INTERVIEW
research topic was a niche field but we managed to publish solid work, though not“ super high-impact.”
What was the next step to build on this successful position? I applied to many positions in Europe and I got 2 + 2-year fellowship in Jena that had exactly the infrastructure I needed: two large optical tables, the right laser, and clean rooms, so I could invest in people. I carried out research in nonlinear nanophotonics, frequency conversion, and microfluidic integration for on-chip analysis. Our third child, my daughter, was born in 2012 in Weimar. The lab was very well equipped, and the four years went by quickly.
How did you transition to ETH Zurich? I obtained an SNSF Professor position at ETH, a four-year, non-tenure-track appointment, hosted by Jérôme Faist. After three years, the department opened a women-only call, and three of us were hired on tenure track in 2018. I was fasttracked, became associate professor in 2021, and full professor in August 2025.
What was your research program at ETH? Scientifically, I remained focused on nonlinear materials and moved into thin-film lithium niobate, TFLN. Early films, homemade in Jena, would break, so I sonicated damaged chips to make nanowires and nanoparticles for nonlinear optics. Back in Zurich, in 2015, we first tried to fabricate films ourselves, and then obtained commercial films around 2016. The obvious first demo on TFLN is the electro-optic modulator based on a Bragg design to add a twist to the typical Mach-Zehnder ones. ETH’ s critical mass of researchers matters a lot: other labs had 100-GHz test benches with unique, very high-end setups. That ecosystem is essential for a young group to ramp up.
What are your main research lines today? I tend to run two main axes and let good ideas in as long as they touch χ( 2) nonlinear materials. That is my core and it keeps the group curious and agile rather than locked into a single track. The first track is a very applied, fabrication-intensive axis: integrated photonics on TFLN, that spans from classical electro-optic modulators to integrated quantum photonic devices with many options of technology transfer into start-ups. The second is a more fundamental axis, nonlinear nanophotonics including scattering in disordered nonlinear media, lighter on fabrication but rich in physics. ETH’ s base funding let us sustain both. And I have always enjoyed writing grants, it is my creative moment, building the architecture, blending ambitious and safer goals, distributing roles, setting milestones. Experimental discovery is wonderful, but the making of a research idea is also very stimulating.
How are those axes organized in practice? Integrated photonics on χ( 2) materials mainly relies on state-of-the-art nanofabrication, to achieve on-chip periodic poling and electro-optic modulation. We work with monolithic integration rather than hybridizing with other materials, at least for now. Target applications are telecom, classical and quantum optics. The platform is transparent from the visible to the mid-IR, which let us go far beyond 1.55 µ m. This is all top-down in-house fabrication in our cleanrooms. I recruit PhD students who are willing to dedicate time in the cleanroom to fabricate for themselves and others. We do not keep know-how with one technician. We pass it peer to peer, from senior PhDs to newcomers, which naturally forms a 4 to 5 person subteam that handles training and handover of process flows. Second, for the bottom-up approach, we do not synthesize nanoparticles and we have a long-running informal collaboration with a team in Italy. Those χ( 2) particles allowed us to study absolute nonlinear responses in complex media, which is unusual because films and random assembly often hide geometry and volume effects. We are also working on solution processing of nonlinear materials suitable for integrated photonics, which represents a convergence of our two research areas.
How do you achieve absolute measurements in scattering media? We wanted to control the probed volume and the structure, not just average over an uncertain film. Classical drop-cast layers suffer from coffee-ring and thickness uncertainty. We therefore built self-assembled microspheres from the particles, measured their volume by electron microscopy, and even made cross-sections to estimate filling fraction and porosity. With precise volume and geometry, we could model the signals and, demonstrate random quasi-phase matching. I am proud that we pursued absolute nonlinear measurements under femtosecond pumping and then matched them to models we developed in-house when no one wanted to do the theory. There was skepticism from colleagues trained on bulk crystals who expected random media to wash things out. Getting the numbers right was a long, careful process, but it convinced people.
How did this expertise in nonlinear optics feed back into integrated photonics? In integrated photonics I initially wanted to stay with electro-optics only, not frequency conversion. One PhD student kept pushing for periodic poling. I resisted, thinking there were groups with decades of head start. This PhD student convinced me eventually and we developed a poling setup. It took about three years to make the process robust. After that, conversion efficiencies were strong, CW operation became natural, and we achieved spontaneous parametric down conversion( SPDC) onchip. In parallel we built the quantum benches, which also took two to three years to reach the level we wanted. Such a sensitive setup with low temperature detector in the near infrared allowed us to explore SPDC from III-V nanowires, whose nonlinear tensors are roughly one order of magnitude larger than LiNbO 3. A single nanowire a few microns long and a couple of hundred nanometers wide can already generate a high rate of entangled photons, which is conceptually very appealing.
14 www. photoniques. com I Photoniques 134