of research being done on 3-D stacking that hopefully will turn into continued value for the end user.
Scaling is always measured in terms of dollars per transistor, and it takes a lot of money to scale, so we should ask ourselves, why should the market pay to scale? Why should Intel or anybody else spend the money to do the next generation of scaling if we cannot sell products to pay for it? In order for us to continue to have the market be interested in our scaling, we have to give them some value. I think that value is gonna come from vertical integration, 3-D stacking and similar stuff. That’s really what I’m excited about, trying to see what’s out there in that area.
What advice would you give to the current students at IIT Madras?
The last few days, I wanted to see what your curriculum is like, so I looked around on the website. I am actually very pleased to see that you had classes on ecology, environment and similar things, and the fact that you had new majors that are beyond the old traditional engineering majors, which is fantastic. So the advice I would give, is really to take advantage of these opportunities to broaden your mind. Back then the Internet was there to send email, it was only a little while after I went to grad school that the world wide web really picked up to a point where knowledge is now easily available. The Internet is the great leveller. The world is a very different place as a student now than it was 20-30 years ago and what you realize is that the problems to be solved are interdisciplinary, whether you talk about climate change, traffic, more efficient cars, solar energy, or bioengineering for artificial limbs. I think those are really the kinds of fields that will make a difference in the next 20-30 years. So the advice I would give students these days is be more generalists at the undergraduate level. Try to learn more things, don’t try to focus too narrowly on things that you think you are going to do, because I guarantee whether you get a job, go to grad school, or do something else, your interests are going to change and a lot of things outside what you perceive as interesting, you will start getting interested in.
So, keep an open mind, take lots of classes, learn things in a variety of different fields, even if it is outside your core area, you can maybe take classes in bio-engineering, in mechanical engineering, in automotives, and similar stuff, that is if your curriculum allows you to take electives. You can do internships slightly outside you core area.
I think that’s where you will find the most gratification, in solving the world’s problems. We didn’t necessarily have those kind of opportunities, you would finish you first year of maths, physics and chemistry, and then jump into your major area of focus and that’s what you keep doing, but I think you guys have more flexibility, so think broadly.
On that note, what do you think of the interdisciplinary dual degree program provided here at IITM?
I think interdisciplinary education is fantastic, so if I had to go do it all over again, that is definitely something I would strongly consider for myself. A lot of problems these days, that have meaningful social impact, are interdisciplinary. This doesn’t mean that the core problems in the various areas are gone completely. Of course, you can always be a pure electrical engineer or a pure mechanical engineer as well, but I think that the most interesting problems are heavily interdisciplinary. Anything that promotes an interdisciplinary way of thinking is a great thing.