“You can’t sit on your laurels.”
Prof. Dr. Thomas MikolajickProfessor Mikolajick, did you ever think you would settle in Dresden?
If you'd asked me 25 years ago, I would have said no! In 1996, I entered employment in the semiconductor industry, initially at Siemens in Regensburg. With the establishment of Infineon Technologies, I then transferred from Regensburg to Munich. In 2001, my work at Infineon pointed me clearly in Dresden's direction. More precisely, Infineon's memory company (which was later broken up and declared insolvent as Qimonda) moved its technology development division to Dresden. Those like me who don't design chips at the computer, but instead work in technology itself, are dependent on having access to billion-dollar high-tech equipment. As such, in order to be able to spearhead the very latest developments, I moved to Dresden - again, with Infineon.
You mention Qimonda's insolvency... it wasn't the only company to go bust in this industry. Are such insolvencies simply par for the course or were people a bit too blasé in Qimonda's case?
When it comes to insolvencies, you have to differentiate between two factors, namely competitive conditions and funding requirements. As regards the former, the semiconductor industry is one that demands considerable investment. When the market is strong, all manufacturers rush to invest in new capabilities, and this then inevitably leads to significant overproduction which must then be purged from the market once more. However, despite these cycles, the semiconductor market has achieved two-digit growth over a period of several decades. The companies which manage to hold out stand to make a very healthy profit in the long run.
But it's already becoming apparent that there is less chance of securing funding today.
Indeed, the second factor is more unfortunate. As the semiconductor industry is so cost-intensive, it can't succeed anywhere in the world without subsidies. That isn't a bad thing in itself, because these investments pay off in the long term. There is, however, real competition for subsidies and every site has its own unique advantages. In Germany, for example, research projects are important factors to be taken into consideration in the funding of risky developments. As such, academic partners are paired with companies and contribute to corporate research projects. Even the Dresden success story of the 1990s would not have been possible without financial support from the state of Saxony and the German government. Indeed, it was thanks to this funding that Siemens and AMD, two big players in the industry, were able to establish operations here and thus attract a number of other companies to the area. The region's high-tech industry employs 50,000 people today, with 20,000 working in microelectronics alone. In the USA and Asia, however, this line of business is often regarded as very much a part of society and is therefore much better funded. This causes many European companies to fall behind their foreign competitors, move their operations abroad or be taken over by their North American or Asian counterparts.
What support do you have in Dresden to help avoid the application of your innovations being siphoned off by the USA or Asia?
Fortunately, things are still looking pretty good in Dresden, because we have the support of the city and the state of Saxony. That said, we rarely receive any support on a national level at the moment. On the EU level, the Key Enabling Technologies (KET) initiative has gradually set the funding ball rolling again.
You are Managing Director of the TUD's Nanoelectronic Materials Laboratory (NaMLab), a professor at the TU Institute for Semiconductors and Microsystems Technology, and also Director of the leading-edge cluster Cool Silicon, supported by the Federal Ministry of Education and Research - so you have three full-time jobs! Are you able to compartmentalise these different roles in your mind?
Each role necessarily feeds into the others; otherwise they would have no meaning at all. The two positions that really go hand in hand are the ones at the NaMLab and university. At the NaMLab, application is the central focus of our work, and it therefore often touches upon what's being researched at the institute where I teach. However, a company can cooperate much more flexibly and in a more structured way with the industry. As such, this approach is extremely successful. My role at Cool Silicon is a little different to what I do in my other jobs, despite the fact that the NaMLab is a key Cool Silicon project partner. Indeed, at Cool Silicon, where science and the local microelectronics industry converge, I'm also involved in site politics as a project coordinator. But that's a concern for me, anyway.
What's your opinion of Dresden as a science hub? Does this cooperation between academia and industry function more successfully here than in other places you've come to know?
Absolutely, at least compared with the other places that I know. The Cool Silicon cluster also contributed directly to the intensification of this collaboration. The people involved now know each other much better, understand in which areas they can work together and can also see where cooperation is not meaningful. As a result, expectations are clear to each of the partners and group projects can be started very quickly here. It is important to note that all parties gain something from the venture. After all, the companies involved are not participating out of the kindness of their hearts; they have their own interests in mind. This cooperation with research establishments is therefore ideally suited to semiconductor companies like Infineon and GlobalFoundries, because semiconductor technology is by its very nature innovative. The number of transistors on a chip doubles every one and a half years, and each time this happens, we have to push the boundaries of physics to the limit. You can't sit on your laurels; you must always remain committed to research.
Are there any other advantages to working in Dresden?
The dynamic environment created here by the many local companies operating at the forefront of the semiconductor industry is extremely exciting; that's the main advantage. But, of course, Dresden's appeal doesn't stop there. The city itself is just fantastic - just look at the Elbe skyline! From Dresden it's also not far to Saxon Switzerland and the Erzgebirge mountains, which increases the city's recreational value significantly. In addition, the city is just the right size; it is compact, but still a large city. And then there's childcare; it's much better here than in West Germany. I always bring that up when I want to convince someone to move to Dresden.
