Marc Jankowski on Nonlinear optics, Supercontinua and Temporal Trapping

Marc Jankowski joined NTT Research as a research scientist in 2020. He has a Ph.D. in electrical engineering from Stanford University and explores the intersection of nonlinear dynamics with optical systems, particularly nanophotonic devices. Dr. Jankowski’s work with the NTT Research PHI Lab addresses such questions as: What are the ultimate limits, in terms of size and energy requirements, of nonlinear photonic devices? What dynamical regimes are still undiscovered? How can these nonlinear devices and processes be used to enable new technologies, such as quantum and classical light sources? For more on his background and ongoing research, please take a look at the following Q&A:

What led you to the field of nonlinear photonics and more specifically to the new approaches to optical frequency comb generation that you explored in your dissertation?

I had a bit of a meandering path to nonlinear photonics. Initially, I was studying alternative energy and sustainable design, which led me to work on biogas reactors and photovoltaics. I later joined an optoelectronics group thinking I would work on LEDs or photovoltaics, but I instead got involved in a project focused on quantum light-matter interactions in microstructures (in particular, an effect commonly referred to as strong coupling in microcavities). This led me to an interest in sources of nonclassical light, such as single photons and squeezed light, which ultimately spurred an interest in nonlinear optics since many common nonlinear interactions can be used to make nonclassical light. When I started my PhD, I found that there were an extraordinary number of problems to be solved in classical nonlinear optics. On the one hand, the field was over 50 years old, and many crucial problems had been solved. On the other hand, many basic questions were still unanswered. Simple things such as the nonlinear coefficients of common optical materials, and the mechanisms for forming short pulses of light in nonlinear resonators were still poorly understood (or at least commonly debated). Ultimately, I became very interested in these problems since even the classical dynamics were quite rich, and in many cases, unresolved. Certainly, one needs to be able to tame these things before they can think about making quantum devices.

Why did you decide to join NTT Research? And how do you see your work within the context of the PHI Lab goal of advancing information processing beyond the current state of the art?

Nonlinear photonics is an extraordinary enabling technology, and the work we do at NTT can revolutionize many fields. Regardless of the goal (information processing, sensing, etc.) the tools we develop may lead to a radically different world than the one we currently live in. However, this field is sufficiently new that very few companies have identified the impact of these devices and built departments focused on this kind of research. I joined NTT because this is the first real attempt to get these kinds of devices out of the lab and into the world.

Do you have any favorites among the academic papers you have co-authored? (The paper on “Ultrahigh-efficiency wavelength conversion in nanophotonic periodically poled lithium niobite waveguides” appears to rank high in terms of citations.)

For better or for worse, my favorite papers are the ones that haven’t been published yet. One paper is currently on the arxiv, called “Supercontinuum generation by saturated second-order nonlinear interactions.” Supercontinua look like rainbows of light, but they comprise coherent laser light, which enables you to use these sources for signal processing and sensing. A few years ago, we found that we could produce supercontinua with record-low power requirements in recently developed nonlinear photonic devices, and later found that the operating principles of these devices were rather different than any of the approaches that had come before. After some time, we have been able to develop some simplified models that made the behavior of these systems much clearer. We now know how these supercontinua form, why they look the way they do, and why they require so little power. We’re hoping these insights will enable more scientists to use this approach- many applications are enabled by supercontinuum generation, so having a new approach with well understood behaviors and low power requirements can be a powerful tool for many of the people in our field.

Do the new capabilities of the Sunnyvale Lab impact your work, or are you still conducting experiments elsewhere, maybe at Stanford?

Currently my experiments are all at Stanford. We have some experiments planned in Sunnyvale for this summer that I think will greatly improve our ability to process and control the materials we use for nonlinear photonics. Lithium niobate (LN), the material we use, has many misbehaviors, and it can be challenging to use LN to build systems for computing things, since the LN itself can change quite a bit from batch to batch and day to day. As we continue to build up our capabilities at Sunnyvale, I believe we can tame these misbehaviors and truly bring LN devices to the wafer scale.

Is there any research you’re currently conducting that you’d like to mention?

I’ve been involved in a project that uses confinement in time, in addition to space, to enhance nonlinear interactions. In my mind, these methods represent the first approach that could utilize nonlinear optics in a meaningful way at the single-photon level. In principle, devices based on temporal trapping could be used not just for extremely low-power classical computers, but also for fault-tolerant quantum computers. We’ve only begun to identify the new types of devices that can be enabled with this approach and our ideas keep evolving very quickly, so I’m naturally excited about this work.

In your brief video on the PHI Lab website, you compare nonlinear optics to the overtones of audible frequencies. Slightly personal question: You wouldn’t have any sort of musical background, would you?

Not really. I played some piano and trombone when I was little and have banged around on the occasional instrument since, but I can’t say I have made anything interesting come out. It’s actually quite common for physicists and engineers to think in terms of frequencies, harmonics, overtones, and octaves since those ideas are applied in many different contexts.  Collectively, this toolset can be called Fourier analysis, or signal processing, and it can be used to understand to anything from vibrating strings and drums (music) to water waves, mechanical systems, electronics, light (both classical and quantum), and matter (both single electrons, which are waves, and solids, which are periodic structures). Outside of engineering and physics most people really only interact with these ideas in the context of music, so it’s a natural point to link what we do back to something more tangible