INTERVIEW
came active devices. Junctions were introduced into the waveguides to create modulators, while germanium epitaxy, already available at imec for electronic applications, could be adapted to photodetectors. At the same time, we never forgot III-V materials. From the beginning, we understood how attractive it would be to combine their direct bandgap and light-generation capabilities with silicon photonics. For roughly the past twenty years, the group has therefore explored many different approaches to III-V / silicon integration, from bonding to micro-transfer printing.
Looking back over your career, which scientific contributions do you consider the most significant? That is not an easy question because we always had a tendency to explore many different directions. There are nevertheless several areas where I believe the Ghent group made important contributions. One is grating couplers. In the early days of silicon photonics, there was a real concern that efficiently coupling light between optical fibres and these extremely small nanophotonic waveguides could become a showstopper for practical applications. We saw this as a risk for the field and invested considerable effort in coupling methods. Starting from relatively simple structures, we progressively improved their performance to coupling losses around 1 dB. III-V / silicon integration is another field in which the group has been very active for many years. We explored different approaches, initially including die-to-wafer bonding and later micro-transfer printing. The objective throughout was the same: to combine the best properties of the different material systems. Another example I particularly like is optical beam steering with phased arrays. Between roughly 2005 and 2010, one of our PhD students worked on silicon-photonic phased arrays using thermo-optic tuning. I believe this was among the earliest demonstrations of beam steering with silicon photonics. But at the time, we had many discussions about whether it would ever really be useful! We were thinking about short-distance free-space communications inside a room; LiDAR was not yet the obvious application it later became. We made the mistake of not continuing the work after that PhD. For five or ten years, the papers were hardly cited, and then suddenly the citations started to explode. It is a nice illustration of how difficult it can be to recognize the future importance of a technology while you are developing it. We also worked extensively on arrayed waveguide gratings and integrated spectrometers. That became particularly interesting to me because these devices could later be used for spectroscopic and biomedical sensing.
In the later part of your career, you increasingly focused on sensing and biomedical applications. What attracted you to this field? To some extent, this goes back to my time at Stanford. I took quite a number of courses in biomedical engineering there, and at that stage I really had two main scientific interests: integrated optics and biomedical engineering. Later in my career, biomedical optics offered a natural way to bring these two interests together. More broadly, I genuinely believe that sensing represents a huge opportunity for integrated photonics. It has not yet materialized on the same scale as communications and interconnects, but I am convinced that it will. There is an important difference between the two markets.
Why has integrated photonics developed more slowly for sensing than for communications? In communications, once you reach a certain combination of data rate and distance, you essentially have to use optics. In sensing, the situation is very different. For almost any sensing problem, there are competing technologies, many of which have nothing to do with light. So if you want to introduce an integrated photonic solution, you need a very convincing reason to do so, a clear advantage in performance, functionality, size or cost. That makes adoption slower. Nevertheless, I remain very optimistic, and I believe that in the coming years we will see several sensing applications scale significantly.
How has your relationship with industry evolved throughout your career? The Ghent team has always occupied a somewhat unusual position because we are both part of Ghent University and part of imec. Compared with imec as a whole, however, our activities have remained much more academic. Our research has relied strongly on public funding, particularly European programmes. We have been fortunate to obtain several ERC grants within the team and to participate in many European collaborative projects. These projects often included industrial partners, but our research agenda was not primarily driven by bilateral industrial contracts. This gave us considerable academic freedom. Of course, collaborative European projects have milestones and deliverables, but they still leave substantial freedom to explore new ideas. At the same time, we are an engineering group. Even when our research is quite fundamental, we like the idea that what we develop may eventually become useful and have societal impact. I think this balance between academic freedom and an engineering mindset has been extremely valuable.
How do you view the role of entrepreneurship in academic research? Whenever PhD students approached me with an idea that might become a product or a spin-off company, I tried to support them as much as possible. Being connected both to Ghent University and imec provided a strong environment for taking an idea from research and intellectual property towards investment and eventually the creation of a company. Seven spin-offs have emerged from the group, with additional initiatives currently developing. I deliberately chose not to join these companies operationally. My academic responsibilities were already substantial, but I was also concerned about possible conflicts of interest if I had one foot in the university and another in a start-up. I have nevertheless remained closely involved with several companies as an advisor or board member. Creating a company is never easy, it is always a challenging adventure, but it is extremely rewarding to see research move
16 www. photoniques. com I Photoniques 140