So it would be difficult to lure you away from Dresden?
There's always the possibility that an opportunity will present itself which is so good that it has to be genuinely considered. But if you want my prediction, I would say that I will stay faithful to Dresden for many years to come.
In your industry cluster you network with other clusters. In such cases, approaches such as "science speed dating" are also used. What is that exactly and how did you come up with this idea?
It was actually a rather spontaneous idea from a member of the board of Cool Silicon, but it was a real success. Our buzzword in the spring was ‘networking', and so we chose this very personal approach whereby two researchers would introduce themselves and their research - it is therefore essentially another form of brainstorming. We are always keen to try new and innovative approaches. Last year we also did an art project...
...the Cool Silicon Art Award. Was that a publicity stunt? Or do you really feel that there are profound relationships between these two creative outlets, art and high technology?
If I'm honest, opinions are divided on that within the cluster. Originally, we wanted to use this channel predominantly to present our research to the general population, but in doing so, you also begin a dialogue with a world which is initially completely foreign to you. For me, that was extremely exciting. Indeed, it opened me up to entirely new ways of thinking. We plan to continue doing it, at any rate.
As a research coordinator, do you ever still get to conduct your own research?
Not as much as I'd like. I rarely enter the lab these days. However, the support available to PhD students through the NaMLab is very structured, so I still manage to get involved in research. Of course, research coordination takes priority over research itself in that role. It's all about striking a balance, but for the moment, I am quite happy.
So that our readers can get an idea of what you do in the NaMLab, would you mind outlining one of your application-orientated projects?
One interesting project (of many) from the NaMLab world is ‘Cool Memory'. Under this project, we have taken a well-known material from the semiconductor industry - hafnium oxide - and modified it in such a way that it is ferroelectric. This process results in a material with two potential stable states of electrical polarisation which can then be used to store data. The process is not actually anything new, but has so far only been implemented with extremely complex materials. Together with GlobalFoundries, we're integrating the material into 28-nanometre technology. This has attracted a lot of attention internationally.
What is the advantage of modified hafnium oxide, exactly?
Every semiconductor chip contains a few bits of non-volatile memory. A variety of technologies are used for this today, but they each have a number of limitations, including the fact that they all work at rather high voltages. This affects the size and energy requirements of the components. However, this is not the case with our material. That's why we named the project ‘Cool Memory'. So, it may well be that GlobalFoundries will soon own the most cost-effective and energy-efficient semiconductor module in the world. This would represent an enormous competitive advantage and a huge success for us in our work.
Why are you cooperating with the industry on this project?
The semiconductor industry is hugely interested in this research, and each attempt to implement this solution into 28-nanometre technology costs more than a small car. Why GlobalFoundries? The innovation is particularly promising for this company because it's a foundry; it's a company which has grown off the back of working exclusively on behalf of other companies. As such, we're not talking about developing an individual product, but a basic technology which can then be used by customers in their own products. A familiar material which has been newly modified, and which can be developed according to a variety of specifications, is therefore highly attractive.
How close are you?
We've successfully proved that the technology works at the 28-nanometre scale. This is no small achievement. Even though everything worked perfectly in the lab - in large structures - these results can often not be integrated directly into a scaled semiconductor process. However, we have demonstrated that this is possible in principle. We have so far only manufactured individual bits for storage technologies, which is naturally not enough for any real-world application of this technology. Now we have to develop a storage matrix and activation system. That's the next step.
How long does this process normally take?
From the point we're at now, it would normally take three to ten years to reach the manufacturing stage of a project. In this case, I would say it will take three to five years, because the material is already being used in manufacturing; we only modified it.
How does this project relate to other Cool Silicon undertakings?
In Cool Silicon's ‘Area 1', which deals with micro- and nanotechnology, there was once a project called ‘Cool Computing'. Under this project, basic transistor technology was developed which we're now using in our Cool Memory work. Likewise, ‘Area 2' projects (information electronics) quite naturally draw upon these basic technologies, even though these projects focus more on system and chip design.
Do you find that your work is more interdisciplinary now than it once was?
Yes. Once upon a time, work in semiconductor technology was much more strongly segregated, into technology on the one hand and systems and design on the other. Today, however, you need an understanding of each of these three areas in order to develop something which will succeed in the market.
Can you give us an example?
We once thought, at least from a technical point of view, that storage technology had to function almost perfectly, else it would be unsalable. Today, however, USB memory sticks no longer work perfectly, but are delivered with errors and defects - and nobody notices this at all, because modern data storage devices contain powerful correction algorithms which ultimately display the data that should have been saved in the first place. In order to gain technical control, a developer must also understand systems; he or she must know what errors can and cannot be remedied. The cross-pollination of both of these areas is therefore essential.
Thank you for the interview.
- Professor of nanoelectronic materials at the Institute for Semiconductor and Microsystems Technology at TUD
- Managing Director of the Nanoelectronic Materials Laboratory (NaMLab) at TUD
- Director of the leading-edge cluster Cool Silicon supported by the Federal Ministry of Education and Research